Home

ModelWorks - Terrestrial Systems Ecology

1. HALT END FALSE PROCEDURE SetCurveAttrsForLegend BEGIN SetCurveAttrForMV SetCurveAttrForMV SetCurveAttrForMV SetCurveAttrForMV SetCurveAttrForMV obsMod SetCurveAttrForMV obsMod END SetCurveAttrsForLegend obsMod obsMod obsMod obsMod yminDash turquoise spotted 0C ymeanDash turquoise dashSpotted 0C ymaxDash turquoise spotted 0C yminDashLn turquoise spotted 0C ymeanDashLn turquoise dashSpotted 0C ymaxDashLn turquoise spotted 0C PROCEDURE Output VAR k INTEGER t0 tend h er c hm REAL BEGIN k TRUNC CurrentTime 0 1 ensures correct rounding IF k gt kmin AND k lt kmax THEN yminDash yminD k ymeanDash ymeanD k ymaxDash ymaxD k yminDashLn Ln negLogDelta tyminDash ModelWorks V2 0 Appendix ymeanDashLn Ln negLogDelta tymeanDash ymaxDashLn Ln negLogDelta ymaxDash SetCurveAttrForMV obsMod ymeanDash turquoise dashSpotted 0C SetCurveAttrForMV obsMod ymeanDashLn turquoise dashSpotted 0C IF k lt limkmax THEN SetCurveAttrForMV obsMod yminDash turquoise spotted 0C SetCurveAttrForMV obsMod ymaxDash turquoise spotted 0C SetCurveAttrForMV obsMod yminDashLn turquoise spotted 0C SetCurveAttrForMV obsMod ymaxDashLn turquoise spotted 0C ELSE SetCurveAttrForMV obsMod yminDash turquoise invisible 0C SetCurveAttrForMV obsMod ymaxDash turquoise invisible 0C SetCurveAttrForMV obsMod yminDashLn turquoise invis
2. this model Selection of all objects of a list is possible by clicking the button All buttons with a down arrow 7 are used to set a current value whereas buttons with a left arrow are used to reset a value to its default as defined by the modeler The button T serves to specify which columns i e current values of the model objects are to be shown in the list R File Edit Settings Windows Simulation UF Monitorable variable names Ident Unit Monitoring x ER Logistic grass growth model I ei Grass G g dry weight m 2 T Y a l i Grass derivative dG dt g dry weight m2 9141919 DO Curves Minimum Hax imum D 0 000 1000 000 Fig T4 Initial screen of the ModelWorks simulation environment obtained immediately after starting the model definition program i e the module which contains the definition of the logistic grass growth sample model All four IO windows for the models the state variables the model parameters the monitorable variables plus the graph and the table window are open The latter two windows have been slightly rearranged from their default size and position in order to give a better view onto the IO windows The menu command Settings All above resets the program to its original state as it was at the beginning of the simulation session If you should loose the orientation during a complex series of interactive changes this command allows you always to resume a well defined state If you sh
3. used for error messages PROCEDURE DataFileNotOk BEGIN WriteMessage 2 3 Data file with observations WriteMessage 3 3 could not be opened END DataFileNotOk PROCEDURE NotEnoughDataInFile BEGIN WriteMessage 2 3 Not enough data in observation file END NotEnoughDataInFile PROCEDURE WrongNum VAR numStr ARRAY 0 3 BEGIN WriteMessage 2 3 Illegal number found CardToString k numStr 0 WriteMessage 2 3 30 numStr WriteMessage 3 3 Attempt to read WriteMessage 3 3 10 curVar END WrongNum OF CHAR year PROCEDURE InitData VAR f TextFile BEGIN Lookup f LBMObsSUE DAT FALSE IF f res done THEN FOR k kmin TO kmax DO r Response year CARDINAL IF EOF f THEN ShowAler GetCardinal f year curVar IF NOT legalNum OR k lt gt year THEN ShowAlert 4 5 t 3 50 NotEnoughDataInFile year t 3 50 NotEnoughDataInFile HALT END 0 WrongNum END HALT END IF EOF f THEN ShowAler GetReal f ymeanD k curVar Ymean IF NOT legalNum THEN ShowAlert 4 50 WrongNum END IF EOF THEN ShowAlert 3 50 NotEnoughDataInFile GetReal f yminD k curVar Ymin IF NOT legalNum THEN ShowAlert 4 50 WrongNum END IF EOF f THEN ShowAlert 3 50 NotEnoughDatalnrFile GetReal f ymaxD k curVar Ymax IF NOT legalNum THEN ShowAlert 4 50 WrongNum END END FOR Close f ELSE ShowAlert 4 40 DataFileNotOk END IF curActive END InitData HALT END
4. FROM SimBase IMPORT IMPORT SimBase FROM SimMaster IMPORT MODULE SubModDisc FROM SimBase IMPORT EXPORT DeclSubModDisc y VAR discM Model step newStep a flip PROCEDURE Dd BEGIN newStep flip step END Dd PROCEDURE Od BEGIN y a step END Od PROCEDURE ObjectsDisc BEGIN DeclSV step newStep 1 SetSimTime SetMonInterval RunSimMaster F F Fe F 2 5 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 5 Dec lM IntegrationMethod DeclSV StashFiling Tabulation Graphing DeclMV DeclP RTCType Model SetSimTime SetMonInterval NoInitialize NoInput NoTerminate NoAbout y REAL 0 1 0E3 1 0E3 Step of discrete time submodel Step DeclMV step 5 0 2 0 Step of discrete time submodel Step notOnFile writeInTable isY DeclP a 1 0 100 0 100 0 rtc Amplitude of step function DeclP flip 1 0 END ObjectsDisc PROCEDURE DeclSubModDisc BEGIN DeclM discM discreteTime ObjectsDisc NoTerminate 1 0 Factor to reverse sign of step function 1 0 EEE DL RE NoInitialize NoInput Od Dd Discrete time submodel DiscSubMod END DeclSubModDisc END SubModDisc FERRER X MODULE SubModCont RRRRRK FROM SimBase IMPORT IMPORT y EXPORT DeclSubModCont VAR contM Model x xDot r u REAL PROCEDURE Ic BEGIN u y END Ic PROCEDURE Dc BEGIN xDot END
5. ModelWorks 2 0 Reference resumes all monitoring exactly as it was before procedure SuppressMonitoring has been called PROCEDURE DeclClientMonitoring initmp mp termmp PROC Installs in ModelWorks a client provided monitoring procedures mp At the begin respectively the end of every simulation run the procedures initmp respectively termmp are called for exact calling sequence see Fig T17 in part Theory During the simulation run the monitoring procedure mp is called every time or integration step once mp will be called as the last monitoring procedure i e after ModelWorks calls the stash file the tabulation and the graph monitoring procedures This allows for instance to draw into the ModelWorks graph window from within the mp 7 5 3 STASH FILING PROCEDURE StashFileName sfn ARRAY OF CHAR Sets the default name of the stash file may contain a path e g Disk Folder MyFile DAT The call of this procedure will have no effect until the stash file name is reset Important notice if a file with the same name should already exist it will be overwritten without any warning PROCEDURE SwitchStashFile newsfn ARRAY OF CHAR Switch the stash file by closing the one currently in use if called between two simulation runs state No run Fig T26 and open a new stash file with the name newsfn may contain a path e g Disk Folder MyFile DAT at the begin of the next simulation run Calling this procedure in the middle of a simulation
6. PROCEDURE SetMissingReal missingReal REAL default 0 0 PROCEDURE GetMissingReal VAR missingReal REAL PROCEDURE SetMissingInt missingInt INTEGER default 0 PROCEDURE GetMissingInt VAR missingInt INTEGER Segments are data untis separated by eosCode PROCEDURE SetEOSCode eosCode CHAR default ASCII us unit seperator PROCEDURE GetEOSCode VAR eosCode CHAR PROCEDURE FindSegment segNr CARDINAL VAR found BOOLEAN first segNr 1 PROCEDURE SkipToNextSegment VAR done BOOLEAN PROCEDURE AtEOL BOOLEAN PROCEDURE AtEOS BOOLEAN PROCEDURE AtEOF BOOLEAN PROCEDURE TestEOF use only where you don t expect EOF shows alert TYPE Relation smaller equal greater PROCEDURE Compare2Strings a b ARRAY OF CHAR Relation CONST negLogDelta 0 01 offset to plot log scale if values lt 0 EEE IR RR IRR e e I IR Kk e He SimGraphUtils FROM SimBase IMPORT Model Stain LineStyle always with dialog normally no dialog used in dataF value used for an integer E e k e e IR KOTOR RI IK IK KK AIK QT value used for a real 37C E e E IIR ROK I He K E I ICR AO IK E A 181 ModelWorks V2 0 Appendix FROM DMWindowIO IMPORT Color TYPE Curve VAR nonexistent Curve read only PROCEDURE SelectForOutputGraph PROCEDURE DefineCurve VAR c Curve st Stain style LineStyle sym CHAR PROCEDURE RemoveCurve VAR c C
7. Stains and color variables from module DMWindowIO correspond to each other They can be paired following this sequence black white red green blue cyan magenta yellow Stain coal is black snow is white ruby is red etc The following line styles are available to connect points in the graph unbroken broken dashSpotted T spotted gt n invisible no drawing at all may be used to stop drawing of a particular curve while others are still drawn purge used to erase already drawn curves autoDefStyle line style will be determined by ModelWorks according to the automatic definition mechanism of curve attributes To set or get defaults respectively current curve attributes for the monitorable variable mv belonging to model m use the following procedures PROCEDURE SetCurveAttrForMV m Model VAR mv REAL st Stain ls LineStyle sym CHAR PROCEDURE GetCurveAttrForMV m Model VAR mv REAL VAR st Stain VAR ls LineStyle VAR sym CHAR PROCEDURE SetDefltCurveAttrForMV m Model VAR mv REAL st Stain ls LineStyle sym CHAR PROCEDURE GetDefltCurveAttrForMV m Model VAR mv REAL VAR st Stain VAR ls LineStyle VAR sym CHAR Where St Stain color is used to draw the lines and or plotting symbols of a curve ls Style of the connecting lines drawn between monitoring points sym Plotting symbol drawn at monitoring points The latter two procedures which affect the defaults req
8. Data from Fischlin A 1982 Analyse eines Wald Insekten Systems Der subalpine L rchen Arvenwald und der graue L rchenwickler Zeiraphera diniana Gn Lep Tortricidae Diss ETH Nr 6977 Swiss Federal Institute of Technology Z rich Switzerland 294pp page 90 Table 10 and from Baltensweiler W and Fischlin A 1987 The larch bud moth in the European Alps In Berryman A A ed Population Dynamics of Forest Insect Systems Plenum Press in print The data are read from a file only once at model declaration and are loaded into memory for subsequent usage This program module contains the model which runs under the simulation environment ModelWorks V0 5 Programming A Fischlin Systems Ecology ETHZ 01 05 87 CONST kmin 1949 first year sampled kmax 1986 last year sampled limkmax 1978 beyond limkmax yminDash ymaxDash no longer available yLL 0 0 minimum used on graph scale for larval densities yUL 600 0 maximum used on graph scale for larval densities negLogDelta 0 01 offset used to plot log scale if values lt 0 The following variables may be freely used in another submodel typically to compare simulation results of a simulation model with the observed values VAR yminDash REAL minimum annual value found in anyone site ymeanDash REAL average annual value for whole valley ymaxDash REAL maximum annual value found in anyone site yminDas
9. PROCEDURE NoDynamic PROCEDURE NoTerminate PROCEDURE NoModelObjects PROCEDURE NoAbout PROCEDURE DoNothing Modifying of models and model objects na ti Ca Ve i pit Ss il i Nn te A a ninth et PROCEDURE GetDefltM VAR m Model VAR defaultMethod IntegrationMethod VAR initialize input output dynamic terminate PROC VAR descriptor identifier ARRAY OF CHAR VAR about PROC PROCEDURE SetDefltM VAR m Model defaultMethod IntegrationMethod initialize input output dynamic terminate PROC descriptor identifier ARRAY OF CHAR about PROC PROCEDURE GetDefltSV m Model VAR s REAL VAR defaultInit minCurInit maxCurInit REAL VAR descriptor identifier unit ARRAY OF CHAR PROCEDURE SetDefltSsV m Model VAR s REAL defaultInit minCurInit maxCurInit REAL descriptor identifier unit ARRAY OF CHAR PROCEDURE GetDefltPp m Model VAR p REAL VAR defaultVal minVal maxVal REAL VAR runTimeChange RTCType VAR descriptor identifier unit ARRAY OF CHAR PROCEDURE SetDefltPp m Model VAR p REAL defaultVal minVal maxVal REAL runTimeChange RTCType descriptor identifier unit ARRAY OF CHAR PROCEDURE GetDefltMV m Model VAR mv REAL VAR defaultScaleMin defaultScaleMax REAL VAR descriptor identifier unit ARRAY OF CHAR VAR defaultSF StashFiling VAR defaultT Tabulation VAR defaultG Graphing PROCEDURE SetDefltMV m Model VAR mv REAL defaultScaleMin defaultScaleMax REAL descri
10. or without the termination condition returning true This command is the only one available to the simulationist to terminate a simulation interactively It may be chosen from both program states Simulating as well as No simulation Execute Experiment E Executes the currently installed structured simulation a so called experiment It enters the program state Simulating and calls the procedure which has been declared as the Experiment procedure by the model definition program see procedure DeclExperiment from module SimMaster If no such procedure has been installed this command will appear dimmed inactive and can not be chosen R 84 ModelWorks 2 0 Reference If the monitoring settings for the stash filing of at least one monitorable variable is set F writeOnFile a stash file with the current stash file name is automatically opened and data are written onto it according to the current recording flags see menu command Project description However this behavior depends also on the current simulation environment mode see above menu command File Peferences In case that the mode Always document run on Stash file is currently active the stash file is opened even if the stash filing is set for no monitorable variable Important notice Incase there exists already a file with the same name as the current stash file name this file will be overwritten without any warning Due to the nature of interactive simulation the overwri
11. 0 19 tempVect 5 40 0 grVect 5 0 26 tempVect 6 50 0 grVect 6 0 25 DeclTabF RTt tempVect grVect 7 TRUE RTt growth rate vs temperature Temperature Growth rate C h 0 0 60 0 0 0 1 0 SetDefltGlobSimPars 0 0 48 0 0 5 0 0001 1 0 1 0 END ModelDefinitions BEGIN RunSimMaster ModelDefinitions END UseTabFunc H 4 A MIXED CONTINUOUS AND DISCRETE TIME SAMPLE MODEL COMBINED MOD The following program code contains a sample model demonstrating the mixing of a continuous time submodel with a discrete time submodel Both models form together a structured model of mixed type see also in the manual Part II Theory in the section Model formalisms especially the subsection Structured models Coupling of submodels MODULE Combined F F u e Fe Fe Fe He de e de Fe e Fe He He Fe Fe e He Fe Fe He He Fe He Fe He Fe 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 20 Structured model built from a continuous and discrete time submodel The continuous time submodel consists of a simple linear differential equation whereby its paramater depends on an input which has been coupled with the output from the discrete time submodel The discrete time submodel contains a simple step function Every submodel is modeled as a local module af 29 Mai 1988 EEREREREREREEERAERE E R ANK K ANK AAE AK A EE REE KE A 155 ModelWorks V2 0 Appendix IMPORT SimMaster
12. 92 104 monitorable variable 20 21 22 93 111 monitoring interval 21 79 109 project description 79 109 recording flags 80 109 scaling 20 69 95 111 simulation start time 77 109 simulation stop time 77 109 table function 85 111 ModelWorks V2 0 Index modifying default value global simulation parameters 108 initial value 110 integration method 31 110 integration step 109 model 110 model parameter 110 run time change 110 monitorable variables 110 monitoring interval 109 project description 108 recording flags 108 scaling 110 simulation start time 109 simulation stop time 109 state variable 110 Modula 2 5 14 module structure of ModelWorks 65 monitorable variable 11 12 20 21 46 55 66 93 monitoring 20 55 58 93 97 117 121 monitoring interval 46 55 79 MS DOS 6 125 126 127 new value of state variable 11 12 26 39 66 101 102 Norun see substate of simulation environment No simulation see state of simulation environment object see model object object selection see selection of object output 37 42 53 54 101 output input coupling 38 page up 57 74 120 parallel model 5 41 164 parameter see model parameter Pause see state of simulation environment personal computer 5 preferences see simulation environment mode recommended preferences 74 removing of model 64 99 113 model object 64 113 reset 12 46 61 77 81 run see simulation run run time c
13. Appendix ytLn Ln negLogDelta tyt springEggs 1 0 cl et END Output PROCEDURE Dynamic PROCEDURE gmstarv x1 x2 REAL REAL BEGIN IF x2 0 0 THEN RETURN 0 0 END IF x2 gt 0 0 THEN RETURN Exp x1 x2 END END gmstarv PROCEDURE grecr def rt REAL REAL CONST eps 0 00001 VAR zrt REAL BEGIN grecr IF def lt p12 THEN IF rt gt p9 eps AND rt lt p9 rt p9 THEN RETURN 1 0 ELSIF rt gt p9 THEN zrt pl0 ABS pll rt rt p9 IF zrt gt rt p9 THEN RETURN p9 rt ELSE zrt lt rt p9 RETURN 1 0 zrt rt END IF ELSE warning rt lt p9 HALT END IF ELSE def gt pl2 IF def lt p13 THEN RETURN 1 0 def pl2 pl1 rt p13 pl2 rt ELSIF def gt p12 AND def gt pl3 THEN RETURN pll rt ELSE def p12 AND def gt pl13 HALT END IF END IF END grecr BEGIN Dynamic def 1 0 gmstarv pl rt p2 p3 rt et p4 et p3 rt et p4 et pl rt p2 rtl grecr def rt rt et1 1 0 gmstarv pl rtt p2 p3 rt et p4 et po rt rt rt po rt rt p7 rt ps et END Dynamic PROCEDURE ModelObjects BEGIN DeclSV rt rtl 15 0 11 99 18 5 Raw fiber content fresh weight rf DeclSV et etl 4765975 0 0 0 1 0E12 Larch bud moth eggs individuals eggs numbers DeclMV rt 10 0 20 0 Raw fiber content fresh weight rf S notOnFile writeInTable notInGraph DeclMV springEggs 0 0 1 0E12 Larch bud moth eggs i
14. AssignString nl nameStateOne AssignString n2 nameStateTwo AssignString n3 nameStateThree FOR is firstState TO lastState DO AssignString Initial prob of state StateToString is istr ident initP IntToString ORD is 1l istr 0 ConcatCh ident SetDefltP m initP is initP is 0 0 1 0 rtc descr ident END FOR FOR is firstState TO lastState DO FOR js firstState TO lastState DO AssignString Prob Transition StateToString is istr Concat descr gt StateToString js istr ident p IntToString ORD is 1 istr 0 ConcatCh ident IntToString ORD js 1 istr 0 ConcatCh ident SetDefltP m p is js p is js 0 0 1 0 rtc END FOR END FOR declaration of monitorable variables FOR is firstState TO lastState DO AssignString State freq descr StateT String is istr Concat descr istr ident fs IntToString ORD is 1 istr 0 ConcatCh ident SetDefltMV m fs is 0 0 1 0 notOnFile notInTable isy END FOR FOR is firstState TO lastState DO FOR js firstState TO lastState DO AssignString Freq transition StateToString is istr Concat descr gt StateToString js istr ARRAY OF CHAR descr Concat descr istr Concat ident istr descr Concat descr istr Concat descr istr Concat ident istr Concat ident istr descr ident Concat ident istr descr ident descr C
15. PROCEDURE ModelDefinitions CONST marg 2 BEGIN DeclM m discreteTime Initialize NoInput NoOutput Dynamic NoTerminate ModelObjects Markov chain simulated m NoAbout SetIntegrationStep 1 0 SetMonInterval 1 0 DeclInitSimSession InitSimSess InstallStartConsistency TestConsistency InstallTerminateCondition AllDead DeclExperiment TheExperiment Tilewindows UseCurWSettingsAsDefault SetDefltwindowPlace GraphW marg marg BackgroundWidth 2 marg BackgroundHeight MenuBarHeight 2 marg A 163 ModelWorks V2 0 Appendix SetDefltWindowPlace TableW marg marg BackgroundWidth 2 marg BackgroundHeight 2 marg DIV 3 InstallMenu myMenu Markov enabled InstallCommand myMenu defMarkovCmd Define DefineMarkov enabled unchecked InstallAliasChar myMenu defMarkovCmd W END ModelDefinitions BEGIN RunSimMaster ModelDefinitions END Markov H 5 2 Population dynamics of larch bud moth LBM The following program code contains a sample model demonstrating the daily use of ModelWorks in a research project This program demonstrates modular modeling the use of a parallel model in order to allow the simulationist to compare simulation results with measured data dynamic setting of curve attributes during simulation runs and dynamic activation respec tively deactivation of models during a simulation session The model system is a structured system consisting of two submodels
16. The descriptions given in this reference are brief and relate only to specific properties of individual commands In order to avoid redundancy they do not explain the general principles behind a class of commands and functions of ModelWorks which are described in the part II ModelWorks Theory in particular in the chapter ModelWorks Functions This part contains two chapters The chapter User interface lists all commands which are available to the simulationist via the user interface The chapter Client interface contains the specifications of the client interface used by the modeler All functions and the use of all program objects exported by the ModelWorks modules SimMaster and SimBase are explained Any serious modeling with ModelWorks requires to read at least the Part II ModelWorks Theory and the second chapter on the client interface of the Part III Reference Reading Hint For easier orientation the pages figures and tables of Part III Reference are prefixed with the letter R Within this part figures and tables are numbered separately starting with Fig R1 respectively Tab R1 R71 ModelWorks 2 0 Reference 6 User Interface The user interface of the various ModelWorks versions differ slightly This text has been made using the standard Macintosh version see Appendix As long as just the appearance is affected by the differing implementations holds in particular for the PC version which has a slightly different appeara
17. a state variable s to the simulation environment is removed not the real variable s itself which remains a part of the model definition program Once removed a model or model object is completely unknown to ModelWorks and has become inaccessible by ModelWorks routines E g removed model objects are no longer listed in IO windows and can no longer be integrated PROCEDURE RemoveM VAR m Model PROCEDURE RemoveAllModels PROCEDURE RemoveSV m Model VAR s REAL PROCEDURE RemoveMV m Model VAR mv REAL PROCEDURE RemoveP m Model VAR p REAL PROCEDURE RemoveTabF VAR t TabFUNC Remove procedures may not be called another time unless the model or the model object has been redeclared in the meantime Remove procedures may not be called in the sub state Running s a part II Theory Fig T26 Calling procedure RemoveM results in an implicit removal of all model objects belonging to this model Note that if RemoveTabF removes the last editable table function the menu Table Functions will also be removed 7 4 Simulation Control and Structured Simulation Runs The following Modula 2 objects serve the control of simulations PROCEDURE SimRun This procedure performs an elementary simulation run with the current parameter and other variable settings Typically this routine is used to execute a series of simulation runs e g in a loop within procedure DeclExperiment see below this section Simulation runs can then be executed
18. ac tually encompass several transitions a fact only visible to the modeler T 67 ModelWorks 2 0 Theory citly in the ModelWorks concept Inputs and outputs were formally defined in chapter Theory and the model definition program is responsible for a correct handling ofthem Auxiliary va riables may be used freely and belong completely to the model definition program Note this implies that ModelWorks does neither initialize nor otherwise maintain auxiliary variables Of ten auxiliary variables are computed in the procedure Dynamic hence they will only hold a cor rect value if the procedure Dynamic has at least been called once i e only after a simulation run has already begun s a Fig T16 T17 and T18 attempts to use them in the procedure nitialize may lead to wrong results since their values may not yet be defined 5 2 6 PROGRAM STATES OF THE CLIENT INTERFACE In addition to the program states of the simulation environment there are two sub states of the state Simulating the sub state No run and the sub state Running Fig T26 They can be determined via the client interface by calling the procedure GetMWSubState Their purpose is to distinguish between a state in which no changes may be made to current values and one in which such changes are allowed For instance in order to avoid inconsistencies with an on going simulation but still to allow changes of global simulation parameters they may be changed in the state
19. curStain Stain curStyle LineStyle curSym CHAR curFiling StashFiling curTabul Tabulation curGraphing Graphing A 159 ModelWorks V2 0 Appendix BEGIN FOR is firstState TO lastState DO FOR js firstState TO lastState DO IF is lt gt three THEN GetCurveAttrForMV m f is js curStain curStyle curSym SetCurveAttrForMV m f is js gold spotted OC ELSE GetMV m f is js curMin curMax curFiling curTabul curGraphing SetMV m f is js curMin curMax curFiling curTabul notInGraph END IF END FOR END FOR ClearGraph END InitSimSess PROCEDURE Transition oldS State VAR newS State VAR u REAL BEGIN u i U IF u lt pacc oldS one THEN newS one ELSIF u lt pacc oldS two THEN newS two ELSE newS three END IF END Transition PROCEDURE Dynamic VAR 1 Individuals is js State BEGIN init state frequencies FOR is firstState TO lastState DO fs is 0 0 END compute new state vars amp compute statistics FOR 1 firstIndiv TO lastIndiv DO Transition x 1 xDash 1 INC c x 1 xDash 1 INC n x 1 fs x 1 f fs x 1 1 0 END FOR update pseudo state vars FOR 1 firstIndiv TO lastIndiv DO x l xDash 1 END FOR FOR is firstState TO lastState DO FOR js firstState TO lastState DO IF n is lt gt 0 THEN f is js FLOAT c is js FLOAT n is ELSE f is js 0 0 END IF END FOR END F
20. min pn mean pn max pn descriptor of pn pn unit pn T 69 ModelWorks 2 0 Theory For instance this parameter file might look like 12 0 109 0 234 0 472 Growth rate r day 1 35 6 42 3 49 8 Half saturation c Ks g l 1 0E5 2 5E5 5 0E5 Initial algal dens x0 cells ml A sensitivity analysis can be easily realized in form of the following program code CONST n 3 TYPE PVal cur min mean max PType RECORD v ARRAY cur max OF REAL descr ident unit ARRAY 0 64 OF CHAR END VAR p ARRAY 1 n OF PType PROCEDURE DeclObjects VAR i l n j min max parFile TextFile BEGIN GetExistingFile parFile Open parameter file FOR i 1 TO n DO FOR j min TO max DO GetReal parFile p i v j END 13 7 SkipGap parFile ReadChars parFile p i descr SkipGap parFile ReadChars parFile p i ident SkipGap parFile ReadChars parFile p i unit END Close parFile FOR i 1 TO n DO WITH p i DO Dec1P v cur v mean 0 0 MAX REAL noRtc descr ident unit END END Dec1SV END DeclObjects PROCEDURE MyExperiment VAR i j k min max BEGIN FOR i min TO max DO FOR j min TO max DO FOR k min TO max DO SetP m p 1 v cur p 1 v i SetP m p 2 v cur p 2 v j SetP m p 3 v cur p 3 v k SimRun END END END END MyExperiment or alternatively the procedure MyExperiment may be programmed in the general recursive vari ant which works for any n PROCEDURE MyE
21. quit the compiler by pressing the return key or clicking the mouse In case compiler errors have been detected you must first correct them before you can continue You can ask the MacMETH shell to insert the error messages at the violating locations in your model definition source file by following this method Select the menu command Merge under menu File of the MacMETH shell immediately after the compilation Then use the text editor to correct your errors in the model definition program by searching for comments containing 4 hash marks these comments contain the compiler error messages pointing with a little arrow t to the IA complete listing of the new model is contained in the Appendix 2 On the IBM PC use the command save file under the Editor Menu of the JPI TopSpeed Modula 2 development environment 3 On the IBM PC use the JPI TopSpeed Modula 2 development environment For more information on how to edit compile and correct programs with it consult the Appendix section installation or your JPI TopSpeed Modula 2 documentation Except for the last sentence ignore the whole section to which this footnote belongs it contains Macintosh specific information only However in contrast to the Macintosh version you will have to add an extra step link the compiled module and rename the resulting application to GRASSAPH EXE to GRASSAPH APP 4 Please refer to the Appendix for the installation of the ModelWorks software and the wo
22. recorded In case of the monitorable variables only the information about those monitorable variables is recorded which are involved in the stash filing F or in the graph X or Y The latter requires also that the corresponding recording flag Graph has been set The recording flag Graph controls whether graphical simulation results are written to the stash file Graph recording not available in Reflex and PC version Note that ModelWorks will record information on the stash file at the begin and end of each simulation run in particular also during experiments The data are written to the stash file in the so called RTF Format gt which can be opened by the Microsoft Word WriteNow or MacWrite II document processing software Opening the file with other text editors is also possible but the graph will not be interpreted correctly and control strings which can not be interpreted will also remain in the text However data from simulation results are written in a format which allows to paste or import them directly into many other applications such as the spread sheet program Excel from Microsoft or the presentation graphics program Cricket Graph The format of the stash file has also been designed to allow for an efficient and simple post run analysis In particular check the recording flags for models and monitoring variables if you wish to produce a stash file which can be interpreted successfully by a post analysis Not
23. s a section on Declaration of table functions The latter is done by retuning the value yy i if x gt MAX xx i respectively yy i if x lt MIN xx i Such extrapolations are allowed only if the function procedure Yie from TabFunc is used read Yie as returns dependent value Y by linear inter or extrapolation In case you use Yi returns dependent value Y by interpolation only any attempt to compute a function value y for an independent value x outside the defined range MIN xx i MAX xx i will result in a program halt Coordinates from table define points of support ka ee 3 5 i me from here on variable y 2 5 2 T 7 linear interpolation for y va 1 5 lues missing in between _ Dependent 0 1 2 3 4 5 6 7 8 9 10 11 12 from here on 4 Independent variable x extrapolation Fig R26 Interpolation and extrapolations computed by the function procedures Yie inter and extrapolation and Yi only interpolation from the optional module TabFunc for a non linearity declared as a so called table function The table function is defined by supporting points given in form of coordinates within the domain of definition Inside the domain ModelWorks computes interpolations outside Yie computes extrapolations R 112 ModelWorks 2 0 Reference 7 3 Removing Models and Model Objects Models and model objects can be removed by calling any of the procedures listed below Note that removing means only that the linkage of e g
24. the name the abbreviated name and the unit of the parameter Note that model parameters must not be implemented as constants since they can be changed interactively during a simulation session they must be Modula 2 variables The next procedure ModelDefinitions declares the model It contains the following call to procedure DeclM Dec1M m Euler NoInitialize NoInput NoOutput Dynamic NoTerminate Objects Logistic grass growth model LogGrowth NoAbout This declares the logistic grass growth model within ModelWorks The first actual parameter is the model variable m Then Euler type IntegrationMethod defines the default integration method for this model The next six parameters are all procedures Modula 2 supports procedure types and therefore it is possible to use procedures as actual parameters when calling a procedure This mechanism has to be used to install in ModelWorks all procedures which describe the model dynamics and perform the model object declarations It is then left to ModelWorks to actually call any of these procedures The procedures No Input No Output Dynamic describe the model s dynamics and the procedures No Initialize No Terminate describe actions to be taken at the begin and end of every simulation run more details on the purpose and usage of these procedures is given in the reference manual and in the definition of module SimBase Some of these procedure identifiers have the prefix No which m
25. 0 Reflex of ModelWorks needs at least 512K Bytes of memory RAM the full standard version V2 0 runs on all Macintosh computers with at least 1 MB of memory RAM The Macintosh II version V2 0 II requires in addition to re quirements of the standard version the Motorola 68020 32 Bit CPU and the arithmetic copro cessor Motorola 68881 or later upward compatible chips If applicable all ModelWorks versions have been thoroughly tested to run on the following Macintosh models Mac Reflex 512KE Mac Plus Mac SE Mac II Mac IIx and Mac IIcx It should also run on the models Mac IIci Mac SE 30 Mac Portable and Mac IIfx All Model Works code is 32 Bit clean and ModelWorks is expected to run under system 7 0 and later versions Currently the fully functional Macintosh ModelWorks software kits are not sold but distributed as public domain software They can be obtained from the Swiss Federal Institute of Technolo gy Z rich ETHZ against a minimum handling charge Besides the user provided hardware the user obtains with the distributed software everything needed to work fully with ModelWorks i e included is even a system software a full Modula 2 development system editor compiler plus utilities such as symbolic debugger application linker the Dialog Machine and Model Works For a detailed description of the ModelWorks software kit see below the section Installation of ModelWorks B 2 IBM PC VERSION The IBM PC ModelWorks versio
26. 2 The monitorable variable grass needs a new upper limit for its clipping range DeclMV grass 0 0 10000 0 Grass G g dry weight m 2 notOnFile writeInTable isY The model parameters c and c2 need new default values DeclP cl 0 4 0 0 10 0 rtc cl growth rate of grass vaL day 1 DeclP c2 8 0E 5 0 0 1 0 rtc c2 self inhibition coefficient of grass c2 m 2 g dw day Secondly insert the procedures declaring the variable aphids as a state and as a moni torable variable plus call the procedures declaring the new parameters DeclSV aphids aphidsDot 20 0 0 0 1000 0 Aphids A g dry weight m 2 DeclMV aphids 0 0 1500 0 Aphids A g dry weight m 2 notOnFile writeInTable isY DeclP c3 1 5E 3 0 0 1 0 rtc c3 coupling parameter c3 m 2 g dw day DeclP c4 0 1 0 0 10 0 rtc c4 ratio of grass net use by aphids c4 DeclP c5 0 2 0 0 10 0 rtc c5 death rate of aphids c5 day 1 You could call these procedures in any order mix declarations of state variables with those of monitorable variables or parameter declarations However consider that the T 30 ModelWorks 2 0 Tutorial sequence of declarations corresponds to the order in which they are listed in the I O windows of ModelWorks simulation environment The model declaration procedure ModelDefinitions remains the same except for minor
27. 2 MB of disk space and significantly more than 500K of free main memory for ModelWorks a minimum of about 350K for the Dialog Machine alone If you do not have enough room clean up the disk and remove unneeded drivers or TSRs Before you start with the installation of GEM please have the following technical specifications ready e the type of monitor you are using e the type of mouse you are using and to which serial port it is connected if it is not a bus mouse e the type of printer you are using and to which port it is connected l Order ModelWorks PC V1 1 PC from the following address Projekt Zentrum IDA re ModelWorks PC Swiss Federal Institute of Technology ETHZ ETH Zentrum CH 8092 Ziirich Switzerland Phone 01 256 5440 A 136 ModelWorks V2 0 Appendix D 1 2 Installation of GEM Desktop First begin with the installation of GEM Start your machine and when it is ready put the GEM Master Disk into drive A and type A and press the lt ENTER gt key Now type GEMSETUP and press lt ENTER gt again The program GEMSETUP will ask you a few questions about your system and will then copy some files to your hard disk Just follow the instructions given in the dialog IMPORTANT The Dialog Machine and ModelWorks in their DOS PC version do not use colors Even if you do have a color monitor you should configure GEM as a monochrome sys tem When you make the selection of monitors in the GEMSETUP dialog you can select
28. 475457 24 7217 12 2294 noRtc c10 slope of fecundity vs rf c10 1lbm DeclP cll 356 72636 264 9847 448 4680 noRtc cll y intercept of fecundity vs rf cll lbm DeclP c12 11 99 11 79 12 19 noRtc c12 minimum rf c12 Dec1P c13 0 425 0 4 0 5 noRtc c13 minimum decrement of rf c13 Dec1l1P c14 18 0 17 5 18 5 noRtc c14 maximum rf c14 DeclP c15 0 4 0 35 0 6 noRtc c15 defoliation threshold c15 DeclP c16 0 8 0 7 1 0 noRtc cl6 defoliation threshold of maximum stress c16 Dec1P c17 91 3 91 3 91 3 noRtc cl7 branches per tree c17 kg END ModelObjects PROCEDURE ActivateLarchLBMModel BEGIN IF NOT curActive THEN DeclM m discreteTime Initialize NoInput Output Dynamic NoTerminate ModelObjects Larch Bud Moth model b1 V3 0 Larch Larch bud moth relationship LWMod3 b1 NoAbout SetSimTime FLOAT kmin FLOAT kmax curActive TRUE END IF END ActivateLarchLBMModel PROCEDURE DeactivateLarchLBMModel BEGIN IF curActive THEN RemoveM m curActive FALSE END IF END DeactivateLarchLBMModel PROCEDURE LarchLBMModelIsActive BOOLEAN BEGIN RETURN curActive END LarchLBMModellsActive BEGIN curActive FALSE END LBMMod The module LBMObs provides a parallel submodel of the measured larval densities of the larch bud moth observations made in the field while studying the lar
29. ARRAY OF CHAR R111 ModelWorks 2 0 Reference xMin xMax yMin yMax REAL It is recommended to avoid the direct modification of state variables parameters etc by assigning them a new value There are two reasons why First there may result a confusing discrepancy in the value actually used for simulations and the one visible in IO windows Secondly ModelWorks is likely to overwrite the value so that the assignment is fictitious and the simulationist may have difficulties to understand subsequent simulation results To avoid any such problems use always the set procedures and they will preserve consistency between the model definition program and ModelWorks In the case of table functions note that the dimensions of the vectors xx and yy may not be changed only the coordinate values but not the number of supporting points may be modified Otherwise remove and redeclare the whole table function Note that if SetTabF disables for the last table function its property of beeing editable modifiable FALSE the menu Table Functions will be removed from the menu bar 7 2 2 c Inter and extrapolation with table functions Table functions as provided by the optional module TabFunc allow to compute function values within the defined domain MIN xx i MAX xx i by linear interpolation or outside this range by extrapolation Fig R26 xx i are the elements of the vector xx containing the indepenent values of the supporting points
30. BOOLEAN tabName xName yName xUnit yUnit ARRAY OF CHAR xMin xMax yMin yMax REAL PROCEDURE SetTabF t TabFUNC XX YY ARRAY OF REAL NValPairs INTEGER modifiable BOOLEAN tabName xName yName xUnit yUnit ARRAY OF CHAR xMin xMax yMin yMax REAL PROCEDURE GetTabF t TabFUNC VAR xx yy ARRAY OF REAL VAR NValPairs INTEGER VAR modifiable BOOLEAN VAR tabName xName yName xUnit yUnit ARRAY OF CHAR VAR xMin xMax yMin yMax REAL PROCEDURE RemoveTabF VAR t TabFUNC PROCEDURE Yi t TabFUNC x REAL REAL interpolate only ELSE HALT PROCEDURE Yie t TabFUNC x REAL REAL inter and extrapolate END u ModelWorks may be freely copied but not for profit A 182 Index application see stand alone application auxiliary variable 12 54 104 calculation order of procedures 42 43 50 52 101 client see client interface 6 10 46 47 65 98 auxiliary library 65 98 181 mandatory part 65 98 179 optional part 65 98 181 coincidence interval 39 40 43 79 coincidence point 40 modeler consistency check see start consistency check continuous time 11 36 38 39 40 108 155 coupling of models 38 40 41 155 164 current value 12 18 60 83 106 curve attribute 21 96 118 119 automatic definition 21 63 96 119 in legend 120 default value 6 11 12 47 60 81 106 declaration of experim
31. Dc r u x A 156 NoAbout Fe He Fe He Fe Fe 2 2 2 2 2 2 2 5 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 9 Fe e Fe He Fe Fe He Fe Fe e He Fe He 8 2 2 2 5 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 20 Dec lM IntegrationMethod DeclSV StashFiling Tabulation Graphing DeclMV DeclP RTCType Model SetSimTime SetMonInterval NoInitialize NoOutput NoTerminate NoAbout ModelWorks V2 0 Appendix PROCEDURE ObjectsCont BEGIN Dec1SV x xDot 1 0 1 0E3 1 0E3 State variable of continuous time submodel x Dec1MV x 0 0 5 0 State variable of continuous time submodel x notOnFile writeInTable isY DeclP r 1 0 100 0 100 0 rtc Intrinsic rate of change for continous time submodel r time 1 END ObjectsCont PROCEDURE DeclSubModCont BEGIN DeclM contM Euler NoInitialize Ic NoOutput Dc NoTerminate ObjectsCont Continuous time submodel ContSubMod NoAbout END DeclSubModCont END SubModCont FER Fe Fe He Fe Fe e Fe He deke Fe e Fe He 2 2 2 2 2 2 2 212 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 27 PROCEDURE StructuredModelDef BEGIN DeclSubModCont DeclSubModDisc SetSimTime 0 0 10 0 SetMonInterval 0 25 END StructuredModelDef BEGIN RunSimMaster StructuredModelDef END Combined H 5 RESEARCH SAMPLE MODELS H 5 1 Third order finite Markov chain Markov MOD The following model definition program Markov MOD demonstra
32. ETH Z rich Institut f r Terrestrische Okologie Grabenstrasse 3 CH 8952 Schlieren Z rich SWITZERLAND e mail sysecol ito umnw ethz ch Dr M Ulrich Institut f r Gew sserschutz und Wassertechnologie ETH Z rich EAWAG CH 8600 D bendorf SWITZERLAND first print1990 System kologie ETH Z rich reprinted 2011 ModelWorks An Interactive Simulation Environment for Personal Computers and Workstations by Andreas Fischlin amp Olivier Roth Dimitrios Gyalistras Markus Ulrich and Thomas Nemecek ModelWorks Version 2 0 Z rich May 1990 Abstract ModelWorks is a modeling and simulation environment in Modula 2 specifically designed to be run interactively on modern personal computers and workstations It supports modular modeling by featuring a coupling mechanism between submodels and unrestricted number of state variables model parameters etc up to the limits of the computer resources It allows for the formulation of continuous time discrete time as well as continuous and discrete time mixed models Finally ModelWorks offers in its interactive simulation environment a handy user interface featuring efficient alterations of model and simulation run parameters Systems Ecology Group Institute of Terrestrial Ecology Department of Environmental Sciences Swiss Federal Institute of Technology ETH Zentrum CH 8092 Z rich Switzerland EAWAG Swiss Federal Institute of Water Resour
33. EditItem stringToFind ARRAY OF CHAR VAR firstCh lastCh INTEGER BOOLEAN PROCEDURE ScrollText textField EditItem dcols dlines INTEGER PROCEDURE ScrollTextWithWindowScrollBars textField EditItem PROCEDURE AddScrollBarsToText textField EditItem withVerticalScrollBar withHorizontalScrollBar BOOLEAN L FEE Hee NE He RE EI He BE IE Te He IE He EE ee DMWPictIOo ERITREA TIEFE RR SIE HHI EET PROCEDURE StartPictureSave PROCEDURE StopPictureSave PROCEDURE PausePictureSave PROCEDURE ResumePictureSave PROCEDURE DisplayPicture ownerWindow Window destRect RectArea PROCEDURE DiscardPicture PROCEDURE SetPictureArea r RectArea PROCEDURE GetPictureArea VAR r RectArea PROCEDURE SetHairLineWidth f REAL PROCEDURE GetHairLineWidth VAR f REAL VERRECHNET FETTE ET TEEN DMWTextIO SO eRe IE He I ee He He EN HE ee NR eI A I ATK He RA Hy PROCEDURE StartTextSave PROCEDURE StopTextSave PROCEDURE PauseTextSave PROCEDURE ResumeTextSave PROCEDURE DisplayText ownerWindow Window destRect RectArea fromLine LONGINT PROCEDURE DiscardText PROCEDURE GrabWText VAR txtbeg ADDRESS VAR curTextLength LONGINT PROCEDURE ReleaseWText PROCEDURE AppendWText txtbeg ADDRESS length LONGINT PROCEDURE SetWTextSize newTextLength LONGINT END Different from Version ModelWorks V2 0 Appendix I 2 MODELWORKS CLIENT INTERFACE AND OPTIONAL MODULES he following listing of the client interface is identical for all ModelWorks versions
34. Fig T2 Max initial value State variable x t _ gt ___ Helga Min initial value x t orx k 1 Max value h Model parameter c Min value Max value of interest Min value of interest Monitorable variable mv Z Clipping range N Fig T2 Model objects 0 of a ModelWorks model A ModelWorks model definition must consist of at least one model and every model usually contains state variables model parameters and monitorable variables Any initial value parameter value minimum or maximum value becomes mandatory if the associated variable or parameter is declared within the model definition ModelWorks maintains the actual values of state variables parameters and monitorable variables and even remembers their initially specified values default values ModelWorks auto matically assigns the initial value to the state variable x at the begin of every simulation run and the value p is assigned to the model parameter c at the beginning of the simulation session or after any interactive change t ModelWorks uses the derivative or new value in order to compute and repeatedly assign newly obtained values to the state variable during the course of a simulation run numerical integration During simulation experiments the unknown values which the monitorable variable mv may obtain shall be drawn in graphs only if they fit within a particular range of interest otherwise ModelWorks will clip them from the di
35. File Execute X and select OBM file to be executed normally model definition program To debug either execute a HALT statement needs recompilation or press interrupt button not reset button of the programmers button at the side of your Macintosh see owner s guide You may look at current values and location of error but you can t change any variable s values Quit debugger by choosing menu command File Quit and resume program Clicking into the source code while having the com mand key pressed allows to debug dynamically program execution will continue till the clicked statement is actually executed and the debugger is automatically reentered don t forget to actually leave the debugger with File Quit when you debug dynamically Please refer to the separate MacMETH manual for general and more detailed information on how to work with the MacMETH software In the following you find some often useful hints and specific information on how to work with MacMETH when using ModelWorks Note that the RMSMacMETH shell reads a configuration file type TEXT with the name User Profile when it is started This file which must have exactly this name and which must reside in the same folder as the shell contains the path names of the folders and disks where the modules and subprograms can be found Make sure that the names of your disks and folders where you keep the tools such as the compiler and the library modules match exactly the names given in
36. Grass is drawn in red broken and Grass derivative in black unbroken This is only because Grass has been activated for monitoring as the second curve after Grass derivative which has remained untouched during the toggling of Grass The convenient the automatic curve attributes definition strategy may be the confusing it may become in complicated situations where the simulationist wishes to run may simulations and to compare the same monitoring variables ModelWorks allows you to gain complete or partial control over the assignment of curve attributes i e you can adopt your own curve attributes assignment strategy You may achieve this by assigning explicitly to monitoring variables their particular curve attributes For instance change the color of the curve Grass to green and draw it with the symbol v which may remind you of real grass tuft _Click on Grass in the window for monitorable variables and click on the button rainbow toggle function Choose the attributes unbroken as line style the stain emerald green and type v in the symbol field then click the OK push button Finally select the command Set Global simulation parameters under the menu Settings and change the monitoring interval to 0 5 then click the OK push button Now run another simulation run and you should see this time a green curve displaying the symbol v at times 0 0 5 1 1 5 2 etc Note that from now on the curve Grass will always be drawn with e
37. In case you have to use them always remember to update the User Profile each time you move or copy the MacMETH system Note also that it is futile to change the User Profile while running the shell unless you execute the utility SetProfile after having changed the User Profile A 130 ModelWorks V2 0 Appendix To work with the MacMETH Modula 2 development shell start it with a double click on the icon of the file named RMSMacMETHI exactly as you do with any other Macintosh application program The file menu of MacMETH offers at least the commands Edit Compile Execute and Ouit the actual content of the commands in this menu depend on the User Profile Edit will only produce a message that you should use the utility Merge to detect compilation errors Execute displays a dialog box in which the program to be started can be chosen e g a ModelWorks model Ouit terminates the MacMETH shell In case that compiler errors were detected use the utility Merge after the compilation Merge inserts the corresponding error messages at the violating locations in your source file Then use MockWrite or any other text editor to correct your errors in the model definition program Search for comments containing 4 hash marks they contain the inserted compiler error messages and point with a little arrow t to the position within the source code line where the error has been detected Note that Merge can insert the error messages only into your
38. MRA IR E EEE KK SK RT HR FERNER TYPE MouseModifiers regular cmded opted shifted ClickKind SET OF MouseModifiers DragProc PROCEDURE INTEGER INTEGER VAR WindowIODone BOOLEAN PROCEDURE PointClicked x y INTEGER maxDist INTEGER BOOLEAN PROCEDURE RectClicked rect RectArea BOOLEAN PROCEDURE PointDoubleClicked x y INTEGER maxDist INTEGER BOOLEAN PROCEDURE RectDoubleClicked rect RectArea BOOLEAN PROCEDURE GetLastClick VAR x y INTEGER VAR click ClickKind BOOLEAN PROCEDURE GetLastDoubleClick VAR x y INTEGER VAR click ClickKind BOOLEAN PROCEDURE GetCurMousePos VAR x y INTEGER PROCEDURE GetLastMouseClick VAR x y INTEGER VAR click ClickKind A 174 ModelWorks V2 0 Appendix PROCEDURE DoTillMButReleased p PROC PROCEDURE Drag duringDragP afterDragP DragProc PROCEDURE SetContSize u Window contentRect RectArea PROCEDURE GetContSize u Window VAR contentRect RectArea PROCEDURE SetScrollStep u Window xStep yStep INTEGER PROCEDURE GetScrollStep u Window VAR xStep yStep INTEGE PROCEDURE GetScrollBoxPos u Window VAR posX posY INTEGER PROCEDURE SetScrollBoxPos u Window posX posY INTEGER PROCEDURE GetScrollBoxChange u Window VAR changeX changeY INTEGER PROCEDURE AutoScrollProc u Window RestoreProc PROCEDURE GetScrollProc u Window VAR scrollP RestoreProc PROCEDURE SetScrollProc u Window scrollP PROCEDURE ScrollContent u Window dx dy INTEGER PROCEDURE M
39. Macintosh computer the instructions are to be executed similarly but may look a bit differently or behave s lightly differently since the IBM PC yersion of the Dialog Machine is only a subset of the Macintosh version A few hints On the IBM PC folders become directories object files ending with the extension OBM become linked GEM applications with the extension APP and in contrast to the Macintosh MS DOS file names are truncated to 8 characters extension excluded note that the latter may also affect module names For more details see the appendix Wherever necessary IBM PC specific information has been added in form of footnotes Please interpret the text accordingly and accept our apology for not being able to offer an IBM PC text version note that we are a research institution not a commercial software company and hence not able to maintain more than that version we use ourselves in our daily research work however you should have no eee in following the tutorial text since all essential features of ModelWorks are available on the IBM PC version as well 2 We recommend Operation of the computer Your owner s guide e g Macintosh owner s guide Modula 2 WIRTH N 1988 Programming with Modula 2 Springer Verlag Heidelberg New York 4th corrected ed Modelling LUENBERGER D G 1979 Introduction to dynamic systems Theory models and applications Wiley New York 446pp T9 ModelWorks 2 0 Tutorial 1 G
40. Modula 2 implementation Purpose Computation of normally distributed variates References Bell J R 1968 Normal random deviates Algorithm 334 Colected Algorithms from CACM Communications of the Association for Computing Machinery 334 P 1 Rl Box G amp Muller M 1958 A note on the generation of normal deviates Ann Math Stat 28 610 Von Neumann J 1959 Various techniques used in connection with random digits In Nat Bur Standards Appl Math Ser 12 US GTovt Printing Off Washington D C p 36 Remark This implementation allows to be completely independent from any particular random number generator see InstallU NOTE The module won t crash if InstallU is never called but it will not be able to produce correct results Imported modules System MathLib Programming Design A Fischlin 17 Dec 87 Implementation A Fischlin 17 Dec 87 Swiss Federal Institute of Technology Zurich Project Centre IDA Pilot Project CELTIA Computer aided Explorative Learning and Teaching with Interactive Animated Simulation ETH Zentrum CH 8092 Zurich Switzerland Last revision 24 Mar 88 A F Fe ke He KEKE ZZ 2 2 2 2 2 212 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 212 2 5 2 2 5 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 TYPE URandGen PROCEDURE REAL PROCEDURE InstallU U URandGen Installs procedure U which returns variates from a random variable uniformally distributed within interval 0
41. No run but not in the state Running The possibility to change current values is essential for structured simulations since the modeler may wish to change current values between two consecutive simulation runs e g initial values parameters during a parameter identification the integration step or the simulation start and stop time see section on Programming structured simulations Experiments Some state transitions are under the control of the simulationist only e g Start run or Execute experiment Resume run or Stop kill run from state Pause all other transitions are controlled by both the simulationist and the modeler e g SimRun or Start run PauseRun or Halt run Pause etc If the modeler calls procedure SimRun the state No run is left and ModelWorks enters the state Running If a simulation run is finished because the simulation time has reached the stop time the termination condition has become true or the simulationist has stopped the simulation the state Running is left and the state No run is entered Fig T26 Note that if the simulationist stops Kills a structured simulation ModelWorks will execute all remaining statements eventually calling procedure SimRun many times till the procedure Experiment is actually finished However note that in the latter case no integration and monitoring will take place since ModelWorks empties the body of the procedure SimRun so that the structured simulation will terminate withou
42. Nov 11 Dec 12 Sun 1 Mon 2 Tue 3 Wed 4 Thur 5 Fri 6 Sat 7 TYPE Months INTEGER WeekDays INTEGER DateAndTimeRec RECORD year INTEGER 1904 1905 2040 month Months day CF Les dl hour 0 23 minute 0 59 second INTEGER 0 59 dayOfWeek WeekDays A 148 ModelWorks V2 0 Appendix END PROCEDURE SetDateAndTime d DateAndTimeRec PROCEDURE GetDateAndTime VAR d DateAndTimeRec PROCEDURE DateToSeconds date DateAndTimeRec VAR s LONGINT PROCEDURE SecondsToDate s LONGINT VAR d DateAndTimeRec s is the number of seconds passed since 1st Jan 1904 All date operations are only valid within lst January 1904 till 31st December 2040 END DateAndTime G 2 4 WriteDatTim DEFINITION MODULE WriteDatTim BER Fe Fe He e He Fe 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 Module WriteDatTim Version 1 0 Copyright 1988 by Andreas Fischlin and CELTIA Swiss Federal Institute of Technology Z rich ETHZ Version for MacMETH V2 4 1 Pass Modula 2 implementation Purpose Writing of date and time Uses DateAndTime Programming Design Implementation A Fischlin 16 Mai 88 Swiss Federal Institute of Technology Zurich Project Centre IDA Pilot Project CELTIA Computer aided Explorative Learning and Teaching with Interactive Animated Simulation ETH Zentrum C
43. P c p will imply that the default value p for the model parameter is assigned to the variable c 3 2 Developing a New Model Instead of just reading an existing model definition program we will develop a new model However the new model will not be written completely anew that would be too cumbersome Instead we will simply modify a copy of the sample model definition program to develop the new model This is an easy hence generally recommended way to develop ModelWorks model definition programs 3 2 1 THE NEW MODEL The new model does not only include grass but also herbivores as the second state variable aphids Aphids feed on the grass and establish an ecological relationship for the sake of simplicity we assume somehow similar to other predator prey relationships The new model will consist of two coupled differential equations each describing the dynamics of the two species according to the Lotka Volterra predator prey model the grass is the prey and the aphids are the predators The model is described with the following nonlinear second order differential equation system note that the parameter and initial values are not the same as in the former model dG t dt c Cl co Gt c3 Git Al 2 dC t dt c3cq4 G t A t c5 A t where State variables Grass g dry weight dw per m2 G t Initial amount of grass initial value G 0 200 g m Aphids g dry weight dw per m A t Initial number of aphids A
44. RAINIER HERE RAIN ET PROCEDURE WriteMessage line col CARDINAL msg ARRAY OF CHAR PROCEDURE ShowAlert height width CARDINAL WriteMessages PROC PROCEDURE ShowPredefinedAlert fileName ARRAY OF CHAR alertID INTEGER strl str2 str3 str4 ARRAY OF CHAR RR Fe IE RIE IE RIE IER SOK HE IIE E RR I DMClipboard ERTEILEN HIE IE RIN LRA RR IIE I SN Ae ON BUR HEE EY TYPE EditCommands undo cut copy paste clear VAR ClipboardDone BOOLEAN PROCEDURE InstallEditMenu UndoProc CutProc CopyProc PasteProc ClearProc PROC PROCEDURE EnableEditMenu PROCEDURE DisableEditMenu PROCEDURE EnableEditCommand whichone EditCommands PROCEDURE DisableEditCommand whichone EditCommands PROCEDURE PutPictureIntoClipboard PROCEDURE GetPictureFromClipboard simultaneousDisplay BOOLEAN destRect RectArea PROCEDURE PutTextIntoClipboard PROCEDURE GetTextFromClipboard simultaneousDisplay BOOLEAN destRect RectArea fromLine LONGINT RRR IRR IC IIR ICRI ICI ICICI ICICI ICICI ICICI CICK IK ke DMEditFields FUORI E e e e ICICI EZ e e e ES II k I e e He k e ICA RCI A IK CONST nonexistent NIL TYPE EditItem RadioBut EditHandler PROCEDURE EditItem ItemType charField stringField textField cardField intField realField pushButton radioButtonSet checkBox scrollBar Direction horizontal vertical VAR EditFieldsDone BOOLEAN PROCEDURE CharField u Window VAR ei EditItem x y INTEGER ch CHAR char
45. Systems Ecology Group re ModelWorks Bug ETH Zentrum CH 8092 Z rich Switzerland Phone 01 256 58 93 G Definition Modules G 1 OPTIONAL CLIENT INTERFACE G 1 1 TabFunc The listing of this definition module has been omitted since its exported objects are already fully described in the part Reference section Client interface G 1 2 SimIntegrate DEFINITION MODULE SimIntegrate 2 H H Fe Fe He e Fe Fe 2 5 2 2 2 2 2 2 5 2 2 2 2 2 2 2 2 2 2 2 2 AAA Module SimIntegrate MW_V2 0 Copyright 1989 by Andreas Fischlin and Swiss Federal Institute of Technology Ziirich ETHZ Version written for Dialog Machine V2 0 User interface MacMETH V2 6 1 Pass Modula 2 implementation Purpose Provides means to integrate an autonomous differential equation system without any monitoring Remarks This module is part of ModelWorks MW_V2 0 an interactive Modula 2 modeling and simulation environment Programming Design A Fischlin 26 06 89 Implementation A Fischlin 26 06 89 Swiss Federal Institute of Technology Zurich ETHZ Department of Environmental Sciences Systems Ecology Group ETH Zentrum CH 8092 Zurich Switzerland Last revision of definition 26 06 89 af Fe ke he Fe He Fe EER 2 Fe He Fe Fe He He Fe Fe He 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 87 FROM SimBase IMPORT Model PROCEDURE Integrate m Model from till REAL Integrate the autonomous system o
46. THEN IF modCB THEN ActivateLarchLBMModel ELSE DeactivateLarchLBMModel END IF obsCB THEN ActivateLBMObsModel ELSE DeactivateLBMObsModel END END IF END Choose PROCEDURE InstallMenus BEGIN InstallMenu modM Models enabled InstallCommand modM modActCmd Activation Choose enabled unchecked InstallAliasChar modM modActCmd L END InstallMenus BEGIN RunSimMaster InstallMenus END LBM A 172 ModelWorks V2 0 Appendix I Quick References I 1 DIALOG MACHINE For details on how to work with the Dialog Machine see this appendix the section above How to Work With the Dialog Machine The following Dialog Machine version 2 0 is currently only available on the Macintosh For the IBM PC the newest version is DM PC V1 6 and its differ ences to DM 2 0 are explained in a separate documentation Dialog Machine Version 2 0 January 1990 1990 Andreas Fischlin CELTIA and Swiss Federal Institute of Technology Zurich KERNEL RRR IRC e E ICI ICI SE ES ICI SE IOI TCA RA ik DMConversions HR KK KR KR KKK KKK KK KEK KK RR KEKE KER KEK RRR KEK TYPE RealFormat FixedFormat ScientificNotation PROCEDURE StringToCard str ARRAY OF CHAR VAR card CARDINAL VAR done BOOLEAN PROCEDURE CardToString card CARDINAL VAR str ARRAY OF CHAR length CARDINAL PROCEDURE StringToLongCard str ARRAY OF CHAR VAR lcard LONGCARD VAR done BOOLEAN PROCEDURE LongCardToString lcard LONGCARD VAR str ARRAY OF C
47. This function does not work on a multiple selection Y Set delete curve Y isY Adds the selected monitorable variable s to the list of variables which are to be drawn as curves in the graph window The function toggles actually the setting 1 e if the current setting is on Y it is disabled otherwise enabled Fig R22 In the column Monitoring of the IO window the symbols F T or Y are shown in the same color as that which is used to draw values of the corresponding variable in the graph If a multiple selection is active the function adds or removes all currently selected variables to or from the list reversing the current setting of the first variable in the selection This function will completely erase and redraw the content of the graph window if the mode RedrawGraphAlwaysMode of the simulation environment is currently active see above File Preferences E Reset graphing Resets the graphing X Y isX isY notInGraph of the selected monitorable variable s to their defaults Y Set scaling Opens an entry form in which minimum and maximum values for the scaling of the selected monitorable variable in the graph can be edited Fig R24 Scaling of the following monitorable variable Grass Min Max 1000 0 Fig R24 Entry form opened by the L button in the IO window Monitorable variables to set the scaling of a particular monitorable variable Only values within these limits will be displayed in the graph If
48. a minimum number of rules so that ModelWorks and the modeler use and access model objects in harmony The client interface has been designed to warrant this harmony automati cally to a large extent as well as to offer the modeler at the same time as much flexibility as possible The modeler may also program a structured simulation a so called ModelWorks experiment Typically an experiment consists of many simulation runs and will call ModelWorks functions similar to the way the simulationist would use them The latter is important if the simulationist is to be relieved from cumbersome repetitive command sequences or if simulations are used as parts of more complicated procedures For instance in order to create a phase portrait the simu lationist would have to assign a series of different initial values to the state variables as well as to start after each assignment a simulation run All this can be easily accomplished by pro gramming a structured simulation which the simulationist then can activate by a single command from within the simulation environment Other examples are After a model has been thoroughly explored interactively it is to be used for a sensitivity analysis or parameter identifi cations both examples requiring the writing of special program sections a task which is well supported by ModelWorks client interface Moreover ModelWorks allows to extend its functions via the client interface Since ModelWorks is based on the Dialo
49. amp HARLEY P J 1983 An Introduction to numerical methods with Pascal London Addison Wesley 300pp BALTENSWEILER W amp FISCHLIN A 1988 The larch bud moth in the Alps In Berryman A A ed Dynamics of forest insect populations patterns causes implications New York a o Plenum Publishing Corporation 331 351 CELLIER F E amp FISCHLIN A 1980 Computer assisted modelling of ill defined systems In Trappl R Klir G J amp Pichler F R eds General Systems Methodology Mathematical Systems Theory Fuzzy Sets Proc of the Fifth European Meeting on Cybernetics and Systems Research Vol VIII 417 429 McGraw Hill Intern Book Comp Washington New York 1982 544pp Copy W J 1981 Analysis of proposals for the floating point standard IEEE Computer 14 3 63 68 ENGELN MULLGES G amp REUTTER F 1988 Formelsammlung zur Numerischen Mathematik mit MODULA 2 Programmen Wissenschaftsverlag Mannheim a o 510pp FISCHLIN A 1982 Analyse eines Wald Insekten Systems Der subalpine L rchen Arvenwald und der graue L rchenwickler Zeiraphera diniana Gn Lep Tortricidae Diss Eidg Tech Hochsch Z rich No 6977 294pp FISCHLIN A 1986 Simplifying the usage and programming of modern workstations with Modula 2 The Dialog Machine Internal report Project Centre IDA Swiss Federal Institute of Technology Z rich ETHZ Switzerland In prep FISCHLIN A 1986 The Dialog Machine for the Macin
50. bold italic underline LaserFont Times Helvetica Courier Symbol FontStyle SET OF FontStyles PROCEDURE AssignWindowFont f WindowFont size CARDINAL PROCEDURE SetWindowFont wf WindowFont size CARDINAL style FontStyle PROCEDURE GetWindowFont VAR wf WindowFont VAR size CARDINAL VAR style FontStyle PROCEDURE SetLaserFont lf LaserFont size CARDINAL style FontStyle PROCEDURE GetLaserFont VAR lf LaserFont VAR size CARDINAL VAR style FontStyle PROCEDURE SetPos line col CARDINAL PROCEDURE GetPos VAR line col CARDINAL PROCEDURE ShowCaret on BOOLEAN PROCEDURE Invert on BOOLEAN PROCEDURE Write ch CHAR PROCEDURE WriteString s ARRAY OF CHAR PROCEDURE WriteL PROCEDURE WriteCard c n CARDINAL PROCEDURE WriteLongCard lc LONGCARD n CARDINAL PROCEDURE WriteInt c INTEGER n CARDINAL PROCEDURE WriteLongInt li LONGINT n CARDINAL PROCEDURE WriteReal r REAL n dec CARDINAL PROCEDURE WriteRealSci r REAL n dec CARDINAL PROCEDURE WriteLongReal lr LONGREAL n dec CARDINAL PROCEDURE WriteLongRealSci lr LONGREAL n dec CARDINAL PROCEDURE SetPen x y INTEGER PROCEDURE GetPen VAR x y INTEGER PROCEDURE SetBrushSize width height INTEGER PROCEDURE GetBrushSize VAR width height INTEGER PROCEDURE Dot x y INTEGER PROCEDURE LineTo x y INTEGER PROCEDURE Circle x y INTEGER radius CARDINAL filled BOOLEAN fillpat Pattern PROCEDURE Area r RectArea pat Pattern P
51. briefly mentioned wherever appropriate This text is subdivided into three parts Part I is a Tutorial containing a little tour to be followed step by step It suffices to learn all basic techniques which are needed in or der to model and simulate simple models with ModelWorks Part II explains the Theory and concepts behind ModelWorks in particular model formalisms and all func tions of ModelWorks Any advanced modeling such as modular modeling requires to study the theoretical part Part III is a Reference manual containing a complete list and description of all features of ModelWorks Finally the Appendix contains detailed instructions for the installation model development cycle and other technical details of interest during the daily work with ModelWorks Included are also listings of Model Works definition modules several listings of sample models convenient quick reference listings and an index Reading Hint Throughout this text italics are used to emphasize that the text is to be taken literally in particular also case sensitive E g in the citation of an identifier such as a module name like SimMaster or if the user has to open a file or directory with a given name such as Logistic OBM or MWSAMPLES For easier orientation the pages figures and tables in Part I Tutorial and II Theory are prefixed with the letter T in part III Reference with the letter R and in the Appendix with the letter A Within some parts figures and ta
52. click on the buttons dl ei and N Note that the effect of this method is exactly the same as if you would have clicked on the buttons with the same pictures in the window Monitorable variables after having selected the model title bold face Logistic grass growth model in the latter window 2 2 6 CHANGING PARAMETERS DURING SIMULATION Now you will learn how you can change model parameters even in the middle of a simulation run We let the model simulate the grass growth as before but when the density has reached its maximal value we increase the self inhibition of the plants the parameter cy this signifies that the carrying capacity of the environment K cj c decreases for instance due to a sudden nutrient depletion or an unknown toxic substance After this change the grass density will tend to a lower equilibrium value To do this start a simulation with the same settings as before When the population has reached its maximal value this happens approximately at time 15 0 watch the time window interrupt the simulation with the menu command Simulation Halt run Pause In the parameter window you can now increase the value of the self inhibition coefficient c2 from 0 001 to 0 002 continue the simulation with Simulation Resume run and observe the reaction of the system Note that on a monochrome screen or a non color printer such as a laser printer both curves are drawn with the same line style i e unbroken and can onl
53. current values of the currently selected model parameters SET Columns set up Activates an entry form in which the display of the columns in the IO window Model parameters can be controlled Fig R21 The columns of which the display can be turned on or off are Parameter names Full names of the model parameters Ident Short identifiers of the model parameters Unit Unit in which to measure the values of the model parameters Value Current value of the model parameters RTC rtc noRtc runtime change no runtime change The value in this column shows whether a parameter may be changed during a simulation run or not In order to warrant consistency the modeler may prevent the changing of a parameter value in the middle of a simulation by disabling this flag noRtc By default this column is not shown Check the columns to be displayed in the window EJ Parameter names EJ Ident Ex Unit EJ Value O RTC Fig R21 Entry form opened by the button in the IO window Model parameters Selects all model parameters All subsequent button functions will operate on the scope All Fig T21 part II Theory i e on all model parameters of all models R 92 ModelWorks 2 0 Reference Y Set model parameter value Opens an entry form in which the current parameter value for the selected model parameter s can be edited ModelWorks rejects any attempt to enter a value out of the range as defined by the mod
54. depends on the integration method used and on the nature of the model R77 ModelWorks 2 0 Reference Start time for simulation Stop time for simulation Integr tion step h Monitoring interval 0 25 Fig R5a Entry form Global simulation parameters 1 for case A only continuous time sub models are present Start time for simulation 1949 Stop time for simulation 1987 Monitoring interval Discrete time step fi o Fig R5b Entry form Global simulation parameters I for case B only discrete time sub models are present Start time for simulation Stop time for simulation Maximum relative local error Discrete time step Maximum integration step h Monitoring interval 0 25 Fig R5c Entry form Global simulation parameters I for case C some continuous as well discrete time sub models are present In addition a variable step length integration method is used R 78 ModelWorks 2 0 Reference differs from case C where not only discrete time but also continuous time sub models are present In the latter case c may be edited in addition to h and becomes the Coincidence interval c The variable c is always an integer Monitoring interval h Interval at which the values of all monitorable variables are either written onto the stash file tabulated in the table or drawn into the graph depending on their current monitoring settings Set Project descr
55. develop new models a par ticular ModelWorks GEM license must be obtained from ETHZ Furthermore it is necessary to purchase a TopSpeed Modula 2 license in order to be able to develop compile and link ModelWorks model definition programs Together with the ModelWorks software you may buy a full ModelWorks development kit for a reduced price from the Swiss Federal Institute of Technology Z rich ETHZ Note that similar to the Macintosh versions the ModelWorks software itself is not sold but distributed as public domain software For a detailed description of the ModelWorks software kit see below the section Installation of ModelWorks C How to Work With ModelWorks on Macintosh Computers All descriptions in this section are valid for all Macintosh ModelWorks versions C 1 INSTALLATION OF MODELWORKS ModelWorks is distributed on two 800 KBytes floppy diskettes They are organized similarly to those shown in Fig Al and Fig A2 The first diskette 1 2 contains the ModelWorks development system for the daily work the second diskette 2 2 a system the documentation and a collection of sample models ModelWorks contains partially the MacMETH Fast Modula 2 Language System software in particular the compiler debugger and linker Files and folders RMSMacMETH User Profile M2MiniLib M2Tools Moreover it contains also a complete copy of the Dialog Machine software folders DMLib AuxLib and RAMSESLib With this software you can fully dev
56. dif ferent color options for the same color monitor e g for an EGA monitor you have the options EGA 16 colors or EGA Monochrome Always choose the monochrome option to get the desired configuration Although the software might run under color GEM you will encounter memory space problems and draw restore of screens will be noticeably slower To test the installation of GEM you can copy the files MAKEPIC APP and WELCOME IMG from the Dialog Machine disk 2 onto the hard disk Then start GEM type GEM at the DOS prompt and start MAKEPIC APP by double clicking its icon You can read and display the file WELCOME IMG with the command GEM Bild einlesen NOTE GEM uses its own mouse drivers You may encounter problems with the mouse movement if any other software using the mouse is installed on your system In case of prob lems check that no other mouse software has its driver installed see the AUTOEXEC BAT file D 1 3 Installation of the Dialog Machine Copy the whole directory TOPSPEED onto your hard disk use either GEM or the DOS XCOPY S command If you you have the 360K disks you must copy the rest of the TopSpeed files from disk 2 of 4 manually into the directory C TOPSPEED Thus everything which is in the directories A TOPSPEED on both 360K disks goes into one directory C TOPSPEED on your hard disk Copy the whole directory DM including its sub directories onto your hard disk If you have the 360K disks you must copy the
57. doubles With increasing density limiting factors such as nutrients light energy or competition by the neighboring plants become more important This results in a decrease of the growth rate expressed as a self inhibition of the plants Finally the grass density reaches a maximum the so called carrying capacity determined by the plant s environment The following nonlinear differential equation describes the model dG t dt c Cll GW 1 where State variable grass g dry weight per m2 G t Initial amount of grass initial value G 0 1 0 g m Model parameters grass growth rate dav c 0 7 day Self inhibition coefficient m2 g day c2 0 001 m g day Let us have a closer look at the model and its equation The model has one state variable the grass density G t which is a function of time Further it has two constant model parameters c and c2 The first term of the differential equation c G t de scribes the exponential growth phase of the plants the second c2G t is responsible for the self inhibition The unknown element in Eq 1 is the function G t During a simulation this function is approximated by calculating a sequence of values G t G t G t gt given the ini tial value G t Since G t is defined by a differential equation these computations correspond to a particular solution of Eq 1 In other words By numerical integration ModelWorks produces the trajectory going through
58. e g called MyModelSet MAKE exactly as you would enter them after the compiler prompt Then each time you want to compile all modules belonging to your model system simply choose the file MyModelSet MAKE after entering Command key Shift 0 For instance all the sample models distributed together with the ModelWorks release can be compiled with this technique The needed file named MAKE Sample Models is distributed together with the sample models MacMETH uses the following extensions in file names MOD Text files Implementation and program modules DEF Text files Definition modules OBM Object files compiled implementation program modules SBM Symbol files compiled definition modules RFM Reference files used for debugging l The exact name may be slightly different e g RMSMacMETH 2 6 indicating also the shells version A 131 ModelWorks V2 0 Appendix Files produced by the compiler in particular SBM OBM and RFM get their names not from the name of the source file but from the module identifier Note also that files names extension included should not be longer than 16 characters Hence module identifiers should not be longer than 12 characters and should be identical to the source file name except for the extensions E g The following model definition program MyModel MODULE MyModel FROM SimBase IMPORT END MyModel should be saved in a text file with the name MyModel MOD In particu
59. even allow to write the two time series onto the same stash file 1On the IBM PC the name of this file is truncated to 8 characters and hence becomes MODELWOR DAT The file resides in the same directory as LOGISTIC APP 2 On the IBM PC use e g the editor of the JPI TopSpeed Modula 2 development environment 3 On the IBM PC use e g MODELWOR DA2 4 This a more advanced technique requiring multiple model declarations For more details on this subject please refer to the reference manual T 23 ModelWorks 2 0 Tutorial 2 2 8 PROGRAM TERMINATION The program can be terminated with the menu command File Quit After that the MacMETH environment becomes active again ready to accept your next command for instance the execution of another model To return to the desktop of the Finder use the menu command File Quit of the MacMETH shell On the IBM PC you will return to the GEM desktop T 24 ModelWorks 2 0 Tutorial 3 Getting Started with Modeling In this chapter you will get a closer look at the way ModelWorks models are defined First it is explained how the Modula 2 program defining the logistic grass growth sample model was written Then you learn how to define anew model by modifying an existing program text by using the MacMETH programming environment Finally the new model is made ready to be executed i e simulated using ModelWorks to start another simulation session Again it is assumed that you know how
60. for the display in small IO windows during a simulation session In particular on small screens IO windows become small in the tiled window position see menu command Tile windows and they will display U only this identifier to denote the state variable s Example sa unit String containing the unit used to measure values of the state variable s This string is displayed in IO windows during a simulation session Example cells ml 7 1 4 DECLARATION OF MODEL PARAMETERS A time invariant model parameter p which the simulationist should be able to change interactively during simulation sessions has to be declared with the procedure DeclP PROCEDURE DeclP VAR p REAL defaultVal minVal maxVal REAL runTimeChange RTCType descriptor identifier unit ARRAY OF CHAR DeclP may not be called another time unless the model parameter has been removed in the meantime DeclP may not be called in the sub state Running s a part II Theory Fig T26 The meaning of the formal procedure parameters are p Real variable to be declared as model parameter Dec P assigns to p its default value defaultVal The real p can be declared anywhere in the model definition program and may be even part of any data structure However make sure that it is declared as a global real variable and does exist as long as the model definition program defaultVal Default value for the model parameter p The default value is re assigned to the current parame
61. format is such that these results can be transferred directly for instance via the clipboard into another application Only horizontal tab characters t ASCII ht octal 11C separate the values and all values at a particular monitoring time are written on the same line terminated with a carriage return p ASCII cr octal 15C E g SRTF stands for Rich Text Format It is based on ASCII characters only but contains coded formatting information and can be interpreted by many commercially marketed text processing applications T 56 ModelWorks 2 0 Theory time t Ident var 1 t Ident var 2 p 0 000000 tT 1 0000000 tT 0 9025031 p 0 200000 tT 1 1764115 T 0 6883310 p 0 400000 T 1 2954322 tT 0 4211738 p 0 600000 tT 1 3516583 tT 0 0882961 p 0 800000 tT 1 3498297 tT 0 3198467 pP Normally the stash file is only opened and written if at least one monitorable variable has been requested for the recording However if the particular simulation environment mode preferences is set appropriately the stash file is opened during every simulation run and data are recorded according to the current settings of the recording flags The table used to tabulate the values of monitorable variables during a simulation appears in a window on the screen Values are written in a similar way as shown above under the stash file monitoring Currently only the values which fit into the window are displayed Once the window is full ModelWorks erases most of
62. get procedure retrieves the objects attributes the set procedure modifies overwrites them Moreover the procedures are grouped into two sets The first set is to modify the defaults the other to modify the current values The meaning of the formal procedure parameters are the same as described under the declaration procedures DeclM DeclSV DeclP DeclMV and DeclTabF Also the parameter lists were kept similar to the ones used by the declaration procedures 7 2 2 a Modification of defaults Setting defaults with any of the listed procedures will not imply a setting of the current values also i e no implicit reset Until the next corresponding reset no changes will become effective or visible Only the change of the descriptors identifiers and the unit strings as well as the change of the range boundaries used during the interactive changing of initial values or model parameter values via IO windows will become effective immediately PROCEDURE GetDefltM VAR m Model VAR defaultMethod IntegrationMethod VAR initialize input output dynamic terminate PROC VAR descriptor identifier ARRAY OF CHAR VAR about PROC PROCEDURE SetDefltM VAR m Model defaultMethod IntegrationMethod initialize input output dynamic terminate PROC descriptor identifier ARRAY OF CHAR about PROC PROCEDURE GetDefltSV m Model VAR s REAL VAR defaultInit minCurInit maxCurInit REAL VAR descriptor identifier unit ARRAY OF CHAR
63. in disk 271K available a System Folder oa Sample Models Docu Mock write W424 Shareware Fig A2 Contents of the Macintosh ModelWorks distribution diskette 2 2 There are two ways to work with ModelWorks with a hard disk or with two floppy diskettes In case you should have a hard disk the installation is very simple Copy the whole contents of the first diskette into a new folder which you may name something like ModelWorks on your hard disk and you are ready to start working In case you should have no hard disk you can work directly with the working copy of the two distribution diskettes never use your original distribution diskettes for this purpose The second diskette contains a system with the desk accessory MockWrite MEdit 4 2a already installed You can use this desk accessory to edit your models In case you prefer a different system than the one provided on disk 2 2 you can freely replace it with your favorite system folder However in this case it is either recommended you install first the desk accessory lYou should have at least 800 KBytes of free disk space For large model development projects it is recommended to have 1 MBytes available in case large data sets are involved enlarge the needed disk space accordingly Models tend to remain small e g within 40 K for source OBM and RFM files fits already quite a sophisticated model definition program A 128 ModelWorks V2 0 Appendix MockWr
64. information partly always included and partly included only selectively by means of the so called recording flags The content consists of General information on the simulation session consisting of a the date and time of the session s begin b date and time of begin and end of simulation runs c project title remarks and footer d date and time when the file was closed It is always written Values of all global simulation parameters Start and stop time of simulation integra tion step respectively maximum integration step plus maximum local relative error discrete time step respectively coincidence interval and monitoring interval The para meters actually written on the stash file depend on the type of models currently pre sent continuous time only discrete time only or both types mixed as well as the used integration methods with or without variable step length methods This type of in formation is written always and in particular also repeatedly for every simulation run Lists of all models and their integration methods of all state variables and their current values of all model parameters and their current values of monitorable variables and their settings curve attributes and scaling are written selectively under control of the recording flags Note that not all monitorable variables are recorded but only those for which either the stash filing is currently set F writeOnFile or those which are present in the grap
65. last table will be kept untouched until the next simulation run is started Only at that time the old table is cleared and a new one will be drawn The number Common rows between page ups in table defines what happens during a page up A page up occurrs when the table window is full but more rows should be written then ModelWorks attempts to erase most of the table and restarts tabulating from the top again This number specifies first how many rows at the bottom are not erased but copied to the top of the next page All remaining space below is then used to add the rows of the new page Thus this number specifies how many rows are common to two consecutive pages Graphing R 74 If the simulation environment mode Once changed immediately redraw graph is active the graph is redrawn immediately after each change in the graph settings Otherwise the last graph will be kept untouched until the next simulation run is started Only at that time the old graph is cleared and a new one will be drawn ModelWorks 2 0 Reference Activating the simulation environment mode Restore graph with colors and high quality vector graphics for printing and clipboard is active the graph is restored with colors when a previously covered graph portion becomes visible again or will be printed or transferred to the clipboard in colors and with high quality If this mode is turned off a bitmap is used instead for restoring printing or transfer into the clipboard p
66. later a complete erase and redrawing of the content of the graph window because it always affects the legend The actual redrawing takes place according to the current settings of the mode RedrawGraphAlwaysMode of the simulation environment E Reset curve attributes Resets the curve attributes of the selected monitorable variable s to their default values R 97 ModelWorks 2 0 Reference 7 Client Interface The client interface consists of a mandatory part an optional part and an auxiliary library Fig T25 part II Theory The mandatory part consists of the two modules SimBase and SimMaster and the optional part of the modules TabFunc SimIntegrate and SimGraphUtils Although every model definition program must import from both modules of the mandatory part only a small subset of the exported Modula 2 objects are really needed always These few Modula 2 objects five procedures and six data types form the core of the ModelWorks client interface Fig T25 part II Theory All other types and procedures also exported from this interface are optional Their purpose is either to serve the convenience of the simulationist or to support the modeler in the programming of advanced structured simulations For instance if the simulationist wishes to run a model in a time range different from the one predefined by ModelWorks the modeler can overwrite the ModelWorks defaults Tab T1 part II Theory with the values the simulationist prefers Th
67. linker is the reverse from that during the dynamic linking loading applied by the MacMETH shell Fortunately there is an easy work around to any file not found problems during linking for modules already prelinked in SimMaster OBM Simply add the imports from module SimMaster at the end of all the imports in the model definition program e g like this FROM FROM SimBase IMPORT FROM SimMaster IMPORT RunSimMaster Recompile the model and link the application as described above Should you encounter any new run time problems with the resulting application which were never visible while running the unlinked program check the contents of the resource fork with the resource editor The following resources should be present ALRT DM error alert proc module 1003 DM simple error alert Bm ErrorAlert Dialog Machine 1003 1004 314 6000 313 410 310 311 312 Prog Stat Big Message Logo Date About MacMETH Modula 2 Error Loader Status Program Status Pos N SSCS mvButtAct mButtAct svButtAct pButtAct simpScrAct aboutResource 2 0 ModelWorks preferences 7418 Any resource listed above but missing in the application will result in an application not func tioning properly However additional resources should cause no harm A 135 ModelWorks V2 0 Appendix D How to Work With ModelWorks on IBM PCs Written by Daniel Keller Project Centre IDA ETHZ 26 April 1990 D 1 IN
68. model M given by Eq 13 and once b modeled as a structured model system consisting of two submodels jz and u according to the Eq 14 and 15 Both submodels are of the type continuous time and case A applies with the equations 6 a till 10a but no global inputs nor global outputs are present Each of these submodels has one input and one output defined according to Eq 7a and 9a These inputs and outputs have only been introduced in order to couple the two submodels They form a structured model system each submodel containing one of the differential equations from Eq 13 Mathematically the structured model system formed with 14 and 15 is equivalent to the one given by Eq 13 However discretisation errors may result in the sample and hold effect described above see also below under Simulation environment of the next chapter Fig T14 This is because no information exchange across submodel boundaries takes place during an integration step Thus coupling among submodels occurs only at the endpoints of an integration step s a Fig T12 In case a higher order integration method is used the coupling of differential equations within a submodel takes place even in the middle of an integration step Hence simulation results of the continuous time part of a structured model might slightly differ for non single step integration routines depending on where the modeler has chosen the submodel boundaries between the differential
69. models this simulation parameter becomes the Maximum integration step h hyax The actual integration step will be determined by the variable step numerical integration algorithm and globally used as the integration step for all other submodels even if they should be solved with a fixed step length method Maximum relative local error e Only if at least one continuous time sub model is using a variable step length integration method this simulation parameter appears in the entry form It determines the maximum relative local integration error estimated by comparing a higher order result with a lower order result If a norm of this error vector lel divided by a norm of the state vector Ixl exceeds e the integration step length h is halved till lel Ixl lt e Otherwise h is doubled unless one of the following two conditions would become true lel Ixl gt e or h gt hyax Discrete time step c If only discrete time sub models are present case B c may be edited in place of the integration step h However in this case the actual value of c is irrelevant since the length of an interval between two discrete time points has no meaning if only discrete time sub models are present This 4The smaller the step h is the more accurate is the calculation unless h gets so small that truncation errors become dominant the larger h is the faster runs the simulation Therefore the simulationist has to select a good compromise which
70. monitorable variables SET Columns set up Activates an entry form in which the display of the columns in the IO window Monitorable variables can be controlled Fig R23 The columns of which the display can be turned on or off are Monitorable variable names Full names of the monitorable variables Ident Short identifiers of the monitorable variables Unit Unit in which to measure the values of the monitorable variables Minimum scaling Lower value used to scale the values of the monitorable variable on the ordinate of the graph By default this column is not shown Maximum scaling Upper value used to scale the values of the monitorable variable on the ordinate of the graph By default this column is not shown Monitoring This column shows the current output settings where R 93 ModelWorks 2 0 Reference gt Monitorable variable is written onto the stash file T Monitorable variable is tabulated in the table X Monitorable variable is used as independent or abscissa variable x values If there is no monitorable variable selected as the abscissa variable ModelWorks provides the so called default independent variable the simulation time Y Monitorable variable is used to draw a curve Its values are drawn as ordinate values y values versus the current independent variable x values Check the columns to be displayed in the window EJ Monitorable variable names EJ Ident S Unit O Minimum
71. parameters listed in Tab R2 The variables listed in Tab R3 are used for the project description The variables in the tables Tab R1 R2 and R3 are referenced in the program texts below with the listed identifiers The symbols listed are the ones which have been used to denote the variables in the manual in particular in the part II Theory Identifier mn Meaning o E CurrentTime Lead ior simulation time or independent variable read only for continuous time sub models CurrentStep Current simulation time or independent variable read only for discrete time sub models CurrentSimNr Number of the current simulation run read only Tab R1 Read only global simulation variables internally used by ModelWorks The first column contains the identifiers used to designate the corresponding variables in the client interface the second the symbol used to denote the corresponding variable in this manual Identifier Symbol BUG TE oa Simon h L Integration step if fixed step length method maximum integration step if at least one variable step length method in use h is only used if at least one continuous time model present otherwise ignored Maximum relative local error er is only used if at least one variable step length method in use zZ Discrete time step if only discrete time models present Coincidence interval if continuous as well as discrete discrete time models present Monitoring interval Tab R2 G
72. problem with the initial values model and global simulation parameters as inputs plus the monitorable variables as outputs The algorithms are given by the run time system of ModelWorks Edit Compile S i mulation Session j Eig T23 Development cycle of the modeler writing ModelWorks model defini tion programs T63 ModelWorks 2 0 Theory The modeling process consists of the model development cycle with the steps editing compila tion and execution of the model definition program Fig T23 5 2 2 STRUCTURE OF MODEL DEFINITION PROGRAMS A model definition program may be built from as many modules as the modeler wishes Ty pically structured models are built from several modules external or library modules each submodel corresponding to a Modula 2 module Fig T24 If the modeler makes use of modular modeling the only thing to pay attention to is to make sure that outputs from one submodel are computed in its procedure Output and the depending inputs of another submodel in its procedure Input Fig R16 The structure of the model The structure ofthe Modula 2 prog ram Output of ModelDeclarations Model 1 Declaration of Model Declaration of Model2 Output of Output of Model1 y1 Model2 y2 Model1 DEF Model2 DEF Output of Input to Model 2 nea 2 odel1 MOD Model2 MOD u1 y2 y2 ju2 y1 y1 Fig T24 Mapping of a structured model comp
73. rest of the DM OBJ files from disk 4 of 4 manually into the directory C DM OBJ analogous to the TopSpeed files above Now edit the AUTOEXEC BAT file on your hard disk The TopSpeed directory must be in cluded in the PATH statement this line in the file should look something like PATH C DOS C MOUSE C TOPSPEED C depending on your system usage After hav ing edited the AUTOEXEC BAT file you must restart the computer To test the installation of the Dialog Machine we will compile and run a little program Dialog Machine programs must be edited compiled and linked with JPI TopSpeed Modula 2 under DOS but the finished programs run under GEM A 137 ModelWorks V2 0 Appendix If you are within GEM exit Atthe DOS prompt type CD DM EXAMPLES to get to the di rectory where the test program is Then type M2 to start the TopSpeed compiler Before compiling the file I would strongly recommend to set the compiler option shortcut to get to the compiler options ALT O C Runtime checks default to ON it s like driving without a seat belt otherwise Options must only be set once per directory the compiler keeps a list of your preferences in the file M2 SES which is read automatically at each start of a session Load F3 the file FIRST MOD and compile it ALT C Then link it ALT L If this was all successful exit the TopSpeed environment ALT X and rename the file FIRST EXE to FIRST APP GEM recognizes a real GEM applicat
74. run a post analysis from the stash file you should at least have the recording flags Models and Monitorable variables active If you rather use the stash file for run documentation purposes it is recommended to have the simulation environment mode Always document run on stash file active The recording flags together with the stash file setting F of the monitorable variables will then solely control the kind of information and data written to the stash file s a menu command Preferences O BUMaster LC DAs O DataWork O Discard 7 DocuPrep D EF Global san Parsi Stash file ModellDorks DAT Fig RR Dialog box Select stash file F Select stash file F Allows to select a stash file Fig R8 with the usual open file dialog box Note that this command will not really open the file so that the simulationist may open it for inspection during the program state No simulation Once a simulation starts i e ModelWorks enters the program state Simulating the stash file will be automatically opened and remains open till the state Simulating will be left Since during a whole structured simulation ModelWorks remains in the state Simulating this means that all results from all simulation runs are normally written on the same stash file The default name used in the file selection dialog box is the current stash file name Note that if there is not at least one monitorable variable present for which the current sta
75. run will have no effect Important notice if a file with the same name should already exist it will be overwritten without any warning PROCEDURE Message m ARRAY OF CHAR Writes the text m onto the stash file and inserts it in the table This procedure surrounds the string m with quotation marks and precedes it with the reserved word MESSAGE This procedure allows to bring state events to the user s attention which would otherwise slip by undetected or it helps the user to locate particular events while reading large stash files PROCEDURE DumpGraph If the stash file is currently open currently stashFiling attribute F for at least one monitorable variable active DumpGraph writes the current graph onto the stash file The data are written in the so called RTF Format which can be opened by several commercially available document processing software s a previous section on recording flags in the entry form Project description under menu Settings Not available in Reflex and PC version 7 5 4 GRAPHICAL MONITORING The following objects allow to control the curve attributes used by ModelWorks for the monitoring of the simulation results in the graph for the specified monitorable variable TYPE Stain coal snow ruby emerald sapphire turquoise pink gold autoDefCol LineStyle unbroken broken dashSpotted spotted invisible purge autoDefStyle CONST autoDefSym 200C R118 ModelWorks 2 0 Reference
76. simulated larval density for whole valley PROCEDURE ActivateLarchLBMModel PROCEDURE DeactivateLarchLBMModel PROCEDURE LarchLBMModelIsActive BOOLEAN END LBMMod IMPLEMENTATION MODULE LBMMod Revision history Author Date Description af Dez 86 First implementation af 12 05 90 ModelWorks 2 0 adaptation now dynamic model activation and de activation supported FROM MathLib IMPORT Exp Ln FROM SimMaster IMPORT RunSimMaster FROM SimBase IMPORT Model DeclM IntegrationMethod DeclSV StashFiling Tabulation Graphing DeclMV DeclP RTCType NoInput NoTerminate NoAbout RemoveM SetSimTime FROM LBMObs IMPORT negLogDelta yLL yUL kmin kmax time domain VAR m Model c1 c2 c3 c4 c5 c6 c7 c8 c9 c10 c11 c12 c13 c14 c15 c16 c17 nrt REAL pl p2 p3 p4 p5 p6 p7 p8 p9 pl10 pl1 pl2 p13 pl4 REAL rt rtl et etl REAL def springEggs REAL curActive BOOLEAN PROCEDURE Initialize PROCEDURE Parameters BEGIN pl c4 p2 c5 p3 c2 c6 1 0 cl p4 c6 1 0 c1 1 0 c3 p5 c2 c7 c9 c10 1 0 cl p6 c9 1 0 c1 c2 c7 c11 c10 c2 1 0 c8 c7 1 0 c3 p7 c9 1 0 c1 c10 1 0 c3 1 0 c8 c11 c2 1 0 c8 c7 1 0 c3 p8 c9 c11 1 0 cl 1 0 c3 1 0 c8 p9 c12 pl0 c13 pll c14 pl2 c15 pl3 c16 pl4 c6 c17 nrt END Parameters BEGIN Initialize Parameters END Initialize PROCEDURE Output BEGIN yt p3 rt p4 et pl4 A 165 ModelWorks V2 0
77. strategy to distribute four colors stains four line styles and four symbols to draw curves see Tab T1 in part II Theory for these monitorable variables This strategy helps the simulationist to tell curves optimally apart regardless whether they are displayed on a color screen or are printed on a black and white printer only Colors line styles and symbols are distributed among the monitorable variables which currently have the automatic definition of curve attributes setting active Which attribute e g color is used for which curve is influenced by the position of monitorable variables within the sequence in which they have been activated for the monitoring For instance if four monitorable variables have been activated for graphing the first activated will automatically be drawn in black the second in red the third in blue and the fourth in green However if the second is removed from the monitoring the previously third will be drawn in red and the previously fourth in blue The user may override this automatic definition and use a particular color line pattern and marking symbol for a curve of a specific monitorable variable This allows e g to use always the same color for the same variable such as green for state variable Grass or blue for a water level Mark symbols are characters drawn exactly at the points as defined by the monitoring line patterns are used to connect these points with lines With this technique it is also poss
78. structured model system composed of three elementary coupled submodel systems The global input u is defined by Eq 6a b or c the inputs of the subsystems uij by Eq 7a b or c the outputs of the subsystems Vij by Eq 9a b or c and the global output y by Eq 10a b or c A A structured model composed of n elementary coupled submodels each continuous time and defined according to Eq 4 x is defined as follows for an example see also Fig T9 The input of the global system is given by u u t global input 6a The input to the submodel i depends on the global input u and on the output y t of the n submodels output input coupling uj t h u t yz t WA input of submodel i 7a The dynamic equations for obtaining the derivatives of the state variables of the submodel i T38 ModelWorks 2 0 Theory dx t dt fj Lun uj t pfit t dynamic equations of submodel i 8a The output of the submodel i depends on the states and the parameter set of the submodel i An output variable must not depend directly on an input variable no direct output input coupling Since the input variables may depend directly on the output variables a direct output input coupling would lead to a circularity which could not be resolved generally yi gj x t Dold t output of submodel i 9a Some of these outputs are not connected to other submodels but are global outputs These elements from
79. that this simulation step can only be read but not changed PROCEDURE CurrentTime REAL Returns the current simulation time i e the current value of continuous time Note the simulation time can only be read but not changed PROCEDURE CurrentSimNr INTEGER Returns the current simulation run number k during structured simulations k 1 2 3 Fig T16 Note that even aborted runs are numbered This procedure is typically called in the client procedure initialize e g to assign parameter values depending on the current run Note k can only be read but not changed 7 2 1 6 Modification of defaults The predefined values ModelWorks uses as defaults are listed in Tab T1 manual part II Theory If the modeler wishes to change i e overwrite them he may access any of the variables listed in the tables Tab R2 or R3 with a SerDefltxyz procedure i e a procedure with an identifier starting with SetDefit PROCEDURE SetDefltGlobSimPars t0 tend h er c hm REAL PROCEDURE GetDefltGlobSimPars VAR t0 tend h er c hm REAL PROCEDURE SetDefltProjDescrs title remark footer ARRAY OF CHAR wtitle wremark autofooter recM recSV recP recMV recG BOOLEAN PROCEDURE GetDefltProjDescrs VAR title remark footer ARRAY OF CHAR VAR wtitle wremark autofooter recM recSV recP recMV recG BOOLEAN Above procedures set or get the defaults for the global simulation parameters the project description or the recordin
80. the bold model title is interpreted as selection of all elements belonging to On the Macintosh no preparations are necessary to follow this tutorial except that you should be using a working copy of the software Working through the tutorial will change the contents of your diskettes so don t use your originals On the IBM PC you are ready only if you have followed exactly the installation procedures described in the Appendix in particular those for the installation of the ModelWorks software You should have an executable GEM application made from the sample model LOGISTIC MOD which is now called LOGISTIC APP 2 On the IBM PC you have to start first the GEM desktop and then to start the application LOGISTIC APP In case you should encounter difficulties up to this point please refer to the GEM documentation and or check your installation s a the Appendix Ignore this and the next paragraph to which this footnote belongs they contain Macintosh specific information only 3 Please refer to the Appendix for more details on the exact organization of the ModelWorks diskettes and how to work with MacMETH 4 In case you should have any difficulties up to this point executing the described steps on your computer please refer to the MacMETH documentation and or check your installation see also the Appendix for detailed instructions how to install ModelWorks and a brief introduction how to work with MacMETH T17 ModelWorks 2 0 Tutorial
81. the data table provided by the module TabFunc While a selection is currently made the simulationist may use within an edit field the key equivalents C for copying the selection into the clipboard X to cut copy plus delete the selection into the clipboard and B to blank delete without copy the selection The current content of the clipboard if text may be pasted into an edit field at the current location of the insertion bar or as a replacement for the current selection by pressing V This technique allows also to transfer textual data between different entry forms and between different applications given they support the clipboard for text If no entry form or other dialog box is currently in use the clipboard accessing keyboard equivalents have the usual meaning see below Menu Edit l On the IBM PC press the Ctrl key simultaneously with the key X 2 On the IBM PC keyboard shortcuts function only if a letter is involved hence the cancel function with Ctrl is not available Use Escape instead R72 6 1 1 OVERVIEW OVER MODELWORKS STANDARD MENUS Page setup pada Set Tile windows Print graph Global simulation parameters 1 Stack windows E E ES Cut 3 H Project description IPP hara nnii Preferences Copy BEL bes en Models EEE E A Paste amp U Select stash file F State variables Quit n Clear 2 9 aina a Model parameters Reset Monitorable variables Global simulation pa
82. the unit used to measure values of the monitorable variable mv This string is displayed in IO windows during a simulation session Example cells ml defaultSF defaultT defaultG Default settings for the kind of monitoring for the monitorable variable mv If defaultSF defaultT defaultG are selected to be written on a file tabulated or to be plotted the values of the variable mv is written in the default stash file resp table or drawn in the graph as a curve versus the current independent variable usually simulation time The defaults for the kind of monitoring are re assigned to the current kind by ModelWorks during the initialization phase of the simulation session or after every reset of the stash filing tabulation respectively graphing During a simulation session the current kind of monitoring may be changed by the simulationist using the IO window for monitorable variables or by the modeler via procedure SetMV 7 1 6 DECLARATION OF TABLE FUNCTIONS Functions given by a table of values may be declared with the procedure DeclTabF which has to be imported from the optional module TabFunc s a section Optional menu Table Functions manual Part III Reference R 105 ModelWorks 2 0 Reference TYPE TabFUNC PROCEDURE DeclTabF VAR t TabFUNC XX YY ARRAY OF REAL NValPairs INTEGER modifiable BOOLEAN tabName xName yName xUnit yUnit ARRAY OF CHAR xMin xMax yMin yMax REAL This procedure can
83. the values of a particular table function The functions are listed with their identifiers Table Function Editor E Growth rate RTt growth rate Hr TemperatureGrowth rate oyo oO oF 20 0 jo 17 wa La a o L Gh A aE U K om o LA ul 0 00 10 00 20 00 30 00 40 00 50 00 Tempera ture Fig R13 Table Function Editor to display Draw and edit values of a table function Secondly a window will be displayed similar to the one shown in Fig R13 It shows a graph of the table function and all its values at the right of the graph The first column in this table contains the values of the independent variable x values and the right column the values of the dependent function value y values All these values may be edited however note that it is required that the x values are always in ascending order and that all x values must be different i e if there are n x values with the index i they must satisfy the condition x lt X lt X3 lt lt Xi lt lt Xp 1 lt Xp The simulationist is asked to correct values which do not satisfy this condition ModelWorks 2 0 Reference Several push buttons at the lower left corner may be used to issue commands The two in the middle Initial or Draw may be used any time Button Jnitial discards the momentary editing and reloads the default values into the value table at the right of the graph Do not confound this with a reset since a reset wou
84. time model If the method discreteTime is specified the model is declared as a discrete time model All other integration methods may be used for the class of the continuous time models The default integration method is re assigned to the current integration method by ModelWorks during the initialization phase of the simulation session or after every reset of the integration methods During a simulation session the current integration method may be changed by the simulationist using an IO window or by the modeler via the procedure SetM for any of the continuous time models For a discrete time model however once declared its integration method may of course never be changed Note that this mechanism makes it possible that every continuous time model uses a different integration method 3 Should DecIM be called for the same model variable m more than once ModelWorks will display an error message and the simulation program will be halted R 100 ModelWorks 2 0 Reference The following five formal procedure parameters are procedures which will be called by ModelWorks during simulations initialize is called only once at the begin of each simulation run Fig T16 part II Theory It may be used freely to execute any task at the begin of a run such as opening a file for writing data during the simulation run or assigning new initial values to state variables by calling SetSV input calculates the input variables of sub model m
85. time or single step integration methods will skip this step Finally the state variables and the independent variable time are updated to their new values Afterwards the termination criteria is evaluated The simulation will be terminated if either the simulation stop time has been reached the simulation run was stopped killed by the simula tionist or if the termination condition from the model definition program has returned true Depending on the result the simulation continues or stops which will result in a state transition from the state Simulating into the state No simulation s a Fig T15 Discrete time models are treated analogous to the continuous time models ModelWorks treats basically both the same except that the discrete time submodels are integrated with a different integration method The new value of a difference equation which is analogous to the derivative of a differential equation is treated like a derivative but the formula to compute the new state differs i e it is just the assignment of the derivative to the new state The situation is more complicated in case of continuous with discrete time mixed simulations since the discrete time step might be several times larger than the integration step needed for the continuous time sub models it is obvious that the two types of models must be treated separately Typically the dis crete time submodels will then not be called as often as the continuous time ones an
86. transitions starting from a state f ARRAY firstState lastState firstState lastState OF REAL frequencies of transitions fs ARRAY firstState lastState OF REAL frequencies of states randomize REAL controls seed randomization PROCEDURE PRED s State State BEGIN PRED RETURN VAL State ORD s 1 END PRED PROCEDURE SUCC s State State BEGIN SUCC RETURN VAL State ORD s 1 END SUCC PROCEDURE Initialize VAR 1 Individuals is js State u REAL BEGIN Initialize IF randomize gt 0 0 THEN Randomize ELSE ResetSeeds END FOR is firstState TO lastState DO fs is 0 0 END FOR FOR 1 firstIndiv TO lastIndiv DO u U IF u lt initP one THEN x l one ELSIF u lt initP one initP two THEN x 1 two ELSE x l three END IF fs x 1 fs x 1 1 0 END FOR FOR is firstState TO lastState DO FOR js firstState TO lastState DO c is js 0 f is js 0 0 END FOR n is 0 END FOR FOR is firstState TO lastState DO pacc is firstState p is firstState FOR js SUCC firstState TO lastState DO pacc is js pacc is PRED js p is js END FOR END FOR fs firstState fs firstState FLOAT lastIndiv firstIndiv 1 FOR is SUCC firstState TO lastState DO fs is fs PRED is fs is FLOAT lastIndiv firstIndiv 1 END FOR END Initialize PROCEDURE InitSimSess VAR is js State curMin curMax REAL
87. y k of the discrete time submodels j gt E A y x t y t yas E3 7 l global outputs 10c y EA Yn 193k Lyn The general definition of the couplinghas two special cases which are often of interest to the modeler Structured model consisting of several but uncoupled submodels The inputs of the submodels do not depend on any output of another submodel u t h u t resp u k hj u k Such submodels coexist as completely independent units Implementing them in such a way offers the advantage that the models can be simulated in parallel at once This is useful to work simultaneously with a set of identical or similar models e g to test different model versions of the same real system or to compare a measured time series parallel model with a simulated trajectory model T41 ModelWorks 2 0 Theory The structured model is composed of hierarchically organized submodels several levels An example of such a hierarchical model system is given in Fig T11 two levels Note however that ModelWorks ignores the hierarchical organization which is only of concern to the modeler ModelWorks treats all models exactly the same way regardless of the level on which they are declared Fig TIl Example of a hierarchically organized structured model system composed of several submodels which are themselves structured model systems consisting of several internally coupled submodels
88. yUL Maximum larval density per site Ymax larvae kg branches notOnFile notInTable notInGraph SetDefltCurveAttrForMV obsMod ymaxDash turquoise spotted 0C DeclMV yminDashLn Ln negLogDelta Ln yUL Ln of minimum larval density per site Ln Ymin larvae kg branches notOnFile notInTable notInGraph SetDefltCurveAttrForMV obsMod yminDashLn turquoise spotted 0C Dec1MV ymeanDashLn Ln negLogDelta Ln yUL Ln of average larval density in valley Ln Y larvae kg branches notOnFile notInTable isY SetDefltCurveAttrForMV obsMod ymeanDashLn turquoise dashSpotted 0C DeclMV ymaxDashLn Ln negLogDelta Ln yUL Ln of maximum larval density per site Ln Ymax larvae kg branches notOnFile notInTable notInGraph SetDefltCurveAttrForMV obsMod ymaxDashLn turquoise spotted 0C END ModelObjects PROCEDURE ActivateLBMObsModel BEGIN IF NOT curActive THEN DeclM obsMod discreteTime NoInitialize NoInput Output NoDynamic Terminate ModelObjects Observations from the Upper Engadine Valley Obs UE NoAbout SetSimTime FLOAT kmin FLOAT kmax curActive TRUE A 170 ModelWorks V2 0 Appendix END IF END ActivateLBMObsModel PROCEDURE DeactivateLBMObsModel BEGIN IF curActive THEN RemoveM obsMod curActive FALSE END IF END DeactivateLBMObsModel PROCEDURE LBMObsModelIsActive BOOLEAN BEGIN RETURN curActive END LBMObsModelIsActive BEGIN InitD
89. 0 20 g m Model parameters Grass growth rate day cy 0 4 day Self inhibition coefficient m2 g day c2 8 105 m g day Grass consumption rate by aphids m2 gl dav c3 1 5 103 m gl day Aphids birth rate per grass consumption g eh c4 01 g gl Death rate of aphids dav c5 0 2 day 1 Early this century these models have first been formulated by LOTKA 1925 and VOLTERRA 1926 Their purpose is to describe the population dynamics of a prey and a predator species 2 The new parameter and initial values are not necessarily realistic since the sole purpose of the model is to help to learn ModelWorks T 28 ModelWorks 2 0 Tutorial Let us have a closer look at the new model and its equations The first equation is the same as before except that the term c3 G t A t has been added This term is responsible for a decrease of the net grass growth due to grass consumption by aphids The second equation describes the dynamics of the aphids They can grow by feeding on the grass which is expressed with the term c3 c4 G t A t The second term c5 A t accounts for the natural mortality of the aphids In the next section it will be explained how to alter step by step a copy of the logistic grass growth sample program to implement this new grass aphids model then the program will be compiled and finally be simulated It is assumed that you know how to edit a program text and that you are familiar with the follow
90. 1 NOTE Always call this procedure before calling N PROCEDURE SetPars mu stdDev REAL PROCEDURE GetPars VAR mu stdDev REAL Set or get the current parameters U mean and the stdDev standard deviation SQRT variance for the normally distributed random variable for which procedure N returns variates PROCEDURE N REAL A151 ModelWorks V2 0 Appendix Returns a variate from a normally distributed random variable with mean mu and the standard deviation stdDev as currently set by procedure SetPars The default values for u mu respectively stdDev are 0 resp 1 0 Method The variates are computed by the method Box and Muller and the Von Neumann rejection technique PROCEDURE ResetN The used method computes each second time two values which are returned by N upon the next call In order to produce completely defined results for instance after setting a new seed value in the basic random number sequence used by U call this procedure in order to reset the internal modes END RandNormal H Sample Models The following sample models have all been implemented and tested with the ModelWorks Macintosh versions V2 0 V2 0 Reflex V2 0 II and some with the PC version 2 0 PC They are distributed in source form together with some additional sample models not listed here on the Macintosh versions distribution diskette ModelWorks 2 2 in the folder Sample Models or on the PC version distributi
91. 13 ModelWorks 2 0 Reference parameters independently from each other entry forms test only syntax and ranges this consistency test is important in case the model equations would become undefined if the conditions were not met Moreover the modeler may use this procedure to compute values of auxiliary variables which depend on the current values of parameters PROCEDURE InstallTerminateCondition tc TerminateConditionProcedure Procedure tc is called at the end of each time integration step during simulation If it returns TRUE the simulation will be terminated This behavior can be used to program state events which lead to the simulation termination Note however that this does not fully conform to a proper handling of state events since ModelWorks performs no iterations to find the exact location of the event You have to program fc such that the value returned is correct even if the current time is not exactly that of the event i e the procedure tc must be able to detect the state event even if it occurs anywhere in the time interval of the current integration step h PROCEDURE PauseRun Makes a state transition from the program state Simulating into the program state Pause Fig T15 and T26 and will only return after the simulationist has chosen the menu command Resume run under menu Simulation This feature allows to temporarily interrupt a simulation run exactly at a particular point such as a state event e g a state varia
92. 23 However if the modeler extends the simulation environment by corre sponding menu commands the dynamic declaration and the removing of models and model ob jects becomes possible and there applies no longer any restriction to the model installation s a part III Reference the chapter Client interface For this purpose the modeler has to use the pro T 64 ModelWorks 2 0 Theory cedures DeclM resp RemoveM from the client interface and to install menu commands using the procedures nstallMenu and InstallCommand from module DMMenus of the Dialog Ma chine Removing a model implies the removing of all its model objects 5 2 4 MODULE STRUCTURE OF MODELWORKS In order to link model definitions to ModelWorks the modeler needs the client interface It consists of amandatory and an optional part and an auxiliary library The mandatory part con sists of the two library modules SimBase plus SimMaster and the optional portion of the mod ules TabFunc SimIntegrate and SimGraphUtils Fig T25 Any model definition program has to import from the mandatory client interface Model Definition Program LEBEN US EI ERE SimGraph Utils SimMaster Simintegrate RAMSES SimBase Aux Lib ModelWorks internal modules Dialog Machine Fig T25 Module structure of ModelWorks programs The model definition pro gram imports from the client interface which consists a
93. A HH A KH A a a a a PROCEDURE Ask question ARRAY OF CHAR butTexts ARRAY OF CHAR butWidth CARDINAL VAR answer INTEGER PROCEDURE ShowPredefinedQuestion fileName ARRAY OF CHAR alertID INTEGER strl str2 str3 str4 ARRAY OF CHAR E 2 2 2 E k e e e e e E E E KE E Re Ae e eee eee a DMTextFields VAR answer INTEGER EEE A177 The Dialog Machine may be freely copied but not for profit 1 0 A 178 ModelWorks V2 0 Appendix TYPE TextPointer POINTER TO TextSegment TextSegment ARRAY 0 32000 OF CHAR PROCEDURE RedefineTextField textField EditItem wf WindowFrame withFrame BOOLEAN PROCEDURE WrapText textField EditItem wrap BOOLEAN PROCEDURE CopyWTextIntoTextField textField EditItem VAR done BOOLEAN PROCEDURE CopyTextFromFieldToWText textField EditItem PROCEDURE SetSelection textField EditItem beforeCh afterCh INTEGER PROCEDURE GetSelection textField EditItem VAR beforeCh afterCh INTEGER PROCEDURE GetSelectedChars textField EditItem VAR text ARRAY OF CHAR PROCEDURE DeleteSelection textField EditItem PROCEDURE InsertBeforeCh textField EditItem VAR text ARRAY OF CHAR beforeCh INTEGER PROCEDURE GetTextSizes textField EditItem VAR curTextLength nrLns charHeight firstLnVis lastLnVis INTEGER PROCEDURE GrabText textField EditItem VAR txtbeg TextPointer VAR curTextLength INTEGER PROCEDURE ReleaseText textField EditItem PROCEDURE FindInText textField
94. AL descriptor identifier unit ARRAY OF CHAR defaultSF StashFiling defaultT Tabulation defaultG Graphing and the real mv passed as actual argument is associated with the ModelWorks monitoring mechanism i e its values may be written onto the stash file tabulated or plotted in the graph from within the simulation environment DeclMVmay not be called another time for the same real variable mv unless it should have been removed in the meantime DeclMVmay may not be called in the sub state Running s a part II Theory Fig T26 The meaning of the formal procedure parameters are The following types control the actual monitoring settings for each kind of monitoring R 104 ModelWorks 2 0 Reference TYPE StashFiling writeOnFile notOnFile Tabulation writeInTable notInTable Graphing isX isY isZ notInGraph The monitoring settings can be independently activated or deactivated and for every monitorable variable the simulationist can control them interactively during simulation sessions The meaning of the formal procedure parameters of DecIMV are mv The variable to be declared as monitorable variable Note DeclMV assigns to mv the value 0 0 The real mv can be declared anywhere in the model definition program and may be even part of any data structure However make sure that it is declared as a global real variable and does exist as long as the model definition program defaultScaleMin defaultScaleMax Default m
95. Ae Se II Ne AE He He Ae Se Meee HEB HE He He Kee He HE OR eA DMLanguage KR KK KEK KEK KEK KEK KKK KEK KEK KK KKK KKK TH HH A A HH HH KKK KK TYPE Language English German French Italian MyLanguagel MyLanguage2 VAR okButtonText cancelButtonText ARRAY 0 15 OF CHAR PROCEDURE SetLanguage l Language PROCEDURE CurrentLanguage Language EER Fe Ie IER iel Ee IE I CIEE IE FE I EEE DMMaster KEK KEK KKK KKK KEK KKK KK KEK KEK KK KKK KKK TH HH A HH A KH A a a a a TYPE MouseHandlers WindowContent BringToFront RemoveFromFront RedefWindow CloseWindow MouseHandler PROCEDURE Window Status normal abnormal VAR DMMasterDone BOOLEAN PROCEDURE InstallSetUpProc suP PROC PROCEDURE GetSetUpProc VAR suP PROC PROCEDURE InstallMouseHandler which MouseHandlers mhP MouseHandler PROCEDURE GetMouseHandler which MouseHandlers VAR mhP MouseHandler PROCEDURE InstallKeyboardHandler khP PROC PROCEDURE GetKeyboardHandler VAR khP PROC PROCEDURE Read VAR ch CHAR PROCEDURE BusyRead VAR ch CHAR PROCEDURE CmdKeyPressed BOOLEAN PROCEDURE OptKeyPressed BOOLEAN PROCEDURE ShiftKeyPressed BOOLEAN PROCEDURE DoTillkeyReleased p PROC PROCEDURE ShowWaitSymbol PROCEDURE HideWaitSymbol PROCEDURE SoundBell PROCEDURE Wait nrTicks LONGCARD 1 tick 1 60 second PROCEDURE InitDialogMachine PROCEDURE RunDialogMachine PROCEDURE DialogMachineIsRunning BOOLEAN PROCEDURE DialogMachineT
96. Condition StashFileName DeclExperiment NoInitialize NoInput NoOutput FROM SimMaster IMPORT RunSimMaster DeclMV NoTerminate NoAbout SimRun FROM TabFunc IMPORT TabFUNC DeclTabF Yie FROM MathLib IMPORT Sin VAR m Model x xDot temp tm ta REAL RTt TabFUNC the table function R T t A 154 ModelWorks V2 0 Appendix PROCEDURE Dynamic BEGIN temp tm ta Sin 6 2832 CurrentTime 24 0 xDot Yie RTt temp x interpolates growth rate for any temp END Dynamic PROCEDURE Objects BEGIN Dec1SV x xDot 1 0 0 0 MAX REAL Bacterial population size x g l Dec1lMV x 0 0 100 0 Population size x g l notOnFile writeInTable isY DeclMV xDot 0 0 100 0 Relative growth rate xDot g l h notOnFile writeInTable isY Dec1MV temp 0 0 100 0 Temperature temp C notOnFile writeInTable isY DeclP tm 20 0 10 0 70 0 rtc mean temperature TAE he DeclP ta 10 0 0 0 40 0 rtc amplitude of temperature fluctuations taty SIC END Objects PROCEDURE ModelDefinitions VAR tempVect grVect ARRAY 0 7 1 OF REAL BEGIN DeclM m Euler NoInitialize NoInput NoOutput Dynamic NoTerminate Objects Bacterial growth varying with temperature temp m NoAbout tempVect 0 tempVect 1 0 0 grVect 0 5 0 grVect 1 0 04 tempVect 2 10 0 grVect 2 0 07 tempVect 3 20 0 grVect 3 0 17 tempVect 4 30 0 grVect 4
97. DeclExperiment e PROC PROCEDURE CurrentSimNr INTEGER TYPE MWState noSimulation no simulation going on simulating during simulation pause current simulation temporarily halted MWSubState noRun simulating but not in SimMaster SimRun running simulating but in SimMaster SimRun noSubState state lt gt simulating PROCEDURE GetMWState VAR s MWState PROCEDURE GetMWSubState VAR ss MWSubState PROCEDURE ExperimentRunning BOOLEAN PROCEDURE ExperimentAborted BOOLEAN OPTIONAL MODULES ERFREUT EE NARA E EEE JulianDays ERHIELTEN REEE EEE CONST Jan 1 Feb 2 Mar 3 Apr 4 Mai 5 Jun 6 Jul 7 Aug 8 Sep 9 Oct 10 Nov 11 Dec 12 Sun 1 Mon 2 Tue 3 Wed 4 Thur 5 Fri 6 Sat KE PROCEDURE DateToJulDay day month year INTEGER LONGINT PROCEDURE JulDayToDate julday LONGINT VAR day month year weekday INTEGER PROCEDURE LeapYear yr INTEGER BOOLEAN PROCEDURE SetCalendarRange firstYear lastYear firstSunday INTEGER ORK RE ER AHR KAR HE S RandGen MERA KIRA KRM KAR AR A KIER EH RHF PROCEDURE SetSeeds z0 z1 z22 INTEGER defaults z0 1 z1 10000 z2 3000 PROCEDURE GetSeeds VAR z0 z1 z2 INTEGER PROCEDURE Randomize PROCEDURE ResetSeeds PROCEDURE U REAL U 0 1 cycle length gt 2 78 E13 220 years for 1000 U sec GR BR RR MI eR aiia RandNormal TYPE URandGen PROCEDURE PROCEDURE PROCEDURE REAL InstallU U URa
98. E GetWindowPlace mww MWWindow VAR x y w h INTEGER PROCEDURE SetDefltWindowPlace mww MWWindow x y w h INTEGER PROCEDURE GetDefltWindowPlace mww MWWindow VAR x y w h INTEGER PROCEDURE CloseWindow w MWWindow TYPE IOWColsDisplay RECORD descrCol identCol BOOLEAN CASE iow MWWindow OF MIOW m RECORD integMethCol BOOLEAN END RECORD VAR isOpen BOOLEAN VAR enabled BOOLEAN SVIOW sv RECORD unitCol sVInitCol BOOLEAN fw dec INTEGER END RECORD PIOW p RECORD unitCol pValCol pRtcCol BOOLEAN fw dec INTEGER END RECORD MVIOW mv RECORD unitCol scaleMinCol scaleMaxCol mVMonSetCol BOOLEAN fw dec INTEGER END RECORD END CASE END RECORD PROCEDURE SetIOWColDisplay mww MWWindow wd IOWColsDisplay PROCEDURE GetIOWColDisplay mww MWWindow VAR wd IOWColsDisplay PROCEDURE SetDefltIOWColDisplay mww MWWindow wd IOWColsDisplay PROCEDURE GetDefltIOWColDisplay mww MWWindow VAR wd IOWColsDisplay PROCEDURE DisableWindow w MWWindow PROCEDURE EnableWindow w MWWindow PROCEDURE UseCurWSettingsAsDefault PROCEDURE SuppressMonitoring PROCEDURE ResumeMonitoring PROCEDURE DeclClientMonitoring initmp mp termmp PROC PROCEDURE StashFileName sfn ARRAY OF CHAR PROCEDURE SwitchStashFile newsfn ARRAY OF CHAR PROCEDURE Message m ARRAY OF CHAR TYPE Stain coal snow ruby emerald sapphire turquoise pink
99. EDURE GetString strRef String VAR s ARRAY OF CHAR PROCEDURE PutString VAR strRef String VAR s ARRAY OF CHAR PROCEDURE DisposeString VAR strRef String PROCEDURE ConcatString VAR strRef String s ARRAY OF CHAR PROCEDURE ConcatChar VAR strRef String ch CHAR PROCEDURE Length VAR string ARRAY OF CHAR INTEGER PROCEDURE Concat VAR dest ARRAY OF CHAR source ARRAY OF CHAR PROCEDURE ConcatCh VAR dest ARRAY OF CHAR ch CHAR PROCEDURE AssignString source ARRAY OF CHAR VAR d ARRAY OF CHAR PROCEDURE ExtractSubString VAR curPosInSrcS INTEGER VAR srcS destS ARRAY OF CHAR delimiter CHAR PROCEDURE LoadString fileName ARRAY OF CHAR stringID INTEGER VAR string ARRAY OF CHAR PROCEDURE StoreString fileName ARRAY OF CHAR VAR stringID INTEGER string ARRAY OF CHAR PROCEDURE GetRString stringID INTEGER VAR str ARRAY OF CHAR PROCEDURE Concatenate first second ARRAY OF CHAR VAR result ARRAY OF CHAR IBM PC DM compatibility PROCEDURE Copy from ARRAY OF CHAR startIndex nrOfChars INTEGER VAR to ARRAY OF CHAR GE He ee eI HIE SEI AE IEE A I RK SE A IER REFEREES DMSystem RR SEN AAI SORTER HA BRR RAMI HER HMA HERR RR Bes CONST startUpLevel 1 maxLevel 5 PROCEDURE CurrentDMLevel CARDINAL PROCEDURE LevelisDMLevel 1l CARDINAL BOOLEAN PROCEDURE TopDMLevel CARDINAL PROCEDURE DoOnSubProgLevel 1l CARDINAL p PROC PROCEDURE InstallInitProc ip PROC VAR done BOOLEAN PROCEDURE ExecutelnitPro
100. Eq 4 4 resp 5 4 It is called only once during a time step but many times during a simulation run Fig T16 output calculates the output variables of sub model m Eq 4 2 resp 5 2 It is called only once during a time step but many times during a simulation run Fig T16 Note the implementation restriction that output variables must not depend directly on input variables see Manual Part II Theory chapter Model formalisms dynamic Contains the the model equations of sub model m Eq 4 1 resp 5 1 or 8a 8b and 8c In the case of a continuous time model it calculates the new derivatives from the current values of the state variables Eq 4 1 or 8a and 8c Depending on the order of the integration method this procedure is called at least once up to several times during a time step In the case of a discrete time model Eq 5 1 or 8b and 8c it calculates the new state vector directly It is only called once during a time step Fig T16 terminate called once at the end of each simulation run Fig T16 May be used freely to execute any task at the end of a run such as closing a file which has been written during the simulation run etc installModelObjects Procedure declaring all model objects i e state variables model parameters and monitorable variables of sub model m Typically this procedure contains calls to the procedures DeclSV DeclP and DecIMV It is also possible to leave the body of this procedure empty e g by us
101. File i INTEGER n CARDINAL PROCEDURE PutLongInt VAR f TextFile li LONGIN n CARDINAL PROCEDURE GetReal VAR f TextFile VAR x REAL PROCEDURE GetLongReal VAR f TextFile VAR x LONGREAL PROCEDURE PutReal VAR f TextFile x REAL n dec CARDINAL PROCEDURE PutRealSci VAR f TextFile x REAL n CARDINAL PROCEDURE PutLongReal VAR f TextFile lr LONGREAL n dec CARDINAL PROCEDURE PutLongRealSci VAR f TextFile lr LONGREAL n dec CARDINAL RRR ER FER RRM EI a a He De Fe IE HE A EE DMPrinting KKK KKK KKK KERR KEKE KEKE KK KKK TH TH TH HH TH TH A HH A HH A a a a a TYPE PrinterFont chicago newYork geneva monaco times helvetica courier symbol PROCEDURE PageSetup PROCEDURE SetHeaderText h ARRAY OF CHAR PROCEDURE SetSubHeaderText sh ARRAY OF CHAR PROCEDURE SetFooterText f ARRAY OF CHAR PROCEDURE PrintPicture PROCEDURE PrintText font PrinterFont fontSize INTEGER tabwidth INTEGER VERFEINERN EEE EAER EREE DMPTFiles KKK KKK KKK ERK KEK KKK TH TH HH TH KH TH HH A TH A HH A KH KH A a a a a VAR PTFileDone BOOLEAN PROCEDURE DumpPicture VAR f TextFile PROCEDURE LoadPicture VAR f TextFile simulDisplay BOOLEAN destRect RectArea PROCEDURE DumpText VAR f TextFile PROCEDURE LoadText VAR f TextFile simulDisplay BOOLEAN destRect RectArea fromLine LONGINT L PERE Ae ee eee He Ee ER A AE Be I SIH SEE RT FEI SERRA Tee SEI SH DMQuestions KEK KKK KKK KKK EK K HT TH HH TH TH TH HH TH
102. H 8092 Zurich Switzerland Last revision 16 Mai 88 A F Fe ke He Fe e KERR Fe 212 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 27 FROM DateAndTime IMPORT DateAndTimeRec TYPE WriteProc PROCEDURE CHAR DateFormat brief only numbers e g 31 05 88 letMonth month in letters e g 31 Mai 1988 full full in letters e g 31st Mai 1988 17 TimeFormat brief24h 24 hour format brief e g 23 15 brief24hSecs 24 hour brief amp secs e g 23 15 02 let24hSecs hour in letters e g 23h 15 02 full24hSecs full in letters e g 23 hours 15 minutes 02 seconds brief12h 24 hour format brief e g 11 15 pm 17 the following procedures write information in English only PROCEDURE WriteDate d DateAndTimeRec w WriteProc df DateFormat PROCEDURE WriteTime d DateAndTimeRec w WriteProc tf TimeFormat END WriteDatTim G 2 5 RandGen For a typical usage of this module see the research sample model Markov MOD A 149 ModelWorks V2 0 Appendix DEFINITION MODULE RandGen RRR Fe Fe Fe ERK e e He Fe Fe He Fe Fe 2 2 2 2 2 2 2 2 2 2 5 2 2 2 2 2 AAA Module RandGen Version 1 0 Copyright 1988 by Andreas Fischlin and Systems Ecology Group ETHZ Swiss Federal Institute of Technology Z rich ETHZ Version for Dialog Machine V2 0 and MacMETH V2 6 1 Pass Modula 2 implementation Purpose Bas
103. HAR length CARDINAL PROCEDURE StringToInt str ARRAY OF CHAR VAR int INTEGER VAR done BOOLEAN PROCEDURE IntToString int INTEGER VAR str ARRAY OF CHAR length CARDINAL PROCEDURE StringToLongInt str ARRAY OF CHAR VAR lint LONGINT VAR done BOOLEAN PROCEDURE LongIntToString lint LONGINT VAR str ARRAY OF CHAR length CARDINAL PROCEDURE StringToReal str ARRAY OF CHAR VAR real REAL VAR done BOOLEAN PROCEDURE StringToLongReal Str ARRAY OF CHAR VAR longReal LONGREAL VAR done BOOLEAN PROCEDURE RealToString real REAL VAR str ARRAY OF CHAR length dec CARDINAL f RealFormat PROCEDURE LongRealToString longreal LONGREAL VAR str ARRAY OF CHAR length dec CARDINAL f RealFormat PROCEDURE HexStringToBytes hstr ARRAY OF CHAR VAR x ARRAY OF BYTE VAR done BOOLEAN PROCEDURE BytesToHexString x ARRAY OF BYTE VAR hstr ARRAY OF CHAR PROCEDURE SetHexDigitsUpperCase upperC BOOLEAN PROCEDURE IllegalSyntaxDetected BOOLEAN FR ee ee Se HII EEE Hee ee FO He I He ISIE I Me IE KI DMErrorMsgs kkkkkkkkkkkkkkkkkkkkkkkkkkkk k kkkkkkkkkkkkkk TYPE ErrMsgDispProc PROCEDURE CARDINAL ARRAY OF CHAR ARRAY OF CHAR ARRAY OF CHAR PROCEDURE DispError nr CARDINAL modIdent procIdent inserts ARRAY OF CHAR PROCEDURE AssignErrMsgProducer emdp ErrMsgDispProc PROCEDURE GetErrMsgProducer VAR emdp ErrMsgDispProc PROCEDURE DfltErrMsgProducer nr CARDINAL modIdent procIdent inserts ARRAY OF CHAR
104. LER T J 1996 How To Write Fast Programs 24 LOFFLER T J FISCHLIN A LISCHKE H amp ULRICH M 1996 Benchmark Experiments on Workstations 25 FISCHLIN A LISCHKE H amp BUGMANN H 1995 The Fate of Forests In a Changing Climate Model Validation and Simulation Results From the Alps 26 LISCHKE H LOFFLER T J FISCHLIN A 1996 Calculating temperature dependence over long time periods Derivation of methods 27 LISCHKE H LOFFLER T J FISCHLIN A 1996 Calculating temperature dependence over long time periods A comparison of methods 28 LISCHKE H LOFFLER T J FISCHLIN A 1996 Aggregation of Individual Trees and Patches in Forest Succession Models Capturing Variability with Height Structured Random Dispersions 29 FISCHLIN A BUCHTER B MATILE L AMMON K HEPPERLE E LEIFELD J amp FUHRER J 2003 Bestandesaufnahme zum Thema Senken in der Schweiz Verfasst im Auftrag des BUWAL 30 KELLER D 2003 Introduction to the Dialog Machine 2 ed Price B editor of 2 ed 31 FISCHLIN A 2008 IPCC estimates for emissions from land use change notably deforestation 32 FISCHLIN A 2007 Leben im und mit dem Klimawandel Lebensgrundlagen in Gefahr 33 _ FISCHLIN A 2010 Andreas Fischlin nimmt Stellung zu Fragen rund um die Klimaproblematik unter Mitwirkung von Markus Hofmann Christian Speicher Betty Zucker Martin L ubli und J rg Fuhrer 34 FISCHLIN A amp FISCHLIN KIS
105. M from module SimBase Typically this function is to inform the simulationist about some model characteristics For instance the written information may consist of the model equations the name of the author s or some help information on the model The Model Help Info window remains open until the simulationist closes it It may be closed moved or resized freely hal Set integration method Opens an entry form in which the integration method for the currently selected model s can be set Fig R17 This is possible only for continuous time models of course the integration method for discrete time models can not be altered The following integration methods are available Integration method for the following model Continuous time submodel i Euler Heun Runge Kutta 4th order Runge Kutta 4 5th order variable step length Fig R17 Entry form to select the numerical integration method with which a continuous time sub model is to be integrated during simulations Euler Simple first order one step integration method with fixed integration step also Euler Cauchy method or Runge Kutta 1st order This method is fast but should be used with care as it can lead to large errors Heun Second order one step integration method with constant integration step also Runge Kutta 2nd order This method is similar to the trapezoidal rule but differs from it inasmuch as it is not iterative Runge Kutta 4th order F
106. MANN H 1992 Think Globally Act Locally A Small Country Case Study in Reducing Net CO Emissions by Carbon Fixation Policies 14 FISCHLIN A GYALISTRAS D ROTH O ULRICH M THONY J NEMECEK T BUGMANN H amp THOMMEN F 1994 ModelWorks 2 2 An Interactive Simulation Environment for Personal Computers and Workstations 15 FISCHLIN A BUGMANN H amp GYALISTRAS D 1992 Sensitivity of a Forest Ecosystem Model to Climate Parametrization Schemes 16 FISCHLIN A amp BUGMANN H 1993 Comparing the Behaviour of Mountainous Forest Succession Models in a Changing Climate 17 GYALISTRAS D STORCH H v FISCHLIN A BENISTON M 1994 Linking GCM Simulated Climatic Changes to Ecosystem Models Case Studies of Statistical Downscaling in the Alps 18 NEMECEK T FISCHLIN A DERRON J amp ROTH O 1993 Distance and Direction of Trivial Flights of Aphids in a Potato Field 19 PERRUCHOUD D amp FISCHLIN A 1994 The Response of the Carbon Cycle in Undisturbed Forest Ecosystems to Climate Change A Review of Plant Soil Models Out of print 20 THONY J 1994 Practical considerations on portable Modula 2 code 21 THONY J FISCHLIN A amp GYALISTRAS D 1994 Introducing RASS The RAMSES Simulation Server 22 GYALISTRAS D amp FISCHLIN A 1996 Derivation of climate change scenarios for mountainous ecosystems A GCM based method and the case study of Valais Switzerland 23 LOFF
107. OCEDURE Dynamic BEGIN grassDot cl grass c2 grass grass c3 grass aphids aphidsDot c3 c4 grass aphids c5 aphids END Dynamic PROCEDURE ModelObjects BEGIN Dec1SV grass grassDot 200 0 0 0 Grass G g dry weight m 2 DeclSV aphids aphidsDot 20 0 0 0 1000 0 Aphids A g dry weight m 2 10000 0 DeclMV grass 0 0 10000 0 Grass G g dry weight m 2 notOnFile writeInTable isY DeclMV grassDot 1000 0 1000 0 Grass derivative dG dt g dry weight m 2 time notOnFile notInTable notInGraph DeclMV aphids 0 0 1500 0 Aphids A g dry weight m 2 notOnFile writeInTable isY DeclP cl 0 4 0 0 10 0 rtc cl growth rate of grass cl day 1 DeclP c2 8 0E 5 0 0 1 0 rtc c2 self inhibition coefficient of grass c2 m 2 g dw day DeclP c3 1 5E 3 0 0 1 0 rtc c3 coupling parameter c3 m 2 g dw day DeclP c4 0 1 0 0 10 0 rtc c4 ratio of grass net use by aphids c4 DeclP c5 0 2 0 0 10 0 rtc c5 death rate of aphids c5 day 1 END ModelObjects PROCEDURE ModelDefinitions BEGIN DeclM m Heun NoInitialize NoInput NoOutput Dynamic NoTerminate ModelObjects Aphid grass model Lotka Volterra AG model NoAbout SetSimTime 0 0 100 0 END ModelDefinitions A 153 ModelWorks V2 0 Appendix BEGIN RunSimMaster ModelDefinitions END GrassAphids H 3 SAMPLE MODEL USING TABLE FUNCTIONS U
108. OLEAN GetWindowPlace returns the current position of the window mww and whether it is currently open or not Since the simulation environment remembers the location and size of a window when it was open the last time this procedure returns meaningful values even if isOpen should be FALSE PROCEDURE SetDefltWindowPlace mww MWWindow x y w h INTEGER SetDefltWindowPlace sets the default position of the window mww PROCEDURE GetDefltwindowPlace mww MWWindow VAR x y w h INTEGER VAR enabled BOOLEAN GetDefltWindowPlace returns the default position of the window mww plus the current IO window status If enabled is TRUE it means that the editing functions of the IO window are currently available to the simulationist This is the case in the state No simulation or partially in the state Pause but editing is disabled in the state Simulating s a Fig T15 in particular the title bars with horizontal lines vs dimmed bars in part II Theory To control the format in which information is displayed in a particular IO window use the following data structure TYPE IOWColsDisplay RECORD descrCol identCol BOOLEAN CASE iow MWWindow OF MIOW m RECORD integMethCol BOOLEAN END RECORD SVIOW sv RECORD unitCol sVInitCol BOOLEAN fw dec INTEGER END RECORD PIOW p RECORD unitCol pValCol pRtcCol BOOLEAN fw dec INTEGER END RECORD MVIOW mv RECORD unitCol scaleMinCol scaleMaxCol mVM
109. OR fs firstState fs firstState FLOAT lastIndiv firstIndiv 1 FOR is SUCC firstState TO lastState DO fs is fs PRED is fs is FLOAT lastIndiv firstIndiv 1 END FOR END Dynamic PROCEDURE CircaEqual x y eps REAL BOOLEAN BEGIN CircaEqual RETURN x eps lt y AND y lt xteps END CircaEqual PROCEDURE TestConsistency BOOLEAN VAR is js State sum REAL sofarOk BOOLEAN BEGIN TestConsistency sum 0 0 FOR is firstState TO lastState DO sum sum initP is END FOR sofarok CircaEqual sum 1 0 1 0E 3 FOR is firstState TO lastState DO sum 0 0 FOR js firstState TO lastState DO sum sum p is js END FOR sofarOk sofarOk AND CircaEqual sum 1 0 1 0E 3 END FOR RETURN sofarOk END TestConsistency PROCEDURE AllDead BOOLEAN BEGIN RETURN CircaEqual fs one 0 0 1 0E 3 A 160 ModelWorks V2 0 Appendix AND CircaEqual fs two fs one 0 0 1 0E 3 END AllDead VAR myMenu Menu defMarkovCmd Command nameStateOne nameStateTwo nameStateThree ARRAY 0 40 OF CHAR PROCEDURE AssignStatesNames n1 n2 n3 VAR 1 Individuals is js State istr descr ident ARRAY 0 40 OF CHAR PROCEDURE StateToString s State VAR str ARRAY OF CHAR BEGIN StateToString CASE s OF one AssignString nameStateOne str two AssignString nameStateTwo str three AssignString nameStateThree str END CASE END StateToString BEGIN
110. P cl 0 7 0 0 10 0 rtc cl growth rate of grass cl day 1 Dec1lP c2 0 001 0 0 1 0 rtc c2 self inhibition coefficient of grass c2 m 2 g dw day END Objects A 152 ModelWorks V2 0 Appendix PROCEDURE ModelDefinitions BEGIN DeclM m Euler NoInitialize NoInput NoOutput Dynamic NoTerminate Objects Logistic grass growth model LogGrowth NoAbout SetSimTime 0 0 30 0 END ModelDefinitions BEGIN RunSimMaster ModelDefinitions END Logistic H 2 THE NEW MODEL GRASSAPHIDS MOD The following program code contains the sample model described in the manual Part I Tutorial in the section Getting started with modeling especially the subsection The new model MODULE GrassAphids tn 15 6 90 Version used in ModelWorks manual FERRER REAR He He Fe Fe 2 5 2 2 2 2 2 5 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 5 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 Model of a predator prey system for aphids feeding on grass with Lotka Volterra equations FERRER ERE Fe Fe 2 2 2 2 3 2 2 2 2 2 2 22 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 FROM SimBase IMPORT DeclM IntegrationMethod DeclSV StashFiling Tabulation Graphing DeclMV DeclP RTCType Model SetSimTime NoInitialize NoInput NoOutput NoTerminate NoAbout FROM SimMaster IMPORT RunSimMaster VAR m Model grass grassDot cl c2 REAL aphids aphidsDot c3 c4 c5 REAL PR
111. PROCEDURE SetDefltSV m Model VAR s REAL defaultInit minCurInit maxCurInit REAL descriptor identifier unit ARRAY OF CHAR PROCEDURE GetDefltP m Model VAR p REAL VAR defaultVal minVal maxVal REAL VAR runTimeChange RTCType VAR descriptor identifier unit ARRAY OF CHAR PROCEDURE SetDefltP m Model VAR p REAL defaultVal minVal maxVal REAL runTimeChange RTCType descriptor identifier unit ARRAY OF CHAR R 110 ModelWorks 2 0 Reference PROCEDURE GetDefltMV m Model VAR mv REAL PROCEDURE VAR defaultScaleMin defaultScaleMax REAL VAR descriptor identifier unit ARRAY OF CHAR VAR defaultSF StashFiling VAR defaultT Tabulation VAR defaultG Graphing SetDefltMV m Model VAR mv REAL defaultScaleMin defaultScaleMax REAL descriptor identifier unit ARRAY OF CHAR defaultSF StashFiling defaultT Tabulation defaultG Graphing Note that there is no procedure available for changing the defaults of table functions To achieve a similar effect first remove it and then declare it anew 7 2 2 6 Modification of current values Most of the so called set procedures affect the corresponding current values immediately However there are some exceptions s If SetMV is called in the middle of a simulation run program state Running Fig T26 part II Theory the call will have no effect at all e SetM should not be called from within procedure dynamic All other set procedures ma
112. PlotSym ch CHAR PROCEDURE UCLineTo xUC yUC REAL RK KR KKK KK KR KKK KK KR KERR KR KR KER KK KK KEK PRR Be d EITHER ERTEILT DMWindows CONST nonexistent NIL TYPE Window WindowKind GrowOrShrinkOrDrag FixedSize FixedLocation FixedLocTitleBar ModalWindowKkind DoubleFrame SingleFrameShadowed ScrollBars WithVerticalScrollBar WithHorizontalScrollBar WithBothScrollBars WithoutScrollBars CloseAttr WithCloseBox WithoutCloseBox ZoomAttr WithZoomBox WithoutZoomBox RectArea RECORD x y w h INTEGER END WindowFrame RectArea WFFixPoint bottomLeft topLeft RestoreProc PROCEDURE Window CloseProc PROCEDURE Window VAR BOOLEAN WindowProc PROCEDURE Window VAR background Window WindowsDone BOOLEAN PROCEDURE NoBackground PROCEDURE OuterWindowFrame innerf WindowFrame wk WindowKind s ScrollBars VAR outerf RectArea PROCEDURE InnerWindowFrame outerf WindowFrame wk WindowKind s ScrollBars VAR innerf RectArea PROCEDURE CreateWindow VAR u Window wk WindowKind s ScrollBars c CloseAttr z ZoomAttr fixPoint WFFixPoint f WindowFrame title ARRAY OF CHAR Repaint RestoreProc PROCEDURE CreateModalWindow VAR u Window wk ModalWindowKind s ScrollBars f WindowFrame Repaint RestoreProc PROCEDURE UsePredefinedWindow VAR u Window fileName ARRAY OF CHAR windowID INTEGER fixPoint WFFixPoint Repaint RestoreProc PROCEDURE RedefineWindow u Windo
113. RAY OF CHAR PROCEDURE SetFileFilter f1 f2 f 3 f4 ARRAY OF CHAR PROCEDURE GetFileFilter VAR fl f2 f3 f4 ARRAY O CHAR PROCEDURE CreateNewFile VAR f TextFile prompt defaultName ARRAY OF CHAR PROCEDURE Lookup VAR f TextFile filename ARRAY OF CHAR new BOOLEAN PROCEDURE Close VAR f TextFile PROCEDURE IsOpen VAR f TextFile BOOLEAN PROCEDURE Delete VAR f TextFile PROCEDURE Rename VAR f TextFile filename ARRA OF CHAR PROCEDURE Reset VAR f TextFile PROCEDURE Rewrite VAR f TextFile PROCEDURE Append VAR f TextFile PROCEDURE FileSize VAR f TextFile LONGINT PROCEDURE EOF VAR f TextFile BOOLEAN PROCEDURE ReadByte VAR f TextFile VAR b BYTE PROCEDURE WriteByte VAR f TextFile b BYTE PROCEDURE ReadChar VAR f TextFile VAR ch CHAR PROCEDURE WriteChar VAR f TextFile ch CHAR PROCEDURE WriteEOL VAR f TextFile PROCEDURE ReadChars VAR f TextFile VAR string ARRAY OF CHAR PROCEDURE WriteChars VAR f TextFile string AR OF CHAR PROCEDURE SkipGap VAR f TextFile PROCEDURE Again VAR f TextFile PROCEDURE GetCardinal VAR f TextFile VAR c CARDINAL PROCEDURE GetLongCard VAR f TextFile VAR c LONGCARD PROCEDURE PutCardinal VAR f TextFile c CARDINAL n CARDINAL PROCEDURE PutLongCard VAR f TextFile lc LONGCARD n CARDINAL PROCEDURE GetInteger VAR f TextFile VAR i INTEGER PROCEDURE GetLongInt VAR f TextFile VAR i LONGINT PROCEDURE PutInteger VAR f Text
114. REAL This procedure allows to free the memory from the declared data to display END SimGraphUtils G 2 AUXILIARY LIBRARY G 2 1 ReadData For a typical usage of this module see the research sample model LBM DEFINITION MODULE ReadData BER Fe Fe He 2 2 2 2 22 2 2 2 2 2 5 2 2 2 2 2 2 5 2 2 2 2 2 2 2 2 2 2 2 2 2 22 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 Module ReadData Version 1 0 Copyright 1989 by Andreas Fischlin and CELTIA Swiss Federal Institute of Technology Z rich ETHZ Version for Dialog Machine V2 0 and MacMETH V2 6 1 Pass Modula 2 implementation Purpose Export of several utilities to read and test data while reading from a file with data in columnar form Programming Design A Fischlin 12 Feb 89 Implementation A Fischlin 12 Feb 89 T Nemecek 9 Sep 89 O Roth 23 Nov 89 Swiss Federal Institute of Technology Zurich Systems Ecology Group ETH Zentrum CH 8092 Zurich A 145 ModelWorks V2 0 Appendix Switzerland Last revision 23 Nov 89 or Fe ke He Fe Fe e He H He e Fe Fe He Fe Fe He He Fe Fe He 2 2 2 2 2 2 2 2 22 2 2 2 2 2 2 2 2 2 2 2 2 5 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 Import list for this module FROM ReadData IMPORT negLogDelta SkipGapOrComment ReadCharsUnlessAComment SetMissingValCode GetMissingValCode SetMissingReal GetMissingReal SetMissingInt GetMissingInt dataF OpenADataFile OpenDataFile ReRea
115. ROCEDURE CopyArea sourceArea RectArea dx dy INTEGER PROCEDURE MapArea sourceArea destArea RectArea PROCEDURE DisplayPredefinedPicture fileName ARRAY OF CHAR pictureID INTEGER f RectArea PROCEDURE StartPolygon PROCEDURE CloseAndFillPolygon pat Pattern PROCEDURE DrawAndFillPoly nPoints CARDINAL VAR x y ARRAY OF INTEGER VAR withEdge ARRAY OF BOOLEAN VAR edgeColors ARRAY OF Color isFilled BOOLEAN fillColor Color fillPattern Pattern TYPE QDVHSelect v h QDVHSelectR v h QDPoint RECORD CASE INTEGER OF 0 v h INTEGER 1 vh ARRAY QDVHSelectR OF INTEGER END END QDRect RECORD CASE INTEGER OF 0 top left bottom right INTEGER 1 topLeft botRight QDPoint END END PROCEDURE XYToQDPoint x y INTEGER VAR p QDPoint PROCEDURE RectAreaToQDRect r RectArea VAR qdr QDRect PROCEDURE SelectRestoreCopy u Window PROCEDURE SetRestoreCopy u Window rcp ADDRESS PROCEDURE Turn angle INTEGER PROCEDURE TurnTo angle INTEGER PROCEDURE Move distance CARDINAL PROCEDURE ScaleUC r RectArea xmin xmax ymin ymax REAL PROCEDURE GetUC VAR r RectArea VAR xmin xmax ymin ymax REAL PROCEDURE ConvertPointToUC x y INTEGER VAR xUC yUC REAL PROCEDURE ConvertUCToPoint xUC yUC REAL VAR x y INTEGER PROCEDURE UCFrame PROCEDURE EraseUCFrame PROCEDURE EraseUCFrameContent PROCEDURE GetUCPen VAR xUC yUC REAL PROCEDURE SetUCPen xUC yUC REAL PROCEDURE UCDot xUC yUC REAL PROCEDURE
116. SETABFUNC MOD The following program code contains a sample model demonstrating the use of the table func tion editor see also in the manual Part III Reference in the section Client interface especially the subsection Declaration of table functions MODULE UseTabFunc BER Fe Fe Fe 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 3 2 2 2 e He Fe He He He He He Fe Fe He He He He e He He Fe 2 2 0 2 2 2 2 2 2 2 2 2 2 2 2 2 2 Modelworks model UseTabFunc Copyright 1989 by Andreas Fischlin and Swiss Federal Institute of Technology Zurich ETHZ Department of Environmental Sciences Systems Ecology Group ETH Zentrum CH 8092 Zurich Switzerland Version written for Dialog Machine V2 0 MacMETH V2 6 ModelWorks V2 0 User interface 1 Pass Modula 2 implementation Modeling amp Simulation Purpose Demonstrates the use of optional ModelWorks module TabFunc References dewit C T amp Goudriaan J 1978 Simulation of ecological processes Simulation monograhs Wageningen Centre for Agricultural Publishing and Documentation ISBN 90 220 0652 2 Implementation and Revisions af 10 03 89 First implementation DM 1 0 MacMETH 2 5 1 EERE EEK E LEE EER EER ERE EERE ERLE SEEKER ER ERE ER EREEE ERE EE RERERERS FROM SimBase IMPORT Model IntegrationMethod DeclM Dec1SV RTCType DeclP StashFiling Tabulation Graphing CurrentStep CurrentTime SetDefltGlobSimPars TerminateConditionProcedure InstallTerminate
117. SLING M M 2011 Limits to Growth and Courseware World Model 2 Update and translation of Unterrichtsprogramm Weltmodell 2 by Fischlin A T Blanke D Gyalistras M Baltensweiler amp M Ulrich 1991 Systems Ecology Report No 1 Erh ltlich bei Download from www sysecol ethz ch publications reports Diese Berichte k nnen in gedruckter Form auch bei folgender Adresse zum Selbstkostenpreis bezogen werden Order any of the listed reports against printing costs and minimal handling charge from the following address SYSTEMS ECOLOGY ETH ZURICH IBZ DEPARTMENT OF ENVIRONMENTAL SCIENCES UNIVERSIT TSTR 16 CH 8092 ZURICH SWITZERLAND
118. STALLATION For the following description of the installation procedure was assumed that you have obtained a full ModelWorks software kit from the Project Centre IDA of the Swiss Federal Institute of Technology D 1 1 Preparing installation The ModelWorks software installation kit consists of 1 The GEM Desktop kit six 360K 5 1 4 in disks or three 720K 3 1 2 in disks plus two booklets GEM is the graphical user interface which provides overlapping windows mouse and menu controls dialog boxes etc You have purchased a license to install GEM on one computer 2 The Dialog Machine kit four 360K disks or two 720K disks plus documentation The Dialog Machine is a set of subroutines which allow to write programs using a mouse menus and windows etc in a machine independent way The resulting programs can be compiled and run practically without change on both Apple Macintoshes and MS DOS PCs The Dialog Machine kit also contains a licensed version 1 17 of the JPI TopSpeed Modula 2 compiler plus its documentation You have the right to install the TopSpeed compiler on one computer 3 The ModelWorks kit optional two 360K disks or one 720K disk plus document ation ModelWorks is a simulation environment that uses the dialog machine for programming As a first thing you should check the amount of free main memory and disk space on your computer At the DOS prompt type CHKDSK or use MAPMEM or Norton SI You should have at least
119. SYSTEMOKOLOGIE ETHZ SYSTEMS ECOLOGY ETHZ Bericht Report Nr 8 ModelWorks An Interactive Simulation Environment for Personal Computers and Workstations Andreas Fischlin amp Olivier Roth Dimitrios Gyalistras Markus Ulrich and Thomas Nemecek ModelWorks 2 0 Zurich Juni June 1990 Eidgen ssische Technische Hochschule Z rich ETHZ Swiss Federal Institute of Technology Zurich Departement f r Umweltnaturwissenschaften Department of Environmental Sciences Institut f r Terrestrische kologie Institute of Terrestrial Ecology The System Ecology Reports consist of preprints and technical reports Preprints are ar ticles which have been submitted to scientific journals and are hereby made available to interested readers before actual publication The technical reports allow for an exhaustive documentation of important research and development results Die Berichte der System kologie sind entweder Vorabdrucke oder technische Berichte Die Vorabdrucke sind Artikel welche bei einer wissenschaftlichen Zeitschrift zur Publi kation eingereicht worden sind zu einem m glichst fr hen Zeitpunkt sollen damit diese Arbeiten interessierten LeserInnen besser zug nglich gemacht werden Die technischen Berichte dokumentieren ersch pfend Forschungs und Entwicklungsresultate von allge meinem Interesse Adresse der Autoren Address of the authors Dr A Fischlin Dr O Roth D Gyalistras T Nemecek System kologie
120. Structuring model systems as defined in Eq 7a b or c requires a particular calculation sequence during simulation which may affect the results in a way which has to be considered by the modeler In particular it must ensure that all input values are calculated first i e at the begin of an integration step Further the results must be independent of the calculation order of the submodels This can be guaranteed given that the following conditions are observed 1 The calculation of a model is split into the following three parts a Calculation of the input variables u t for the submodel i Eq 7a b or c b Calculation of the derivatives resp the new values of the state variables of the submodel i integration Eq 8a b or c c Calculation of the output variables y t of the submodel i Eq 9a b or c 2 The calculation order is that shown in Fig T12 ModelWorks guarantees that the prerequisit under point two is always met but cannot ensure that none of the model equations are misplaced e g that a derivative is calculated in a part reserved for the calculation of inputs T 42 ModelWorks 2 0 Theory Note also that the calculation order shown in Fig T12 has a further consequence to be considered by the modeler It may affect the precision of the numerical results depending on how the equations are distributed among the continuous time submodels Differential equations coupled within a submodel are integrated differ
121. The first submodel module LBMMod describes the ecological interaction of the host plant larch Larix decidua MILLER with the herbivorous insect larch bud moth Zeiraphera diniana GN Lep Tortricidae FISCHLIN 1982 BALTENSWEILER amp FISCHLIN 1988 The second submodel module LBMObs is a parallel model mimicking the real system by using field data A master module the program module LBM combines all modules to a model definition program LBMMod LBMObs ModelWorks Dialog Machine DMFiles Fig A4 Module structure of the research sample model The next two listings show the definition and the implementation parts of the module LBMMod containing the discrete time submodel describing the relationship between the host plant and the insect DEFINITION MODULE LBMMod Purpose Simulates Larch Bud Moth population dynamics for the Upper Engadine valley from 1949 till 1977 Model b local dynamics larch larch bud moth interaction Reference Fischlin 1982 Analyse eines Wald Insekten Systemes A 164 ModelWorks V2 0 Appendix Der subalpine L rchen Arvenwald und der Graue L rchenwickler Zeiraphera diniana Gn Lep Tortricidae Diss ETHZ No 6977 Remark This program module contains the model which runs under the simulation environment ModelWorks V0 5 Programming A Fischlin Systems Ecology ETHZ Dez 1986 VAR yt REAL output simulated larval density for whole valley ytLn REAL output In of
122. The meaning of the formal procedure parameters are listed in the tables Tab R2 and R3 Note that the call of procedure SetGlobSimPars in the sub state Running will have no effect until the next simulation run s a part II Theory Fig T26 R 109 ModelWorks 2 0 Reference ModelWorks solves equations e g differential equations by using an independent variable which is normally time It also needs the independent variable if no monitorable variable has been selected as abscissa variable X isX PROCEDURE SetIndepVariIdent descr ident unit ARRAY OF CHAR SetDefltIndepVarldent overwrites the defaults of the descriptor descr identifier ident and the unit unit of the independent variable The predefined default values ModelWorks uses are the descr time ident t and no unit empty string The call of this procedure will have no effect until the next simulation run 7 2 2 INSTALLED MODELS AND MODEL OBJECTS Once declared model and model objects may be modified in any way except for their binding to a particular variable in the model definition program In order to break even this binding you have to remove the model or model object completely by calling a remove procedure see below section Removing models and model objects Modifications affect attributes and values associated with a model or model object To support model and model object editing there exists for each object class a procedure pair a get and a set procedure The
123. V2 0 V2 0 Reflex V1 1 PC and V2 0 I For differences in behavior see part III Reference ModelWorks Version 2 0 Ulrich and Swiss Federal Institute of Technology Zurich ETHZ May 1990 1989 1990 Andreas Fischlin Switzerland cLIENT INTERFACE MODULES Olivier Roth Dimitrios Gyalistras and Markus EE RR Me EH ER HR HIM HS adda SimBase RAR RRS RON I HFK ICR BRN BRK RY Declaration of models and model objects ta tak A e Sl lp SS lh a a a i fa fe Sd TYPE Model IntegrationMethod Euler Heun RungeKutta4 RungeKutta45Var stiff discreteTime RTCType rtc noRtc StashFiling writeOnFile notOnFile Tabulation writeInTable notInTable Graphing isX isY isZ notInGraph PROCEDURE Dec1M VAR m Model defaultMethod IntegrationMethod initial input output dynamic terminal PROC installModelObjects PROC descriptor identifier ARRAY OF CHAR about PROC PROCEDURE DeclSV VAR s ds REAL initial minRange maxRange REAL descriptor identifier unit ARRAY OF CHAR PROCEDURE DeclP VAR p REAL defaultVal minVal maxVal REAL runTimeChange RTCType descriptor identifier unit ARRAY OF CHAR PROCEDURE DeclMV VAR mv REAL defaultScaleMin defaultScaleMax REAL descriptor identifier unit ARRAY OF CHAR defaultSF StashFiling defaultT Tabulation defaultG Graphing PROCEDURE SelectM m Model VAR done BOOLEAN PROCEDURE Nolnitialize PROCEDURE NoInput PROCEDURE NoOutput
124. VAR ei EditItem x y length INTEGER sbd Direction minVal maxVal REAL smallStep bigStep REAL curVal REAL actionProc PROC PROCEDURE SetChar ei EditItem newCh CHAR PROCEDURE SetString ei EditItem newStr ARRAY OF CHAR PROCEDURE SetText ei EditItem VAR text ARRAY OF CHAR i i PROCEDURE SetCardinal ei EditItem newValue CARDINAL PROCEDURE SetLongCardinal ei EditItem newValue LONGCARD PROCEDURE SetInteger ei EditItem newValue INTEGER PROCEDURE SetLongInteger ei EditItem newValue LONGINT PROCEDURE SetReal ei EditItem newValue REAL PROCEDURE SetLongReal ei EditItem newValue LONGREAL PROCEDURE SetRadioButtonSet ei EditItem checkedRadioButton RadioBut PROCEDURE SetCheckBox ei EditItem boxChecked BOOLEAN PROCEDURE SetScrollBar ei EditItem newValue REAL PROCEDURE IsChar ei EditItem VAR ch CHAR BOOLEAN PROCEDURE GetString ei EditItem VAR str ARRAY OF CHAR PROCEDURE GetText ei EditItem VAR text ARRAY OF CHAR PROCEDURE IsCardinal ei EditItem VAR c CARDINAL BOOLEAN PROCEDURE IsLongCardinal ei EditItem VAR c LONGCARD BOOLEAN PROCEDURE IsInteger ei EditItem VAR i INTEGER BOOLEAN PROCEDURE IsLongInteger ei EditItem VAR i LONGINT BOOLEAN A 176 ModelWorks V2 0 Appendix PROCEDURE IsReal ei EditItem VAR r REAL BOOLEAN PROCEDURE IsLongReal ei EditItem VAR r LONGREAL BOOLEAN PROCEDURE GetRadioButtonSet ei EditItem VAR c
125. a 2 Error Error __ Module Name__ Memory Address Other 1 What circumstances led to the problem or list the last action before the error Mouse click Where __ Keyboard strike Key ___ Menu command Which 2 Have you a suggestion as to the cause of the problem 3 Describe any additional details necessary to reproduce the error Please send a copy of the program together with all necessary files in a condition which produces the error Software Description Indicate the software used when the software error occurred 1 Finder Version ____ Multifinder System Version ___ File System HFS or MFS MS DOS Version __ 2 Software used Switcher JMacMETH Version___ JPI TopSpeed Modula Version ___ Dialog Machine Version____ ModelWorks Version___ 3 Other please list name version Hardware Description Indicate the hardware used when the software error occurred Macintosh JReflex Plus SE 1 Mix llcx SE 30 Portable llci IIfx IBM PC or compatible Model configuration Documentation Error Found 1 Where is the exact location of the error 2 Describe why you believe it is an error 3 Do you have any suggestions for the correction Use back Thank you for your report and helping us to correct and improve the software you are using For internal use only Date Received Date Resolved __ Action Taken A 140 ModelWorks V2 0 Appendix Mail to
126. a multiple selection of monitorable variables has been made not only one but a series of entry forms will be offered one form for each monitorable variable This sequence can be terminated by pressing the push button Cancel Note however that in the latter case all changes which have been made to variables before will not be reversed Only eventual changes made to the monitorable variable currently in display will be discarded This function causes sooner or later a complete erase and redrawing of the content of the graph window because it always affects the R95 ModelWorks 2 0 Reference legend The actual redrawing takes place according to the current settings of the mode RedrawGraphAlwaysMode of the simulation environment E Reset scaling Resets the scaling minimum and maximum of the selected monitorable variable s to their default values K2 Set curve attributes Opens an entry form in which the attributes for the drawing of a monitorable variable s curve in the graph window can be edited Fig R25 This editing serves to set or override the automatic definition of curve attributes by ModelWorks Since no colors are available in the PC version any settings of the curve attribute stain color is without effect Curve attributes of the following monitorable variable State var y automatic definition of curve attributes or use follawing curve attributes unbroken coal O snow broken ruby sapphire C7 dashS
127. a procedure programmed by the modeler typi cally it calls several elementary simulation runs by calling the procedure SimRun from module SimMaster Since it is implemented as a client provided procedure where the modeler has anyway already full control no experiment specific initialization and termination procedures for experiments are offered by ModelWorks Structured simulations are optional and have to be installed first by the modeler via the client interface before they can be executed by the simulationist If no experiment is known to Model Works this is the same as leaving this level out completely Note also the simulationist may always bypass this level by directly starting an elementary simulation run The simulationist can execute experiments an arbitrary number of times n A structured simula tion calls elementary simulation runs k times i e structured simulations are only of some inter est if k gt 1 The total number of simulation runs then becomes k n To facilitate the orientation of the simulationist ModelWorks displays in the time window visible in the state Simulating in the upper right corner of the screen not only the current simulation time but also the number k of the current simulation run The state of ModelWorks during a structured simulation is always Simulating If an experiment is stopped by the simulationist not only the currently running elementary simulation is termi nated but also all subsequently eventual
128. ables of the currently selected model s to their default values E Reset parameters Resets all parameters of the currently selected model s to their default values E Reset stash filing Resets the monitoring settings for the stash filing F writeOnFile notOnFile of all monitorable variables of the currently selected model s to their default settings Reset tabulation Resets the monitoring settings for the tabulation T writelnTable notInTable of all monitorable variables of the selected model s to their default settings E Reset graphing Resets the monitoring settings for the graphing X Y isX isY notInGraph of all monitorable variables of the selected model s to their default settings E Reset scaling Resets the scaling minimum and maximum on ordinate of all moni torable variables of the selected model s to their default values aj Reset curve attributes Resets all curve attributes color or stain line style and symbol of all monitorable variables of the selected models to their default values 6 2 2 IO WINDOW STATE VARIABLES State variables State variable names Ident Unit Initial value Potato model Physiological age Fhys age Fhys age unit 0 000 Leaf g dw plant 0 050 Stem g dw plant 0 050 Root g dw plant 0 020 Tuber g dw plant 0 000 Fig R18 IO window State variables showing the list of all state variables and offering a palette of button functions operating on the current values of t
129. actual parameter of the procedure SimMaster RunSimMaster Any object needed from the Dialog Machine can be imported and used in the usual way one writes Dialog Machine programs This method is easy straight for ward and should cause no problems to any programmer developing models For examples see also the sample models in particular Lorenz MOD and the reasearch sample models Markov MOD and LBM MOD The other method however only available in the Macintosh versions is to call a full Dialog Machine program as a subprogram This will activate the Dialog Machine a second time result ing in a multilevel Dialog Machine program which imposes certain restrictions on the pro gramming In particular one should note that any object e g an allocated portion of the heap space does belong only to a certain program level This implies that such an object will only be able to exist as long as the level exists In particular the Dialog Machine as well as MacMETH will automatically remove objects upon leaving a program level The programmer should make sure that this does not lead to any inconsistencies e g pointer variables from a low program level pointing to a memory block no longer existing Apart from that this method should work as well as the first one although it is likely that a full Dialog Machine program installs too many menus so that the menu bar will not be able to hold all menus ModelWorks uses already most of the menu bar for its own menu
130. ad END ModelObjects VAR recordF TextFile PROCEDURE WriteOnFile ch CHAR BEGIN Write WriteChar recordF ch END WriteOnFile PROCEDURE TheExperiment CONST TAB 11C VAR is js State dt DateAndTimeRec curMin curMax REAL curFiling StashFiling curTabul Tabulation curGraphing Graphing dummyStr ident ARRAY 0 63 OF CHAR i maxRuns CARDINAL PROCEDURE AskForNrRuns BOOLEAN CONST lem 5 tab 35 VAR ef FormFrame ok BOOLEAN cl INTEGER BEGIN cl 2 WriteLabel cl lem How many runs CardField cl tab 7 maxRuns useAsDeflt 1 MAX CARDINAL INC cl ef x 0 ef y 1 display entry form in middle of screen ef lines cl 1 ef columns 55 A 162 ModelWorks V2 0 Appendix UseEntryForm ef ok RETURN ok END AskForNrRuns PROCEDURE RecordDateTime s ARRAY OF CHAR dt BEGIN RecordDateTime WriteChars recordF s WriteDate dt WriteOnFile letMonth WriteChars recordF at WriteTime dt WriteOnFile brief24hSecs WriteEOL recordF END RecordDateTime BEGIN TheExperiment maxRuns 100 IF AskForNrRuns THEN FOR is firstState TO lastState DO FOR js firstState TO lastState DO GetMV m f is js curMin curMax curFiling curTabul curGraphing SetMV m f is js curMin curMax notOnFile curTabul curGraphing END FOR GetMV m fs is curMin curMax curFiling curTabul curGraphing SetMV m fs is curMin curMax notOnFile curTabul curGraphin
131. allel model to compare measured with simulated behavior In this case you would need to declare only a monitorable but not a state variable T 26 ModelWorks 2 0 Tutorial The first parameter denotes the real variable which will be monitored Next there are two real constants the default values for the scaling of the graphics output The three strings are the same as for the state variables the name abbreviated name and the unit of the monitorable variables The next three elements are default settings for file table and graph output e g isY means that by default the variable grass will be plotted on the y axis ordinate of the graph These elements are imported with the enumeration types StashFiling Tabulation Graphing DeclP is the procedure for the declaration of model parameters Since we have two model parameters c and c3 it is called twice DeclP cl 0 7 0 0 10 0 rtc cl growth rate of grass CL day 1 Dec1P c2 0 001 0 0 1 0 rtc c2 self inhibition coefficient of grass c2 m 2 g dw day The parameter list contains first the real variable of the parameter Next there are three reals the default value of the parameter and the upper and lower limit of its range within which the simulationist may enter a new parameter value A parameter declared as rtc RTCType means that its value may be changed even in the middle of a simulation not only before or after a run The three strings are again
132. alues within the procedure Output T 33 ModelWorks 2 0 Tutorial T 34 Part II Theory This part of the ModelWorks manual contains a description and functional specification of every feature ModelWorks offers However it contains only little information on the elementary and typical usage of ModelWorks In case you should not be familiar with the basic concepts of ModelWorks please read first the ModelWorks tutorial In particular you should read the first chapter of the tutorial General Description Neither does this part of the manual contain technical information on the actual use This part of the manual explains the principles behind ModelWorks not the details on the actual implementation and version of ModelWorks Implementation dependent details are listed and explained in Part III Reference of this manual The latter has been written to support you during your daily work with ModelWorks this part is typically studied once at the begin of any serious work with ModelWorks This part of the ModelWorks manual contains two chapters The chapter Model Formalisms presents the mathematical formalisms in which ModelWorks models are to be formulated The first section of this chapter treats elementary models the second structured models which are built from several elementary coupled submodels The chapter ModelWorks Functions describes all basic functions of ModelWorks First it describes the functionality of the simulation e
133. ame place it was before its closing R 82 ModelWorks 2 0 Reference Windows Tile windows Stack windows Models State variables Model parameters Monitorable variables Clear table Clear graph Fig R9 Menu Windows Tile windows All IO windows plus the table and graph window are closed and reopened so that they do no longer overlap On small screens the IO windows for the state variables model parameters and monitorable variables are shown beside each other on top of the screen on larger screens all four IO windows are displayed in two rows on top of the screen The remaining windows are fit into the bottom portion of the screen making the graph window as large as possible The column display in the IO windows is also affected Only the short identifiers ident and the current value columns are shown current integration method for models current initial values for state variables current values for model parameters and current monitoring settings for the monitorable variables Stack windows All IO windows plus the table and graph window are closed and reopened in a stacked fashion The locations and sizes of all windows plus the columns displayed in the IO windows are the same as at the begin of a simulation session However in contrast to that situation the table and the graph window are also opened Models M The IO window Models is opened respectively brought to the front State variables S The IO windo
134. aph restoration graph printing and transfer of graph into clipboard If the mode color and vector graph saving is activated crg is TRUE each time the graph window needs to be redrawn the graph will be reconstructed in colors Restoration is necessary after some parts of it become visible again after they have been covered by another window see also description of restore or update mechanism in module DMWindows of the Dialog Machine Deactivation of this mode results in storing graphical output in a hidden bitmap without colors with a coarser resolution and more modest memory requirements Note that this mode won t affect the very first drawing of the graph i e on a color screen you may still get colored curves even if this mode should be turned off Since the full reconstruction in colors for complicated graphs may be slow especially on monochrome monitors it may be preferable to deactivate this mode trade off between colors and speed In addition to the colors all graphical output is stored as vectorized objects This allows printing and copying to the clipboard of graphs in high resolution quality but requires a corresponding amount of memory Not available in Reflex and PC version R 121 ModelWorks 2 0 Reference R 122 ModelWorks 2 0 Literature Literature The following list contains references of cited literature as well as references for further reading on the subject of modelling and simulation ATKINSON L V
135. ard if no other window than a ModelWorks window is the frontmost window Otherwise e g if a desk accessory is the frontmost window this command will perform the standard Cut command as described in the Macintosh owner s guide The quality of the transferred graph depends on the current simulation environment mode as described under menu command File Preferences 3Note that these simulation environment modes have no default values The current values are written in the resource fork of the MacMETH Shell They are read from there at the startup of your model definition program R75 ModelWorks 2 0 Reference Copy C Copies the graph into the clipboard if no other window than a ModelWorks window is the frontmost window Otherwise e g if a desk accessory is the frontmost window this command will perform the standard Copy command as described in the Macintosh owner s guide The quality of the transferred graph depends on the current simulation environment mode as described under menu command File Preferences Paste V Not available unless another window than a ModelWorks window is the frontmost window For instance if a window of a desk accessory is the frontmost window the content of the clipboard will be pasted into the desk accessory such as the scrapbook Clear B Clears the graph if no other window than a ModelWorks window is the frontmost window Otherwise e g if a desk accessory is the frontmost window this command
136. ard maxLCard LONGCARD PROCEDURE IntField line col CARDINAL fw CARDINAL VAR int INTEGER du DefltUse minInt maxInt INTEGER PROCEDURE LongIntField line col CARDINAL fw CARDINAL VAR longInt LONGINT du DefltUse minLInt maxLInt LONGINT PROCEDURE RealField line col CARDINAL fw CARDINAL VAR real REAL du DefltUse minReal maxReal REAL PROCEDURE LongRealField line col CARDINAL fw dig CARDINAL fmt RealFormat VAR longReal LONGREAL du DefltUse minLReal maxLReal LONGREAL PROCEDURE PushButton line col CARDINAL buttonText ARRAY OF CHAR buttonWidth CARDINAL pushButtonAction PROC PROCEDURE DefineRadioButtonSet VAR radioButtonVar RadioButtonID PROCEDURE RadioButton VAR radButt RadioButtonID line col CARDINAL text ARRAY OF CHAR PROCEDURE CheckBox line col CARDINAL text ARRAY OF CHAR VAR checkBoxVar BOOLEAN PROCEDURE UseEntryForm bf FormFrame VAR ok BOOLEAN d Aadal A ae I ee eI AG Bee BEI HE eA I BEI I DMFiles KR KK KEK KEK KEK KEK KEK KKK KEK KKK KKK KEK KKK KK HH HH HH A KEKE K CONST EOL 36C TYPE Response done filenotfound volnotfound cancelled unknownfile toomanyfiles diskfull memfull notdone HiddenFileInfo IOMode reading writing TextFile RECORD res Response filename ARRAY 0 255 OF CHAR path ARRAY 0 63 OF CHAR curIOMode IOMode curChar CHAR fhint HiddenFileInfo END VAR legalNum BOOLEAN PROCEDURE GetExistingFile VAR f TextFile prompt AR
137. ask PROCEDURE CallSubProg module ARRAY OF CHAR VAR status Status PROCEDURE QuitDialogMachine PROCEDURE AbortDialogMachine RRR RIO ICI ICI ICI ICICI CI ICR ICICI CIO IO k e e 2 MathLib KK KK EK KI KR KR KR KEIR KICK KER KR KK KERR KR KER KEK KR RK PROCEDURE Sqrt x REAL REAL PROCEDURE Exp x REAL REAL PROCEDURE Ln x REAL REAL PROCEDURE Sin x REAL REAL PROCEDURE Cos x REAL REAL REAL REAL PROCEDURE ArcTan x PROCEDURE Real x INTEGER REAL l Available from the following address Projekt Zentrum IDA re Dialog Machine Swiss Federal Institute of Technology ETHZ ETH Zentrum CH 8092 Ziirich Switzerland A 173 ModelWorks V2 0 Appendix PROCEDURE Entier x REAL INTEGER EEE EE EI RE REEE E E ERLERNTE DMMenus ERINNERTE NER REE EE E EL IER ER TYPE Menu Command AccessStatus enabled disabled Marking checked unchecked Separator line blank QuitProc PROCEDURE VAR BOOLEAN VAR MenusDone BOOLEAN PROCEDURE InstallAbout s ARRAY OF CHAR w h CARDINAL p PROC PROCEDURE NoDeskAccessories PROCEDURE InstallMenu VAR m Menu menuText ARRAY OF CHAR ast AccessStatus PROCEDURE InstallSubMenu inm Menu VAR subm Menu menuText ARRAY OF CHAR ast AccessStatus PROCEDURE InstallCommand m Menu VAR c Command cmdText ARRAY OF CHAR p PROC ast AccessStatus chm Marking PROCEDURE RemoveCommand m Menu cmd Command PROCEDURE InstallAliasChar m Menu c C
138. ata END LBMObs Excerpt middle portion missing from data file LBMObsUE DAT accessed by module LBMObs year y y MIN Y MAX 1949 0 018 0 006 0 041 1950 0 082 0 006 0 232 1951 0 444 0 001 1 266 1952 4 174 0 191 10 464 1953 68 797 16 667 128 490 1954 331 760 163 340 933 524 1955 126 541 25 048 317 868 1956 21 280 9 888 41 974 1957 2 246 1 330 4 538 1958 0 085 0 000 0 359 1985 0 120 N N 1986 0 690 N N 1987 2 279 0 445 4 866 1988 39 029 4 149 88 146 Legend f mean observed larval density y MIN minimum observed larval density y MAX maximum observed larval density The following module is the main program module LBM Its sole purpose is to start the simu lation environment procedure RunSimMaster and to install a menu procedure nstallMenus which gives access to the actual models The latter menu contains a command which asks the simulationist which submodel s he she wishes to load activate or to remove deactivate procedure Choose The master module imports from the modules LBMMod population model and LBMObs exports the parallel observation model the procedures ActivateLarchLBMModel and DeactivateLarchLBMModel resp ActivateLBMObsModel and DeactivateLBMObsModel These procedures will declare or remove the desired models thus allowing the simulationist to drop or load a model anytime during the simulation session MODULE LBM af 1 5 87 12 5 90 Module LBM Larch Bud Moth Purpose master module modeling the larc
139. ationStep h REAL Sets the default integration step only not the current value The call of this procedure will have no effect until the global simulation parameters are reset PROCEDURE SetSimTime t0 tend REAL Sets the defaults for the simulation start and stop time as well as the current simulation start and stop time It differs in this respect from all other parameter setting routines which affect either only the defaults or only the current values Do not call this procedure from within a model during a simulation run or an experiment since the simulation time must not be changed during a simulation run This procedure is ineffective if called in the sub state Running s a part II Theory Fig T26 7 2 1 c Modification of current values Current values of the parameters and variables listed in Tab R2 and R3 can be accomplished by the following Setxyz procedures i e procedures whose identifiers start with Ser PROCEDURE SetGlobSimPars t0 tend h er c hm REAL PROCEDURE GetGlobSimPars VAR t0 tend h er c hm REAL PROCEDURE SetProjDescrs title remark footer ARRAY OF CHAR wtitle wremark autofooter recM recSV recP recMV recG BOOLEAN PROCEDURE GetProjDescrs VAR title remark footer ARRAY OF CHAR VAR wtitle wremark autofooter recM recSV recP recMV recG BOOLEAN Above procedures set or get the current values for the global simulation parameters the project description or the recording flags
140. ault missingInt 0 PROCEDURE SetMissingInt missingInt INTEGER PROCEDURE GetMissingInt VAR missingInt INTEGER PROCEDURE SkipHeaderLine PROCEDURE ReadHeaderLine VAR labels ARRAY OF String VAR nrVars INTEGER IMPORTANT NOTE labels must be initialized to NIL before first use PROCEDURE ReadLn VAR txt ARRAY OF CHAR PROCEDURE GetChars VAR str ARRAY OF CHAR PROCEDURE GetStr VAR str String In the following procedures the two first parameters desc and loc are only needed for the display of error messages and help the user to identify an erronous location within the data file desc a string describing the kind of data to be read e g N population density or number of individuals loc a location number indicating where the error has A 146 ModelWorks V2 0 Appendix been found e g a line number PROCEDURE GetInt desc ARRAY OF CHAR loc INTEGER VAR x INTEGER min max INTEGER PROCEDURE GetReal desc ARRAY OF CHAR loc INTEGER VAR x REAL min max REAL 13 Working with data segments EOS means End Of Segment PROCEDURE SetEOSCode eosCode CHAR PROCEDURE GetEOSCode VAR eosCode CHAR PROCEDURE FindSegment segNr CARDINAL VAR found BOOLEAN PROCEDURE SkipToNextSegment VAR done BOOLEAN Testing PROCEDURE AtEOL BOOLEAN PROCEDURE AtEOS BOOLEAN PROCEDURE AtEOF BOOLEAN PROCEDURE TestEOF use only where you don t yet exp
141. ay The following type enumerates all windows of the ModelWorks simulation environment except for the window in the top right corner of the main screen used to display the current simulation time while in substate running TYPE MWWindow MIOW SVIOW PIOW MVIOW TableW GraphW AboutMW MIOW SVIOW PIOW and MVIOW designate the IO windows for the models state variables model parameters and the monitorable variables TableW GraphW and AboutMW are the table graph and the about model window with the title Model Help Info The latter window is displayed if the simulationist clicks into the question mark button of the models IO window The following procedures operate on ModelWorks windows PROCEDURE SetWindowPlace mww MWWindow x y w h INTEGER SetWindowPlace places the window mww with its lower left corner at the position x y and resizes it to the width w and height h size of outer frame including title bar frame and shadows The point x y is given in pixel coordinates with an origin at the lower left corner of the main computer screen If this procedure is called in case the window should not already be open it will open the window in the proper size at the specified location R 115 ModelWorks 2 0 Reference PROCEDURE CloseWindow w MWWindow CloseWindow closes the window w and remembers the location plus size for the next reopening PROCEDURE GetWindowPlace mww MWWindow VAR x y w h INTEGER VAR isOpen BO
142. be called any number of times but should be called for each table function only once unless it should have been removed in the meantime DeclTabFmay not be called in the sub state Running s a part Il Theory Fig T26 The meanings of its formal parameters are as follows t The variable is a variable of the opaque type TabFUNC which is exported by the module TabFunc It is used to identify and update the table function values and parameters when accessing the table function This may be for the linear interpolation procedures for reading or changing the function s current values or for removing the function XX yy The vector xx contains the independent and yy the dependent values of the table function being defined Note that the xx vector must be given sorted ascendingly otherwise the program will halt execution NValPairs Contains the number of the first elements of the vectors xx and yy which hold a valid value modifiable If this formal parameter is set to TRUE the table function may be modified by means of the table function editor from within the simulation environment s a section Optional menu Table Functions manual Part III Reference tabName Is used for the identification of the table function in the simulation environment e g for its selection in the table function editor s entry form s a section Optional menu Table Functions manual Part III Reference xName yName xUnit yUnit The names of the table function s axis
143. ble becomes negative and allows the simulationist to take some action e g changing a parameter value before resuming the simulation PROCEDURE DeclExperiment e PROC Installs an experiment which may be executed by the user by selecting the menu command Execute experiment under menu Simulation which corresponds to the call of procedure e The procedure e is provided by the modeler and contains typically calls to the procedure SimMaster SimRun If the procedure DeclExperiment has at least been called once in the course of a simulation session the menu command Execute experiment under menu Simulation will no longer appear dimmed but as active and can be chosen by the simulationist in the state No Simulation TYPE MWState noSimulation simulating pause PROCEDURE GetMWState VAR s MWState The current state of the simulation environment can be determined by calling procedure GetMWState from SimMaster The meaning of the returned value s either noSimulation simulating or pause corresponds exactly to the program states shown in Fig T15 part II Theory TYPE MWSubState noRun running noSubState PROCEDURE GetMWSubState VAR ss MWSubState The current substate of the simulation environment while a structured simulation experiment is currently in execution can be determined by calling procedure GetMWSubState from SimMaster The meaning of the returned value ss either noRun running or noSubState corresponds exactl
144. bles are numbered separately starting e g with Fig R1 respectively Tab Al 1 See the appendix for availability and installation of the Dialog Machine 2 See the appendix for availability and installation of MacMETH 6 ModelWorks 2 0 Acknowledgements The authors wish to express many thanks to Prof Dr Walter Schaufelberger not only for his substantial support but also for his unceasing encouragement which made this research and development only possible Current address Project Centre IDA or Institute of Automatic Control and Industrial Electronics Swiss Federal Institute of Technology Z rich ETHZ ETH Zentrum CH 8092 Z rich Switzerland 7 ModelWorks 2 0 Part I Tutorial This tutorial describes the elementary usage of ModelWorks i e you learn how to de velop and simulate models using ModelWorks The first chapter General Description describes the general fundamental con cepts of ModelWorks The second chapter Getting Started with the Simulation Environment contains a step by step explanation for running an existing model and getting familiar with the simulation environment of ModelWorks The third chapter Getting Started with Modeling teaches how to develop new models Having read this tutorial you will be able do develop and simulate your own simple models However if you are interested in more complex models and more advanced techniques this tutorial is not sufficient In order to l
145. ces Water Pollution and Water Control CH 8600 D bendorf Switzerland ModelWorks 2 0 Contents ABOUT MODELWORKS AND THIS TEXT ccssscccessceceseceesseceesseecesceceseecesaeesetaeeesees 5 ACKNOWLEDGEMENTS 1 sau das ein la es ass 7 Part I Tutorial T GENERAL DESCRIPTION sn en nein 10 2 GETTING STARTED WITH THE SIMULATION ENVIRONMENT 16 Dols The Sample Model este esse 16 2 2 Simulating the Sample Model a 17 2 2 1 Default simulation anne 18 2 2 2 Changing initial values asus a hir 19 2 2 3 CHANGING parameters ass re E A EEE RESE 19 224 Changing eT 20 2 2 5 Changing monitoring sese eee eee eee eee eee eee 20 2 2 6 Changing parameters during Simulation sss sees sese eee eee eee eee 22 2 2 7 Changing integration methods sss sese 23 22 8 Program termination bee 24 3 GETTING STARTED WITH MODELING sese 25 3 1 The Model Definition Program of the Sample Model 25 3 2 Developing a New Model sss 28 3 2 1 Thenewmodel 12 4 2282 ee else 28 3 2 2 Model definition program for the new model sss 29 3 2 3 Compilation of the new model sees 31 3 2 3 Simulation of henewmoldl u a 32 Part II Theory 4 MODEL FORMALISMS SET EEE ESG 36 41 Elem ntary Models unse ea ea 36 4 2 Structured Models Coupling of Submodels u 2 0002200ssssnennnennnnnn 38 5 MODEL WORKS FUNC TICING 255 cies hae ernennen 46 3 1 Simulation Environment ee 47 5 1 1 Program states of the simulation environment sese eee ee ee
146. ch bud moth system in the Upper Engadine valley in Switzerland from 1949 till the presence BALTENSWEILER amp FISCHLIN 1988 This allows to compare the observations with simulated values At the begin of the simulation session this parallel model simply reads the observations stored in the data file into an array and will assign the measured values during any simulations to a monitoring variable which the simulationist can display from within the simulation environment In case the simulationist should set the global simulation time such that it lies outside the range 1949 and 1988 the values produced by this module are no longer valid The module has been programmed such that it visualizes missing values in the graph by letting portions of the curve s disappear This is accomplished by setting the curve attribute to invisble as soon as values have become undefined yet the legend is drawn with the attributes normally used if val ues are available The next three listings show the definition and the implementation parts of the module LBMObs which reads the data from the text file LBMObsUE DAT DEFINITION MODULE LBMObs Module LBMObs Purpose Simulates the real larch bud moth system in the A 167 ModelWorks V2 0 Appendix Method Remark Upper Engadine Valley as a parallel model Observed larval densities in larvae kg larch branches as sampled from the Upper Engadine Valley are simulated by means of a ModelWorks submodel
147. changes in the actual parameters of the call to procedure DecIM As the new model requires a better integration algorithm we change the default method from Euler to Heun further we change the model name strings Dec1M m Heun NoInitialize NoInput NoOutput Dynamic NoTerminate Objects Aphid grass model Lotka Volterra GrassAphids NoAbout Change the defaults for the simulation start and stop time as follows SetSimTime 0 0 100 0 The main program needs no changes Once you have finished editing the new model save it with the command Save in the MockWrite menu and close the MockWrite window 3 2 3 COMPILATION OF THE NEW MODEL Use the MacMETH shell for the compilation and execution of ModelWorks models If you are not already in the MacMETH shell start it now the same way as you have done before when you have started the simulation environment Select the command File Compile and type the name of your just written source program GrassAphids MOD or hit the TAB key to choose the file with the ordinary Macintosh file opening dialog box Press return or click the mouse to start the compilation or click into the Open button Should you encounter any problems e g the message File not found check your installation If the compiler finds no errors in your program the compilation will end with the message Work GrassAphids OBM size else you will see the message Work GrassAphids RFM errors detected In both cases
148. complex models for the expert finally and not the least the language is available in efficient implementations on many machines as e g Apple Macintosh IBM personal computers or Sun workstations Therefore we have chosen Modula 2 as the programming language to be used for ModelWorks currently meeting all the listed requirements closest WIRTH 1988 Due to this approach ModelWorks could be designed as a fully open system which can be expanded or customized by the user to any purpose he desires 1 ModelWorks does not force the modeler to discard the simulation software together with all other investments in learning implementation and testing time or any compatibility issues when he reaches the limits of the simulation language on the contrary ModelWorks avoids the risk of having to restart with the model implementation all over again in a high level programming language since it does so from the very beginning In contrast to a simulation language a well designed general purpose high level programming language guarantees that anything which can be computed on a computer can be realized It appears that one of the reasons why so many experienced researchers almost never use simulation software but use instead general purpose high level programming languages is that they avoid the risk to have to switch techniques in the middle of a project 2 Macintosh is a registered trademark of Apple Computer Inc 3IBM is a register
149. cs PROCEDURE InstallTermProc tp PROC VAR done BOOLEAN PROCEDURE ExecuteTermProcs PROCEDURE MainScreen INTEGER PROCEDURE MenuBarHeight INTEGER PROCEDURE TitleBarHeight INTEGER PROCEDURE ScrollBarWidth INTEGER PROCEDURE ScreenWidth INTEGER PROCEDURE ScreenHeight INTEGER PROCEDURE NumberOfColors INTEGER PROCEDURE HowManyScreens INTEGER PROCEDURE GetScreenSize screen INTEGER VAR x y w h INTEGER PROCEDURE NumberOfColorsOnScreen screen INTEGER INTEGER PROCEDURE SuperScreen VAR whichScreen x y w h nrOfColors INTEGER colorPriority BOOLEAN CONST unknown 0 Mac512 1 MacIIx 457 MacIIci 9 SUN 101 IBMPC 201 MacPlus 27 MacIIcx 6 MacIIxi 10 SUN3 102 IBMAT 202 MacSE 3 MacSE30 7 SUN3 102 IBMS2 203 MacII 4 MacPortable 8 PROCEDURE ComputerSystem INTEGER CONST CPU68000 Ig CPU8088 201 CPU8186 203 CPU68010 2 CPU8086 202 CPU8286 204 CPU68020 33 CPU8386 205 CPU68030 4 CPU8486 206 CUP68040 5 PROCEDURE CPU INTEGER PROCEDURE FPU BOOLEAN CONST MacKeyboard 1 AExtendKbd 4 StandPortableISOKbd 7 MacKbdAndPad 2 ADBKeyboard 5 EastwoodISOKbd 8 MacPlusKbd 3 StandPortablekbd 6 SaratogaISOKbd 9 PROCEDURE Keyboard INTEGER CONST rom64k 1 roml128k 2 rom256k 3 rom512k 4 PROCEDURE RomVersion INTEGER PROCEDURE SystemVersion REAL FRE HEH HDI BEN HE Se IE IE IIE IR HER Ne IE HE HE AMIE A ete DMWindowIO
150. ctions are available in two ways First by the simulationist in the simulation environment of ModelWorks and second by the modeler via the client interface Tutorial Fig T1 The simulation environment allows to access the ModelWorks functions interactively by the simulationist the client interface to access them in a static predefined way i e through the writing of a model definition program Tutorial Fig T3 by the modeler Both techniques have their unique advantages and disadvantages The simulation environment of ModelWorks provides the run time environment to define simu lation experiments e g by changing parameter values to execute simulation runs and to control the monitoring of the simulation results A simulation run of ModelWorks corresponds to the numerical solution of an initial value problem of a set of ordinary differential equations or difference equations alone or in any mixed form In particular note that this means that the cur rent implementation of ModelWorks does neither provide a direct support for the solution of boundary value problems nor does it offer backward numerical integration All activities from the starting of the simulation environment till its quitting are termed a simula tion session In essence it is nothing else than the execution of the procedure RunSimMaster from module SimMaster Values or attributes of the models plus their objects and the monitor ing can be changed interactively from within the simu
151. cular states File Settings Windows Simulation Page setup Set Tile windows Start run Print graph Global simulation parameters Stack windows Halt run Pause or Preferences Project description Models en re Quit Select stash file State variables en Reset Model parameters Execute Experiment Global simulation parameters Monitorable variables Project description Table Stash File Name Clear table Windows Graph AN model s integration methods Clear Graph All model s initial values AI model s parameters All model s stash filing All model s tabulation All model s graphing All model s scaling All model s curve attributes Initialize session All above a Menubar and menu commands b Program states and transition commands of menu Simulation No simulation Menu status O window status Models State variables EMonitor variabl Parameters 5E Startrun or Stop kill run or Execute Experiment Stop time reached or Termination condition true imulating O window status Stop kill run Resume run Halt run Pause Parameters Fig T15 Menu commands a and state transition diagram b of the simulation environment of ModelWorks The simulation environment is always in one of the following three sta
152. d default values originally specified by the modeler via the client interface This allows the simulationist to re turn any time to a well defined state of all parameters and settings Further details on how to use ModelWorks during a simulation session can be found below in the section on the Simulation environment of this chapter The client interface is used to define i e to declare the models the model objects the model equations and the default values for all objects so that the run time system of ModelWorks may access and maintain them Tutorial Fig T2 It is important to note that ModelWorks does only know about those objects which have been made known to ModelWorks i e which have been explicitly declared Otherwise ModelWorks does not care what the modeler is doing with these objects nor whether they are part of a complicated structure For instance a state variable might be computed by retrieving values from a data base or might be part of a structured type such as a Modula 2 record or an array On the other hand it is also important to understand that ModelWorks will perform calculations on objects i e will assign values E g at the begin of a simulation run ModelWorks assigns automatically the initial values to all state variables or up dates the values of state variables during the simulation by assigning to them the results of nu merical integrations In order to use ModelWorks meaningfully it is necessary that the modeler obeys
153. d the Output of the discrete time submodels will have to be computed at the begin the Input plus Dynamic only at the end of the coincidence interval Time steps usually vary even if a fixed step integration method is used This is because the time steps depend not only on the integration step but also on the coincidence interval and or the monitoring interval ModelWorks is computing values exactly at any of the time points given by the current values of these global simulation parameters E g a fixed integration step of h 0 75 and a monitoring interval m 1 0 will result in the following sequence of integration step lengths 0 75 0 25 0 75 0 25 0 75 0 25 5 1 3 MONITORING ModelWorks displays simulation results only via a monitoring concept It is based on the so called monitorable variables which are declared in the model definition program Once a vari able has been declared as monitorable variable via the client interface it can be selected interac tively in the corresponding IO window in order to activate a certain kind of monitoring Any variable can be monitored as long as it is a real number In this way the simulationist may ob serve the values of any variable might it be an input state auxiliary or output variable Monitorable variable might be understood as nothing else than probes attached to any informa tion flow circulating within the model system They measure anytime anywhere any quantity without any disturbance
154. dDataFile CloseDataFile SkipHeaderLine ReadHeaderLine ReadLn GetChars GetStr GetInt GetReal SetEOSCode GetEOSCode FindSegment SkipToNextSegment AtEOL AtEOS AtEOF TestEOF Relation Compare2Strings FROM DMStrings IMPORT String FROM DMFiles IMPORT TextFile CONST negLogDelta 0 01 offset to plot log scale if values lt 0 File handling VAR dataF TextFile PROCEDURE OpenADataFile VAR fn ARRAY OF CHAR VAR ok BOOLEAN opens a file using the standard open file dialog PROCEDURE OpenDataFile VAR fn ARRAY OF CHAR VAR ok BOOLEAN opens a file specified by fn automatically and calls OpenADataFile if fn couldn t be found PROCEDURE ReReadDataFile PROCEDURE CloseDataFile Reading and number testing PROCEDURE SkipGapOrComment skips all characters lt and all text enclosed in comment brackets as used in Modula 2 i e This procedure is used in this module PROCEDURE ReadCharsUnlessAComment VAR string ARRAY OF CHAR reads a string beginning from the current position until a character lt or a comment is encountered Missing values default missingValCode N PROCEDURE SetMissingValCode missingValCode CHAR PROCEDURE GetMissingValCode VAR missingValCode CHAR default missingReal 0 0 PROCEDURE SetMissingReal missingReal REAL PROCEDURE GetMissingReal VAR missingReal REAL def
155. density reached 1200 g m T 19 ModelWorks 2 0 Tutorial 2 2 4 CHANGING SCALING As the grass curve exceeds the maximum value of 1000 ModelWorks clips these values In order to avoid this clipping and to have also a look at the clipped portions of the curve you should rescale the monitorable variable Grass You may achieve this by increasing the upper limit of interest for the grass This can be done in the window for monitorable variables Bring it to the front select Grass and click on the button el In the appearing entry form you can enter the new scaling value for the upper limit of interest type 1200 and click into the OK button ModelWorks writes the values automatically into the legend in the graph window Perform another simulation run This time the curve should be fully visible and no longer be clipped 2 2 5 CHANGING MONITORING ModelWorks uses the expression monitoring for any kind of display of simulation re sults Any variable which can be monitored is called a monitorable variable Every monitoring definition is done in the window for monitorable variables ModelWorks uses one window for numerical display tabulated and one window for the graphical display line charts of results called the table window respectively the graph window Storage of numerical results is also supported on the so called stash file for the use of the data by other programs e g a spread sheet program like Microsoft Excel or a program f
156. dly For instance the installation of an additional menu should normally only be done once at the very begin of a simulation session see also chapter Simulation environment section Simulations in the previous part II Theory and the next chapter Client interface of this part Reset All above Encompasses a reset of all reset commands listed above in particular resetting of the global simulation parameters of the project description of the stash file name of all integration methods for all models of all initial values of all parameters and of all monitoring settings Stash filing tabulation graphing See Resetting in the part Theory In addition the windows are all closed and reopened at their original position at which they were at the begin of the simulation session and the eventually installed DeclInitSimSession procedure to initialize the simulation session is called If the modeler has not changed any defaults by calling a SetDeflt procedure see module SimBase the status and the current values of all objects will be exactly the same as it was right after the start up of the model definition program 6 1 5 MENU WINDOWS This menu contains all commands which operate on windows Fig R9 These menu com mands rearrange size and position of all ModelWorks windows open the corresponding window or bring it to the front If the simulationist closes a window ModelWorks remembers its size and position and will reopen it at exactly the s
157. dure often contains also calls to procedures which set defaults for the global simulation parameters such as SetDefltGlobSimPars However it is also possible to use this procedure only for the installation of extra menus and menu commands in the Dialog Machine to expand the standard simulation environment InitDialogMachine from module DMMaster has already been called For instance the procedure may contain no calls to any model object declarations at all but the menu commands it installs allow for the activation of models since they are bound to procedures which call the model and model object declaration procedures DecIM DeclSV DeclP and DecIMV s a below section Declaration of model There may also be menu command procedures installed which remove models so that a full dynamic model loading and unloading becomes possible during a simulation session for an example see the research sample model Population dynamics of larch bud moth in the Appendix Any menu installation called within procedure md will be placed on the right side of the menu bar as it is installed by ModelWorks simulation environment After R 99 ModelWorks 2 0 Reference the execution of the md procedure RunSimMaster starts the Dialog Machine by calling procedure RunDialogMachine from module DMMaster The modeler can declare a procedure issp to ModelWorks by calling DeclInitSimSession from SimMaster ModelWorks will then call this procedure at the beginning of eac
158. e the function adds or removes all currently selected variables to or from the list reversing the current setting of the first variable in the selection Reset tabulation Resets the tabulation of the currently selected monitorable variable s to their defaults R 94 ModelWorks 2 0 Reference Note that in case that the user has made any fundamental changes to the graph such as a redefinition of scales or the activation of new curves the current version of ModelWorks redraws the graph latest at the begin of the next simulation run The actual redrawing time is determined by the current settings of the mode RedrawGraphAlwaysMode of the simulation environment x Set cancel variable as x axis in the graph X isX Sets the selected variable as independent or abscissa variable x values for the graph The function toggles actually the setting i e if the current setting is on X it is disabled otherwise enabled In the monitoring column of the IO window the monitorable variable to be used as abscissa variable is marked with an X If another variable was already selected as the abscissa variable that is automatically deselected since ModelWorks allows only one independent variable at a time In case that no monitorable variable is selected as abscissa variable ModelWorks uses the default independent variable time This function will completely erase and redraw the content of the graph window This is because it has to redraw the x axis
159. e ee ee eee eee 47 ModelWorks 2 0 5 1 2 Si SLT TTT 49 3 1 2 2 Smmulation SESSO riene ior atia E E RE A RE E 51 5 1 2 b Structured simulation Experiment sese eee eee eee eee 51 5 1 2 c Elementary simulation run eee eee eee 53 5 1 2 d Integration or time step nei el ee 53 5 1 5 S Le Te 55 5 1 4 IO windows Input Output windowS sesssesssesesseesseesseressereseresseesseessee 58 5 1 5 Predefinitions defaults current values and resetting sese eee eee ee 60 9 2 M de INN nee cells 63 5 2 1 The model development cycle use 63 5 2 2 Structure of model definition programs ea 64 9 2 3 Model Installation seesinane esanai a a ems E NETE et 64 5 2 4 Module structure of ModelWOrS sese 65 5 2 5 ModelWorks objects and the run time system sese eee eee eee 66 5 2 6 Program states of the client interface sese eee eee eee 68 5 2 7 Programming structured simulations Experiments 0 0 0 0 sese eee ee eee 68 Part III Reference 6 USERNTERF GE u 242 RR RER OEE Ta Rain lan 72 6 1 Menus and Menu Commands in eee eee 72 6 1 1 Overview over ModelWorks standard menus sss eee eee 73 6 1 2 Men 5 T ponuka a E A A E R 73 0 1 3 Diem BO in ee lee 75 eB Men SEH NOS ENTE 76 0 12 39 Menu WIRBOWS a an ee Bine ale 82 61 6 Men Simulation a Sides haa estat TONS SXT 22 reelle 84 6 1 7 Optional menu Table FUNctions sese esse eee eee 85 6 2 IO Windows Input Output WindowS sese sees eee eee 88 6 241 I1O window Models
160. e g the simulation time this might lead to inconsistencies Hence the modeler should observe some restrictions and consider the following rules e The current program state determines which procedures may be called effectively The program state during an experiment is always Simulating However ModelWorks makes a difference between the two sub states No run still executing procedure Experiment but not procedure SimRun and Running executing procedure SimRun In analogy to the state No simulation the state No run allows the modeler to program changes to current values in a similar way the simulationist might do it in state No simulation Fig T26 In the state Running the modeler may not change any of the following classes of current values Global simulation parameters procedures SetGlobSimPars SetDefaultIndep Varldent Stash file procedure SwitchStashFile minimum and maximum for the scaling and monitoring settings stash filing tabulation and graphing for monitorable variables procedure SetMV e Attempts to call procedures modifying current values of the listed kinds in the state Running will have no immediate effect In the case of the global simulation parameters they will affect the next simulation run in all other cases the attempts to change will have no effect at all e Attempts to call procedures modifying current values of other values than the listed kinds will have an effect as soon as these values are act
161. e procedures nitialize and Terminate may be called many i e n or n k times during a single simulation session The actual number depends on how many times the simulationist starts a simulation run directly or via an experiment and is not known to the modeler Whenever ModelWorks calls client procedures such as procedures nitialize or Terminate the calling sequence is the same as the declaration order in the model definition program 5 1 2 d Integration or time step The lowest level is that of an integration or time step ModelWorks supports this level by re quiring the modeler to specify for each model an Input Output and Dynamic client procedure ModelWorks will execute for every model the Output Input and Dynamic procedures during every integration step at least once Only Dynamic may be called from once up to times the order of the current integration method during a single integration step Note also that in con trast to the procedures nitialize and Terminate which are called only once per simulation run the procedures Input Output and Dynamic are called many i e i times during an elementary simulation run Fig T16 In the integration loop user commands such as pausing or stopping the simulation are pro cessed first This enables an interactive control of the simulation After that the client proce dures of the models are called Their calling sequence guarantees a correct coupling of more than one model independently
162. e that in case the simulation environment mode Always document run on stash file is currently not active the recording flag settings are irrelevant if no monitoring variable has currently the stash file setting active F Only as soon there is at least one variable the stash file will be actually opened and the recording flags then control which information is wirtten onto the stash file in SRTF stands for Rich Text Format It is based on ASCII characters only but contains coded formatting information and can be interpreted by many commercially marketed text processing applications 6Microsoft Word is available from Microsoft Corporation WriteNow has been written by Anderson D J Tschumy B amp Stinson C and is available from NeXT Inc MacWrite II is available from Claris Corp 7Cricket Graph is a program to edit and produce presentation graphics for science and business by Rafferty J amp Norling R and is available from Cricket Software Inc 8The current version of ModelWorks does not feature a post analysis session However RAMSES Research Aids for the Modeling and Simulation of Environmental Systems a software package which encompasses not only the full functionality of ModelWorks but also interactive modeling and experimental frame definition will contain also a post analysis tool capable of interpreting ModelWorks stash files R 80 ModelWorks 2 0 Reference addition to the simulation results If you plan to
163. eOne nameStateTwo nameStateThree END END DefineMarkov PROCEDURE ModelObjects VAR 1 Individuals is js State istr descr ident ARRAY 0 40 OF CHAR BEGIN Objects declaration of parameters FOR is firstState TO lastState DO ee 1 0 FLOAT ORD lastState 1 0 0 1 0 LEC eg ys END FOR Dec1P randomize 0 0 0 0 1 0 rtc Randomize option 0 don t 1 do randomize randomize 0 1 plone two 0 2 plone three 0 05 p one one 1 0 p one two p one three pltwo two 0 3 pltwo three 0 1 p two one 1 0 p two two p two three p three two 0 0 p three one 0 0 plthree three 1 0 FOR is firstState TO lastState DO FOR js firstState TO lastState DO Dec1P p is js plis js 0 0 1 0 rtc END FOR END FOR compute initial states Initialize declaration of monitorable variables FOR is firstState TO lastState DO Dec1MV fs is 0 0 1 0 notOnFile notInTable isY END FOR SetDefltCurveAttrForMV m fs one emerald unbroken SetDefltCurveAttrForMV m fs two ruby unbroken o SetDefltCurveAttrForMV m fs three coal unbroken FOR is firstState TO lastState DO FOR js firstState TO lastState DO Dec1MV f is js 0 0 1 0 writeOnFile writeInTable isY SetDefltCurveAttrForMV m f is js autoDefCol autoDefStyle OC END FOR END FOR AssignStatesNames healthy ill de
164. eans that these procedures have actually just empty bodies and are needed here only to call DeclM properly The next procedure Objects declares all model objects as explained above The next two elements are strings for the name and an abbreviated name of the model The last procedure in our case No About could be used to write information about the model in the help window of ModelWorks this window is activated by clicking on the button in the model window The procedure SetSimTime sets the default values for the simulation start and stop time Finally we come to the short body of the program module BEGIN RunSimMaster ModelDefinitions END Logistic T27 ModelWorks 2 0 Tutorial The only action performed by this program is to call the procedure RunSimMaster This starts the ModelWorks simulation environment and passes the program control to ModelWorks Its parameter the procedure ModelDefinitions contains the complete definition of the sample model Note how the procedures are nested First ModelWorks will activate the simulation environment and call the procedure ModelDefinitions Later on it will call the procedure Objects which will result again in calls to the procedures DeclSV DeclMV and DeclP This mechanism ensures that it is clear which objects belong to which model Note also that while declaring an object this object will also be immediately initialized with the given values E g returning from procedure Dec
165. ear must be an year following immediately a leap year The day of the first Sunday in January in the first year firstSunday must be specified otherwise weekdays won t be computed correctly If faulty values are specified this routine will lead to an error condition The default range is firstYear 1949 lastYear 5282 firstSunday 2 since the 2nd January 1949 is a Sunday Other possibilities Sunday 6 Jan 1805 Note that calling this procedure may be useful in order to use Julian days of type INTEGER instead of LONGINT Then the calendar routines can cover fully 137 years without causing an overflow when assigning the LONGINT result of procedure DateToJulDay to an INTEGER variable END JulianDays G 2 3 DateAndTime The module DateAndTime is only present in the Macintosh versions It allows to access the clock which is built into any Macintosh computer In the context of simulations it is typically used in conjunction with module WriteDatTim see below to record data and time at the begin and end of a long several hours lasting structured simulation see sample model Markov MOD for such a use Hence it is only available for the ModelWorks Macintosh versions ModelWorks uses this module too and the compiled implementation DateAndTime OBM is linked into file SimMaster OBM DEFINITION MODULE DateAndTime A F 3 7 89 CONST Jan 1 Feb 2 Mar 3 Apr 4 Mai 5 Jun 6 Jul 7 Aug 8 Sep 9 Oct 10
166. earn the more sophisticated fea tures of ModelWorks you should read part II ModelWorks Theory and the second chap ter Client interface of part III Reference They contain a full and complete description of all possibilities ModelWorks offers This tutorial is best read while having access to a computer and the described steps are actually executed This requires that the reader is already familiar with his her compu ter and the usage of its software in particular the choosing of menu commands clicking on objects i e object selection and the dragging of objects e g moving the scroll box in a scroll bar Moreover it is assumed that the user knows how to operate a simple programming editor e g the desk accessory MockWrite has a basic knowledge of the programming language Pascal or Modula 2 and is familiar with the mathematics in volved with modeling and simulation of differential and difference equation systems No particular information is provided on these topics Please refer to other texts if you should have any difficulties with any of these subjects The appendix contains infor mation on how to proceed in order to install the ModelWorks software Reading Hint For easier orientation the pages figures and tables of Part I Tutorial are prefixed with the letter T Note that the following text assumes that you will work with the original ModelWorks version as available on the Macintosh computer If you have no access to a
167. ect EOF shows alert TYPE Relation smaller equal greater PROCEDURE Compare2Strings a b ARRAY OF CHAR Relation END ReadData G 2 2 JulianDays DEFINITION MODULE JulianDays BERR 2 5 2 2 2 22 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 5 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 Module JulianDays Version 1 0 Copyright 1989 by Andreas Fischlin and Swiss Federal Institute of Technology Z rich ETHZ Version written for Dialog Machine V2 0 User interface MacMETH V2 6 1 Pass Modula 2 implementation ModelWorks V2 0 Modeling amp Simulation Purpose Translates back and forth dates into a number of days Julian days in order to allow the computing with dates Remark This implementation is based on the Gregorian calendar which is valid after 15 0ct 1582 Note that this date followed immediately after 4 0ct 1582 to correct for accumulated errors in the Julian calendar introduced by Julius C sar ab urbe condiata the foundation of Rome i e 753 BC Gregorian calendar correction by Pope Gregor XIII The Gregorian calendar will need no corrections for 3333 years Note there is also the so called Julian Period which is used in astronomy as proposed by Joseph Justus Scaliger 1581 First Julian Date J D is middle noon 1 Jan 4713 BC The Julian time is calculated in days and is a real defining hours minutes plus seconds Note that in this method a day starts at no
168. ected models and or their objects by clicking on the buttons in the window Ident Integration method Log model Euler Fig R15 IO window Models showing the list of all sub models and offering a palette of button functions operating on models and model objects SET Columns set up Activates an entry form in which the display of the columns in the IO window Models can be controlled Fig R16 The columns of which the display can be turned on or off are Model names Full names of the models Ident Short identifiers of the models Integration method Current integration method Check the columns to be displayed in the window EJ Model names EJ Ident EJ Integration method Fig R16 Entry form opened by the button in the IO window Models Selects all models All subsequent button functions will operate on the scope All Fig T21 part II Theory i e on all models respectively all objects of all models R 88 ModelWorks 2 0 Reference Help respectively model information Opens or brings a window with the title Model Help Info to the front and executes the help or about procedure for the currently selected model works only on single model and not on multiple selections ModelWorks opens only the window but writes nothing into it It is up to the procedure Abour to write information into this window The latter procedure has been installed by the modeler while declaring the model see procedure Decl
169. ed Programming Programming and Implementation O Roth 12 09 89 Implementation O Roth 12 09 89 Swiss Federal Institute of Technology Zurich ETHZ Department of Environmental Sciences Systems Ecology Group ETH Zentrum CH 8092 Zurich Switzerland Last revision of definition 12 09 89 or Fe He Fe He Fe e He Fe He 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 FROM SimBase IMPORT Model Stain LineStyle FROM DMwindowIO IMPORT Color TYPE Curve VAR nonexistent Curve read only LX Procedure to access the ModelWorks Graph WINDOW PROCEDURE SelectForOutputGraph This procedures brings the ModelWorks Graph window to front and makes it the current output window This allows subsequently calls to almost all of the I O procedures of the Dialog Machine module DMwindowIO LX ES EEE he Procedures to access the GRAPH in the Graph window m PROCEDURE DefineCurve VAR c Curve col Stain style LineStyle sym CHAR Every curve has it own plotting style and color This allows for the simultaneous drawing of an arbitary number of curves within the ModelWorks graph sym specifies a character which is drawn repeatedly at the data points they help identifying a curve sym 0C no mark is plotted Use this procedure a
170. ed are already imported from the modules SimMaster respectively SimBase 1 On the IBM PC an editor is part of the JPI TopSpeed Modula 2 development environment 2 Please note that shareware does not mean it is free but you owe the author a payment in case you use the software on a regular basis 3 On the IBM PC use the editor of the JPI TopSpeed Modula 2 development environment 4 On the IBM PC use the command write to under the Editor Menu of the JPI TopSpeed Modula 2 development environment 5 On the IBM PC name the file e g GRASSAPH MOD T 29 ModelWorks 2 0 Tutorial Next declare the new state variable aphids its derivative aphidsDot and the three new parameters c3 c4 and c5 VAR m Model grass grassDot cl c2 REAL aphids aphidsDot c3 c4 c5 REAL Change the procedure Dynamic by adding the consumption term into the first statement plus inserting a second statement corresponding to the second differential equation PROCEDURE Dynamic BEGIN grassDot cl grass c2 grass grass c3 grass aphids aphidsDot c3 c4 grass aphids c5 aphids END Dynamic Now edit the procedure Objects Since several default values will be different from the ones of the old model first change the parameters of the declaration procedures already present The behavior of the state variable grass is different It needs another initial value DeclSV grass grassDot 200 0 0 0 10000 0 Grass G g dry weight m
171. ed trademark of International Business Machines Corporation 4 Sun isa registered trademark of Sun Microsystems Inc 5 See the appendix for cited literature ModelWorks 2 0 ModelWorks consists of a set of library modules written in Modula 2 which contain the program parts common to any simulation such as numerical integration algorithms and the tabular plus graphical display of the simulation results or the interactive chan ging of model or other simulation parameters The variable portion the model of inte rest is to be supplied by the user in the form of a standard Modula 2 program It descri bes the model s behavior and installs the model in the simulation environment by means of the elements provided by the so called client interface of ModelWorks Modeling and simulating with ModelWorks includes therefore three steps a Writing a Modula 2 program the model definition b compilation and c execution of the program run ning simulation experiments with the model within the simulation environment Interactive modeling and interactive simulations are supported in ModelWorks in several ways The user interface of ModelWorks in the simulation environment allows to change interactively all settings including any simulation parameters such as the integration method or the step length model parameters and or initial values of the state variables plus selection of the display of simulation results Simulation results are made visible t
172. ed window position see menu command Tile windows and they will display only this identifier to denote the model parameter p Example Ks unit Unit in which to measure values of the model parameter p This string is displayed in IO windows during a simulation session Example ug l Thereafter the value of the parameter p can be changed within the range minRange maxRange and be reset to its default value defaultVal A parameter change in the middle of a simulation run can lead to data inconsistencies It can selectively be allowed or prevented with the parameter runTimeChange of the type TYPE RTC Type rtc noRtc Normally only time invariant parameters are declared as model parameters Time variant parameters are often better treated as input variables or if the simulationist wishes to edit their values interactively during a simulation session it is advisable to use the series of values as a table function depending on time and the parameter is treated as an auxiliary variable see section on Declaration of table functions 7 1 5 DECLARATION OF MONITORABLE VARIABLES Every real variable may be declared as a monitorable variable This allows the simulationist to monitor or observe its values from within the simulation environment There apply no restrictions nor does the monitoring exert any influence on the variables monitored Simply call procedure DeclIMV PROCEDURE Dec1lMV VAR mv REAL defaultScaleMin defaultScaleMax RE
173. efltCurveAttributesForMV ModelWorks adopts the so called automatic definition of curve attributes It has been designed so that curves can be optimally told apart on black and white as well as color devices such as monochrome or color screens on laser printers or on color ribbon matrix printers on slide recorders plotters etc However this has the disadvantage that for a particular monitorable variable the curve attributes may change too often i e as soon as the automatic curve attribute of another previously activated monitorable variable is changed To avoid the latter the user has to override the auto matic definition Note that the curve attributes assigned by the automatic definition are prede fined by ModelWorks only and can not be changed by the user ModelWorks uses the values listed in Tab T1 Attributes are distributed according to the position 7 in the sequence in which the monitorable variables have been activated for graphical monitoring 5 2 Modeling 5 2 1 THE MODEL DEVELOPMENT CYCLE Starting with a mathematical model given in form of the Equ 4 or 5 respectively 6 7 to 10 the modeler or client has to write first the so called model definition program It solves numerically the initial value problem of the system of differential and or difference equations This corresponds to a translation process of the initial value problem of the mathematical model to a simulation model The latter may also be termed a numerical
174. elWorks will call it also during a full reset hence if programmed accordingly this would allow to resume initial start up conditions even if defaults should have been modified Tol ModelWorks 2 0 Theory Symbol Meaning of variable Predefined default Global simulation parameters Start time for simulation Stop time for simulation Fixed integration step or maximum integration step for continuous time sub models Maximum relative local integration error Discrete time step for discrete time sub models or coincidence interval for mixed time structured models 1 Monitoring interval 0 25 Descriptor identifier and unit for independent variable time tho Project description Project title string Use project title string in graph TRUE Remarks string PS Use remarks string in graph TRUE Footer string dd mon yyyy hh mm Run 1 7 Automatic update of date time and run in footer TRUE Recording of data on models in stash file TRUE Recording of data on state variables in stash file FALSE Recording of data on model parameters in stash file FALSE Recording of data on monitorable variables in stash file TRUE Recording of graph in stash file FALSE Stash filing ae Stash file name ModelWorks DAT 8 Automatic definition of curve attributes Predefined value colors and line styles i MOD 4 coal unbroken ruby broken emerald dashSpotted turquoise spotted Tab T1 Predefined defaults and values Unless overwritten by the
175. eler in the model definition program If a multiple selection of model parameters has been made not only one but a series of entry forms will be offered one form for each model parameter This sequence can be terminated by pressing the push button Cancel Note however that in the latter case all changes which have been made to model parameters before can no longer be reversed Only eventual changes made to the model parameter currently in display will be discarded E Reset model parameter Resets the parameter values of the currently selected model parameter s to their default values 6 2 4 IO WINDOW MONITORABLE VARIABLES The IO window Monitorable variables displays information name identifier unit minimal and maximal value for scaling current output selection about the installed monitorable variables of all models Fig R22 Furthermore it offers a mechanism to select monitorable variables and to execute functions which operate on the current monitoring settings of the selected monitorable variables by clicking on the buttons in the window Monitorable variables Honitorable variable names Ident Unit Monitoring Logistic grass growth model Grass G g dry weight m 2 T Y Grass derivative dG dt g dry weight m 2 a a Lo Fig R22 IO window Monitorable variables showing the list of all monitorable variables and offering a palette of button functions operating on the current monitoring settings of the currently selected
176. elop your own model definition programs without any limits However if you should wish to obtain the full MacMETH Fast Modula 2 Language System together with its library and a manual you have to purchase a separate license for MacMETH Furthermore if you should wish to obtain the original Dialog Machine software consisting of a fully described interface to its library an optional auxiliary library sample programs and texts explaining concepts and typical usage you have to order a separate Dialog Machine distribution s a appendix section How to work with the Dialog Machine Before you continue please lock the distribution diskettes and make first a working copy Then store the distribution diskettes in a safe place Order a full ModelWorks Development Kit V1 1 PC from the following address Projekt Zentrum IDA re ModelWorks Swiss Federal Institute of Technology ETHZ ETH Zentrum CH 8092 Z rich Switzerland 2 Order MacMETH from the following address Institut f r Informatik ETHZ Swiss Federal Institute of Technology ETH Zentrum CH 8092 Z rich Switzerland 3 Order the Dialog Machine from the following address Projekt Zentrum IDA re Dialog Machine Swiss Federal Institute of Technology ETHZ ETH Zentrum CH 8092 Z rich Switzerland A 127 ModelWorks V2 0 Appendix Modelllorks 172 T258 in disk ZE available S M RMSMacMETH 2 ov Ea Bu User Profile M2Tools R FAMSESLib ModelWorks 2 2 4 items SOSE
177. em Accad Nazionale Lincei ser 6 2 31 113 WIRTH N 1988 Programming in Modula 2 Springer Berlin a o 4th corrected edition WIRTH N GUTKNECHT J HEIZ W SCH R H SEILER H amp VETTERLI C 1988 MacMETH A fast Modula 2 language system for the Apple Macintosh User Manual 2nd ed Institut fiir Informatik ETH Ziirich Switzerland 100pp WYMORE A W 1984 Theory of Systems in VICK C R RAMAMOORTHY C V EDS Handbook of Software Engineering Van Nostrand Reinhold Company New York 1984 ZEIGLER B P 1976 Theory of Modelling and Simulation John Wiley amp Sons ZEIGLER B P 1984 System Theoretic Foundations of Modelling and Simulation in OREN T I ZEIGLER B P ELZAS M S EDS Simulation and Model Based Methodologies An Integrative View Springer Verlag R 124 Appendix Reading Hint For easier orientation the titles pages figures and tables of this Appendix are prefixed with the letter A Within this part figures are numbered separately starting with Fig Al A ModelWorks Version and Implementations The ModelWorks version described in this text is version V2 0 made in May 1990 There exist in fact four slightly differing implementations or versions 1 The standard Macintosh version V2 0 Runs on all Macintosh computers with at least 1 MBytes of memory and offers all ModelWorks functions without restrictions 2 The Reflex Macintosh version V2 0 Reflex It is a reduced subset from t
178. eneral Description ModelWorks is an interactive modeling and simulation environment to study the behavior of dynamic models which are described by differential or difference equations Any system described by a set of coupled ordinary differential or difference equations can be modeled using ModelWorks Since ModelWorks features modular modeling it is also possible to mix models of different types and even simultaneously integrate them with different integration methods ModelWorks has two interfaces to communicate with the human user the user interface of the simulation environment for the simulationist and the client interface for the modeler who builds models Fig T1 Modeler develops moc Client Interface ModelWorks Simulationist uses existing MO User Interface Fig T1 The two interfaces of ModelWorks The modeler uses the client interface for the model development the simulationist uses the user interface of ModelWorks simulation environment to perform simulation experiments with an already existing model Typically the modeler and the simulationist are one and the same person changing just roles Typically the modeler and the simulationist are one and the same person However their roles are distinct and should be clearly separated The modeler defines all properties of a simulation model i e he specifies a model definition This includes the specification of the model s mathematical properties and its objects s
179. ent 114 model 27 64 100 model object 26 31 66 102 model parameter 25 27 103 monitorable variable 25 26 104 state variable 25 26 102 table function 105 derivative of state variable 11 12 25 38 54 66 101 102 Dialog Machine 15 47 65 99 127 137 139 difference equation 10 11 36 101 differential equation 10 11 23 36 101 discrete time 11 37 38 39 40 100 155 experiment see simulation structured 183 GEM desktop 127 136 137 global simulation parameter 60 63 77 107 graph 32 56 57 73 84 95 98 105 118 120 hardware requirements 6 126 hierarchical system 38 42 independent variable 39 55 57 94 95 107 109 initial value 11 12 13 36 37 46 66 81 91 101 103 input 36 37 42 53 54 101 integration method 23 44 46 53 89 100 integration step 23 43 46 53 54 IO window 17 58 83 88 115 117 JPI TopSpeed Modula 2 136 138 MacMETH 6 17 31 127 129 132 mixed model 12 38 39 40 155 model 11 25 36 38 46 88 model definition program 10 13 14 25 28 63 64 65 98 99 130 model development 10 29 63 64 model object 11 46 58 66 model parameter 11 12 37 46 66 91 modeler 10 modifying current value curve attribute 21 96 global simulation parameter 69 77 108 109 initial value 19 91 111 integration method 23 89 111 integration step 23 77 model 111 model parameter 19 91 93 111 run time change 22 27 48 84
180. ent modes a een 120 IETTERA TURE eau raa a r a A Ec a a a a 123 Appendix A ModelWorks Version and Implementations sss sees eee eee eee eee eee 125 B Hard and Software Requirements uueessesssssesssssessnnnessnnnnsnnnnnnnnnenennnannnnnannnnn 126 B 1Ma intosh Versions scinsioni anian a 126 BZ IBM RC TTT 126 C How to Work With ModelWorks on Macintosh Computers sese esse eee 127 C 1 Installation of ModelWorks a na ee 127 C 2 How to Work With MacMETH ne a au 129 Co Conteris Model OLKS TT 132 C 4 How to Make a Stand alone Application sss esse eee eee 134 D How to Work With ModelWorks on IBM PCS sss 136 ISN aT 136 D 1 1 Preparing installation een nee 136 D 1 2 Installation of GEM Desktop u Ha ee 137 D 1 3 Installation of the Dialog Machine sss eee 137 D 1 4 Installation of ModelWorks 2220022200022s0nsnnnensnnnnesnnenennnennnnnn 138 D 2 How to Develop Models 35 5 4 sss sees 138 E How to Work With the Dialog Machine sss 139 F B g Report Form na nl 139 G Definition Modules un augen 141 Gil Optional Client Interface nn as 141 Gl 1 PGE UIC ass ara e a n ne 141 AD Simne orate cenena e nenne E ARES 141 7 1 3 SEU HADI nils ana aege aae a eigen 142 E Auxiliary Library anni rE RE E E EN R ES 145 E L IT E T 145 DD JulanDays nnahus means 147 7 2 3 DoleAndh mern nunmal 148 G 2 4 WriteDat Tim unse anne une 149 E RONG OR sera eier 149 G 2 6 RandNormal au ehilehendesshs 151 ModelWorks 2 0 F
181. ented by means of the Dialog Machine For instance under graduate students at the ETHZ have been able to work successfully with ModelWorks model definition programs within a learning time of only a few minutes The Dialog Machine has been designed by Andreas Fischlin implemented by Andreas Fischlin Klara Vancso and Alex Itten during the pilot project CELTIA under the auspices of Walter Schaufelberger from the Swiss Federal Institute of Technology ETHZ Ziirich Switzerland T 15 ModelWorks 2 0 Tutorial 2 Getting Started with the Simulation Environment When you read this chapter and follow the instructions given you learn step by step how to run simulation experiments with ModelWorks In particular you learn how to produce behavior trajectories of a sample model and how to change a model s initial and parameter values using the ModelWorks simulation environment It is assumed that you know how to operate the computer you are using its operating system and typical application software and that you have ModelWorks installed and are ready in order to actually perform the described procedures on your computer while reading this chapter 2 1 The Sample Model The sample model is a simple growth model for grass It models in a crude way the growth of real grass by assuming logistic growth In the first phase the plants grow exponentially under optimal conditions Within a given constant time interval doubling time the density
182. ently from those coupled via submodel boundaries when using higher order integration methods This fact should be considered when subdividing a model into several submodels unless the simulationist should restrict himself to single step integration methods only s a the following example and Fig T14 t lt t lt t h Initial Output of all Input of all Integration of conditions K subsystems a subsystems ial subs el t t h Fig TI2 Calculation order applied by ModelWorks during integration The larger loop corresponds to a single time step h current integration step the inner loop is used only by integration methods with order gt 1 e g Heun Runge Kutta 4th order Fig T12 shows how coupling within a single submodel i e formulated within the equation section dynamic is defined at every point in time whereas the coupling between submodels i e formulated within the equation sections output respectively input takes place only at the end points of an integration step Note also that this phenomenon is different from the coupling between continuous and discrete time submodels where the coupling is usually happening even more rare i e only at the coincidence points They are mostly much further apart than the current size of the integration step h Both kinds of coupling the one at the end points of the discretisation interval h as well as the one at the end points of the coincidence interval c are of the same type i e Mode
183. equations However the smaller the integration step the smaller becomes this effect E g in order to make the effect clearly visible case ii of Fig T14 has been computed with a rather large integration step of h 0 15 T 44 0 0 Curves xia sas Hag eee x 16 Fig T14 Simulation results of two mathematically equivalent model variants a and b as given by Eq 13 respectively Eq 14 15 Results obtained using i the first order Euler ii the second order integration method Heun with steplength h 0 15 Although the two model variants s a Fig T13 ought to behave identically their two implementation variants a and b yield the same results only in case i but differing ones in case ii This is a consequence of the calculation order within a simulation step Fig T12 and the order of the integration method In case ii the information exchange between submodels is not so often done for variant b than for variant a because it takes only place at the begin of not during an integration step 2 0 10 0 15 0 Minimum prey variant a 0 000 predator variant a 0 000 prey variant b 0 000 predator variant b 0 000 ModelWorks 2 0 Theory 20 0 25 0 mer 20 0 25 0 Max imum 27000 000 1200 000 27000 000 1200 000 Xia X24 State variables of variant a Xin Xp of variant b 30 0 time 30 0 time Unit HEHE T45 ModelWorks 2 0 Theory 5 ModelWorks Functions ModelWorks fun
184. ermined by the modeler via the client interface by calling the procedure GetMWSrate Transitions from one state to another are accomplished in several ways Either the simulationist decides to issue a menu command available under the menu Simulation or ModelWorks leaves the state Simulating due to one of the following two reasons The simulation time has reached the stop time or the terminate condition provided by the modeler returns true Note that this implies that state transitions are under the control of either the simulationist or of the modeler and or are automatically controlled by the ModelWorks run time system In case of conflict the commands of the simulationist have the highest priority followed by the conditions programmed by the modeler It is possible to quit the simulation environment from within every ModelWorks state menu command Quit 5 1 2 SIMULATIONS After starting an existing model definition program i e execution of the procedure RunSimMaster the ModelWorks simulation environment appears on the screen with its menubar and the four IO windows one for each kind of model objects Tutorial Fig T4 They display the information current settings and values of all model s and all model s objects Any model object is recognized by ModelWorks only if it has been declared during execution of the corresponding declaration procedure The latter is given by the actual parameter used while calling procedure RunSimMaster Choosing t
185. erw und korr Aufl 1993 Unterrichtsprogramm Weltmodell2 2 FISCHLIN A amp ULRICH M 1990 Unterrichtsprogramm Stabilitat 3 FISCHLIN A amp ULRICH M 1990 Unterrichtsprogramm Drosophila 4 ROTH O 1990 Maisreife das Konzept der physiologischen Zeit 5 FISCHLIN A ROTH O BLANKE T BUGMANN H GYALISTRAS D amp THOMMEN F 1990 Fallstudie interdisziplin re Modellierung eines terrestrischen Okosystems unter Einfluss des Treibhauseffektes 6 FISCHLIN A 1990 On Daisyworlds The Reconstruction of a Model on the Gaia Hypothesis 7 i GYALISTRAS D 1990 Implementing a One Dimensional Energy Balance Climatic Model on a Microcomputer out of print 8 FISCHLIN A amp ROTH O GYALISTRAS D ULRICH M UND NEMECEK T 1990 ModelWorks An Interactive Simulation Environment for Personal Computers and Workstations for new edition see title 14 9 FISCHLIN A 1990 Interactive Modeling and Simulation of Environmental Systems on Workstations 10 ROTH O DERRON J FISCHLIN A NEMECEK T amp ULRICH M 1992 Implementation and Parameter Adaptation of a Potato Crop Simulation Model Combined with a Soil Water Subsystem 1 NEMECEK T FISCHLIN A ROTH O amp DERRON J 1993 Quantifying Behaviour Sequences of Winged Aphids on Potato Plants for Virus Epidemic Models 12 FISCHLIN A 1991 Modellierung und Computersimulationen in den Umweltnaturwissenschaften 13 _ FISCHLIN A amp BUG
186. es of the auxiliary library distributed with the current version of ModelWorks have been included in the release since they are often used in the context of modeling and simulation Besides there is also the possibility to use any object from the Dialog Machine The majority of the latter is contained in the kernel of ModelWorks and resides together with any model definition program already in the memory Hence this part of the Dialog Machine can be used by the modeler without any penalty Those modules of the Dialog Machine which are not in use by ModelWorks can then be considered as a particular extension of the auxiliary library Fig T25 in part II Theory 7 1 Declaring Models and Model Objects This section describes the core of the ModelWorks client interface i e those procedures and types which are used by every model definition program 7 1 1 RUNNING A SIMULATION SESSION The simulation session is started whenever a program calls the procedure RunSimMaster from the client interface module SimMaster The simulation session is terminated upon returning from the procedure RunSimMaster PROCEDURE RunSimMaster md PROC This is the only statement which the model definition program must execute in order to use ModelWorks The argument md refers to a procedure which typically declares the models including all their model objects by calling the procedure DeclM from module SimBase s a below section Declaration of model The proce
187. f model equations m within the interval from till of the independent variable simulation time with the current integration method associated with model m The integration will be performed for every state variable belonging to model m As initial values ModelWorks will use the current initial values associated with the declared state variables Either stopping the simulation permanently kill or encountering the termination condition will stop the integration END SimIntegrate A 141 ModelWorks V2 0 Appendix G 1 3 SimGraphUtils For a typical usage of this module see the sample model Lorenz MOD DEFINITION MODULE SimGraphUtils 2 F F Fe Fe Fe ke Fe Fe 2 2 2 5 2 2 2 2 2 2 2 2 2 2 2 2 2 2 5 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 5 2 2 2 2 2 2 2 2 2 2 2 2 2 200 Module SimGraphUtils MW_V2 0 Copyright 1989 by Olivier Roth and Swiss Federal Institute of Technology Ziirich ETHZ Version written for Dialog Machine DM V2 0 User interface MacMETH_V2 6 1 Pass Modula 2 implementation Modelworks MW_V2 0 Modeling amp Simulation Purpose provides some utilities to make I O to the graph window and the graph of the modeling and simulation environment ModelWorks Remarks most procedures behave similar to those of the module DM2DGraphs and may now be combined with many procedures from DMWindowIO The window and its associated graph are objects of the ModelWorks environment and should therefore not be remov
188. file if the path of its location is also contained in the User Profile You don t have to remove the inserted error messages your self the next call to Merge removes any messages no longer valid and all messages are removed after an error free compilation There are three methods to tell the compiler where to find the input file containing the source code Type the name and hit return or click the mouse Note that this method may require that you type not only the file name but also its path preceding the name E g the source code is stored in a file MyFile MOD within the folder Subfolder1 which again is within the folder Work which is in the same folder as the MacMETH shell in this case you would have to enter Work Subfolder1 MyFile MOD unless your profile contains already the path Work Subfolder1 Press the TAB key instead of typing a file name This method will let you choose the file containing the source code with the ordinary Macintosh file selection dialog box Press simultaneously the Command key the shift key and the key 0 on the numeric keypad instead of typing a file name This method lets you choose a text file with the ordinary Macintosh file selection dialog box however this file will not contain the source code but the name of the file s you wish to compile This method is most use ful if you have to compile a whole set of modules Instead of typing each time several module names you enter them into a text file
189. g SetP m initP is 0 0 END FOR SetP m initP one 1 0 SetP m randomize 1 0 CreateNewFile recordF Record results on file Markov DAT IF recordF res done THEN GetDateAndTime dt WriteChars recordF Run WriteChar recordF TAB FOR is firstState TO lastState DO FOR js firstState TO lastState DO GetDefltMV m f is js curMin curMax dummyStr ident dummyStr curFiling curTabul curGraphing WriteChars recordF ident WriteChar recordF TAB END FOR END FOR WriteChars recordF N WriteEOL recordF i 1 WHILE i lt maxRuns AND NOT ExperimentAborted DO SimRun PutInteger recordF i 0 WriteChar recordF TAB FOR is firstState TO lastState DO FOR js firstState TO lastState DO PutReal recordF f is js 8 4 WriteChar recordF TAB END FOR END FOR PutInteger recordF CurrentStep 0 WriteEOL recordF INC i END WHILE WriteChars recordF Legend WriteEOL recordF H DateAndTimeRec WriteChars recordF Run Number of simulation run within experiment WriteEOL recordF WriteChars recordF f i j Relative frequency of transition from state i to state j WriteEOL recordF WriteChars recordF N Number of transitions used to estimate f i j WriteEOL recordF RecordDateTime Experiment started on dt GetDateAndTime dt RecordDateTime Experiment ended on dt Close recordF END IF END IF END TheExperiment
190. g Machine the modeler may also access Dialog Machine routines himself and mix them with the functions provided by ModelWorks alone For instance the modeler might want to have an additional kind of monitoring not offered by the standard ModelWorks functions To accomplish this he may add a new menu with commands to open a window in which simulation results are to be displayed in a problem specific graph e g by coloring areas in a map according to the current values of model variables Besides since model objects belong to the model definition program and are fully accessible the latter poses no problem for the modeler Further details on the use of ModelWorks via the client interface are given below in the section Model development of this chapter 5 1 Simulation Environment 5 1 1 PROGRAM STATES OF THE SIMULATION ENVIRONMENT During a simulation session ModelWorks is always only in one of three states No simulation Simulating or Pause Fig T15 States are characterized by the availability of certain com mands In the state No simulation model and model object attributes can be interactively changed In the state Simulating user interactions are limited e g IO windows are inactivated In the state Pause the simulation is temporarily brought to a halt to allow for interactive changes T47 ModelWorks 2 0 Theory of parameters Limitations have been introduced to ensure consistency and because certain commands are meaningful only in parti
191. g flags The meaning of the formal procedure parameters are listed in the tables Tab R2 and R3 R 108 ModelWorks 2 0 Reference The call of procedure SetDefltGlobSimPars or SetDefltProjDescrs will have no effect until the global simulation parameters respectively the project description are reset ModelWorks solves equations e g differential equations by using an independent variable which is normally time It also needs the independent variable if no monitorable variable has been selected as abscissa variable X isX PROCEDURE SetDefltIndepVarIdent descr ident unit ARRAY OF CHAR SetDefltIndepVarldent overwrites the defaults of the descriptor descr identifier ident and the unit unit of the independent variable The predefined default values ModelWorks uses are the descr time ident t and no unit empty string The call of this procedure will have no effect until the global simulation parameters are reset The following procedures are actually only kept for convenience and upward compatibility with previous versions of the ModelWorks client interface In ModelWorks versions later than V1 1 their functions are also available by using the procedures SetDefltGlobSimPars respectively SetGlobSimPars PROCEDURE SetMonInterval hm REAL Sets the default of the monitoring interval only not the current value The call of this procedure will have no effect until the global simulation parameters are reset PROCEDURE SetIntegr
192. given in form of a closed analytical form but Table Functions Edit Reset Reset all SRE 3559230132 Fig R11 Menu Table Functions which are given by a table of values so called table functions Fig R13 ModelWorks will then linearly interpolate or extrapolate missing values during simulations As soon as the modeler imports any objects from the optional client interface TabFunc Fig T25 part II Theory ModelWorks will add an additional menu to the right of the menu Simulation Fig R11 This menu serves the editing and resetting of the declared table functions Edit Lets the simulationist edit a table function 9When using the menu command Select stash file the simulationist will always be first asked if he really wants to overwrite in case there should already exist a file with the same name as the stash file R 85 ModelWorks 2 0 Reference R 86 First the table function has to be selected by clicking into the corresponding radio button in an entry form similar to the one shown in Fig R12 Every table function will be associated with a short identifier which will be used by ModelWorks in this entry form EDIT one of the following Table Functions BRCMT CH BREMT O BRMMT O BRPMT O CFIFRT O CIMT O CIOR O DRCMT O DREMT O DRMMT C DRPMT O FEMT OFPEIT CH FPMT O NREMT O NRMMT C POLATT O POLCMT GO QLET CO QLFT O OLMT O OLPT Fig R12 Entry form to select a table function for the editing of
193. gold autoDefCol LineStyle unbroken broken dashSpotted spotted invisible purge autoDefStyle CONST autoDefSym 200C PROCEDURE SetCurveAttrForMV m Model VAR mv REAL st Stain ls LineStyle sym CHAR PROCEDURE GetCurveAttrForMV m Model VAR mv REAL VAR st Stain VAR ls LineStyle VAR sym CHAR PROCEDURE SetDefltCurveAttrForMV m Model VAR mv REAL st Stain ls LineStyle sym CHAR PROCEDURE GetDefltCurveAttrForMV m Model VAR mv REAL VAR st Stain VAR ls LineStyle PROCEDURE ClearGraph PROCEDURE DumpGraph Preferences and simulation environment modes Cz i at jk a a a Ail i a yl lg at a Sata pl ll a Al lye wy PROCEDURE SetDocumentRunAlwaysMode dra BOOLEAN PROCEDURE GetDocumentRunAlwaysMode VAR dra BOOLEAN PROCEDURE SetRedrawTableAlwaysMode rta BOOLEAN PROCEDURE GetRedrawTableAlwaysMode VAR rta BOOLEAN PROCEDURE SetCommonPageUpRows rows CARDINAL PROCEDURE GetCommonPageUpRows VAR rows CARDINAL PROCEDURE SetRedrawGraphAlwaysMode rga BOOLEAN PROCEDURE GetRedrawGraphAlwaysMode VAR rga BOOLEAN A 180 VAR sym CHAR PROCEDURE SetColorVectorGraphSaveMode cvgs BOOLEAN PROCEDURE GetColorVectorGraphSaveMode VAR cvgs BOOLEAN ModelWorks V2 0 Appendix L E E EEE SimMaster BIRR E EE E E RR RIK SK RR BY PROCEDURE RunSimMaster md PROC must always be called PROCEDURE DeclInitSimSession issp PROC PROCEDURE SimRun PROCEDURE PauseRun PROCEDURE
194. h given that the recording flag for graph dumping is currently set Tabulation of numerical simulation results for those monitorable variables for which the stash filing is currently set F writeOnFile Dumping of graphical simulation results encoded only machine readable under con trol of a particular recording flag ModelWorks can handle only one stash file at a time In the state No simulation it is always closed to allow for the inspection of its content by the simulationist ModelWorks automatically opens respectively closes the stash file at the begin respectively at the end of an elementary simulation run of a structured simulation experiment Unless the stash file name is changed its default name is ModelWorks DAT ModelWorks will use always the same file i e if a file with the same name already exists that file s old content will be lost and completely rewritten The stash file is written in a format which is not too difficult to read by a human being as well as easy to scan by a computer program post run analysis Furthermore it is also possible to transfer the results into another program e g a spreadsheet program or into a document pro cessing program which understands the RTF gt format These formats can not be changed At the heart of the information written to the stash file is the writing of the values of the moni torable variables for which the stash file monitoring has been set at every monitoring time La The
195. h as drawing of additional curves and displaying of validation data at discrete time points with or without error bars The following two modules belong to the auxiliary library of ModelWorks and are only briefly described here ReadData Allows to read data from an input file and to test the various conditions such as a minimum maximum range in an easy way It is typically used to enter measurements stored on text data files into ModelWorks for instance to compare simulated with observed data lFor detailed description of the functions of these modules refer to the listings of the definition modules in the appendix 2For detailed description of the functions of these modules refer to the listings of the definition modules in the appendix R 98 ModelWorks 2 0 Reference JulianDays provides calendar procedures to convert calendar dates into julian days and vice versa Julian days can be used for calculations e g to compute an elapsed time between two dates DateAndTime and WriteDatTim can be used to access the built in clock and to write dates and times RandGen contains a pseudo random number generator for variates uniformly distributed within interval 0 1 RandNormal provides a pseudo random number generator for normally distributed variates Nu 0 Since ModelWorks has been designed as an open architecture and is based on Modula 2 the modeler is free to add any module he wishes to the auxiliary library The modul
196. h bud moth system by means of ModelWorks V0 3 simulating the system behavior for the Upper Engadine Valley References Fischlin A 1982 Analyse eines Wald Insekten Systems Der subalpine Larchen Arvenwald und der graue Larchenwickler Zeiraphera diniana Gn Lep Tortricidae Diss ETH Nr 6977 Swiss Federal Institute of Technology Z rich Switzerland 294pp FROM DMLanguage IMPORT A171 ModelWorks V2 0 Appendix Language SetLanguage CurrentLanguage FROM DMMenus IMPORT Menu Command AccessStatus Marking InstallMenu InstallCommand InstallAliasChar FROM DMEntryForms IMPORT FormFrame WriteLabel DefltUse CheckBox UseEntryForm FROM SimMaster IMPORT RunSimMaster FROM SimBase IMPORT DoNothing FROM LBMMod IMPORT ActivateLarchLBMModel DeactivateLarchLBMModel LarchLBMModellsActive FROM LBMObs IMPORT ActivateLBMObsModel DeactivateLBMObsModel LBMObsModellsActive VAR modM Menu modActCmd Command PROCEDURE Choose CONST lm 6 VAR bf FormFrame ok modCB obsCB BOOLEAN cl INTEGER BEGIN cl 2 WriteLabel cl lm 1 Check models to be activated INC cl obsCB LBMObsModelIsActive modCB LarchLBMModellsActive CheckBox cl 1lm Observations Parallel Model Upper Engadine obsCB INC cl CheckBox cl lm Larch Larch Bud Moth Model b1 modCB INC cl bf x 0 bf y 1 display dialog window in middle of screen bf lines cl 1 bf columns 50 UseEntryForm bf ok IF ok
197. h simulation session i e after the start up of the simulation environment or when the simulationist chooses the command Settings Initialize session PROCEDURE DeclInitSimSession issp PROC 7 1 2 DECLARATION OF MODELS A model or a submodel is declared to ModelWorks or installed in the simulation environment with the procedure DeclM TYPE Model PROCEDURE Dec lM VAR m Model defaultMethod IntegrationMethod initialize input output dynamic terminate PROC installModelObjects PROC descriptor identifier ARRAY OF CHAR about PROC This procedure can be called any number of times but should be called for each model only once unless it has been removed in the meantime DeclM may not be called in the sub state Running s a part II Theory Fig T26 m is a variable of the opaque type Model exported by SimBase It may be used for further references to the model e g when accessing a model in order to change its integration method with procedure SerM It must be declared in the model definition program It does not matter where but m must be a global variable which exists as long as the model definition program IntegrationMethod Euler Heun RungeKutta4 RungeKutta45Var stiff discreteTime defaultMethod is the default integration method with which the model will be solved during simulations Moreover the modeler defines with this parameter also the type of model i e whether it is a continuous time or a discrete
198. hLn REAL In of minimum annual value found in anyone site ymeanDashLn REAL In of average annual value for whole valley ymaxDashLn REAL In of maximum annual value found in anyone site PROCEDURE ActivateLBMObsModel PROCEDURE DeactivateLBMObsModel PROCEDURE LBMObsModellsActive BOOLEAN END LBMObs IMPLEMENTATION MODULE LBMObs Revision history Author Date Description af 01 05 87 First implementation af 12 05 90 ModelWorks 2 0 adaptation now dynamic model activation and de activation supported Curve attributes set in particular if no observations available lineStyle is set to invisible FROM SimBase IMPORT Model DeclSV DeclM IntegrationMethod DeclMV StashFiling Tabulation Graphing CurrentTime SetSimTime GetGlobSimPars NoInitialize NoInput NoOutput NoDynamic NoTerminate NoAbout RemoveM SetDefltCurveAttrForMV SetCurveAttrForMV Stain LineStyle FROM DMFiles IMPORT Response TextFile Lookup Reset Close EOF GetCardinal GetReal legalNum A 168 ModelWorks V2 0 Appendix FROM DMAlerts IMPORT WriteMessage ShowAlert FROM DMConversions IMPORT CardToString FROM MathLib IMPORT Ln VAR storage for observations yminD ymeanD ymaxD ARRAY obsMod Model current larval densities used as a state variables yminDashl ymeanDashl ymaxDashl REAL k CARDINAL curVar ARRAY 0 20 OF CHAR curActive BOOLEAN kmin kmax OF REAL
199. hange see modifying current value model parameter Running see substate of simulation environment scaling 11 20 57 selection of object 17 58 59 sensitivity analysis 69 70 Simulating see state of simulation environment simulation environment 13 14 17 28 46 simulation environment mode 74 75 80 120 simulation run 18 46 50 52 53 84 113 simulation session 14 46 50 51 82 99 100 simulation time 18 27 54 77 107 108 simulationist 10 source see model definition program stand alone application 134 start consistency check 52 113 stash file 23 56 74 80 81 108 118 120 recording flags 80 108 109 state of simulation environment 48 67 114 No simulation 47 48 67 Pause 22 47 48 67 Simulating 47 48 67 84 state variable 11 12 54 55 66 91 structured model 36 38 41 44 155 structured simulation 47 50 51 67 68 84 113 114 115 see experiment submodel 36 38 42 43 44 subsequent monitoring 117 substate of simulation environment 67 68 114 No run 67 ModelWorks V2 0 Index Running 67 table 22 56 57 83 user interface 10 72 table function 85 87 105 112 154 versions of ModelWorks 125 user see simulationist terminate condition 52 114 185 186 BERICHTE DER FACHGRUPPE SYSTEM KOLOGIE SYSTEMS ECOLOGY REPORTS ETH Z RICH Nr No 1 FISCHLIN A BLANKE T GYALISTRAS D BALTENSWEILER M NEMECEK T ROTH O amp ULRICH M 1991
200. he currently selected state variables R 90 ModelWorks 2 0 Reference The IO window State variables displays information name identifier unit current initial value about the installed state variables of all models Fig R18 Furthermore it offers a mechanism to select state variables and to execute functions which operate on the current initial values of the selected state variables by clicking on the buttons in the window SET Columns set up Activates an entry form in which the display of the columns in the IO window State variables can be controlled Fig R19 The columns of which the display can be turned on or off are State variable names Full names of the state variables Ident Short identifiers of the state variables Unit Unit in which to measure the values of the state variables Jnitial value Current initial value of the state variable used to initialize the state variables at the begin of each simulation run Check the columns to be displayed in the window EJ State variable names E Ident EJ Unit EJ Initial value Fig R19 Entry form opened by the button in the IO window State variables Selects all state variables All subsequent button functions will operate on the scope All Fig T21 part II Theory i e on all state variables of all models Y Int Set initial value Opens an entry form in which the current initial value for the selected state variable s can be edited M
201. he entry form the value for the integration step and the monitoring interval to 1 0 The monitoring interval is the interval at which simulation results are displayed If this is smaller than the integration step the former is automatically reduced to generate the requested result display With these settings you can perform a simulation run The integration method Euler is the simplest integration algorithm therefore it is fast but not very precise After the integration the stash file ModelWorks DAT contained in the same folder as the MacMETH shell resides is ready for inspection and you can open it with any text editor you have available e g the desk accessory MockWrite2 It should contain the same simulation results as the table window look for look for DATA BEGIN of Run 1 For the next simulation choose another algorithm Bring the model window to the front select the model and click Choose the more precise algorithm Runge Kutta 4 To prevent overwriting of the stash file we change its name Therefore before starting the next simulation run choose first the menu command Settings Select stash file and give the stash file a new name e g ModelWorks DAT2 Only now start the simulation by choosing the menu command Simulation Start run The new run will give different results as you can easily verify in the graph For a more detailed numerical analysis you could use the values on the two stash files ModelWorks would
202. he menu command Quit results in the termination of the procedure RunSimMaster i e the simulation session is terminated and the simulation environment is left All models and all model objects will cease to exist The reverse is true also The modeler must be careful to ensure that all models and model objects exist as long as the simulation session lasts otherwise ModelWorks will not function properly In particular this means that models or model objects variables such as state variables must never be declared as variables local to a Modula 2 procedure After startup the simulationist has the choice either to start immediately a simulation with the predefined default values or to change interactively any settings The client interface has been designed such that the modeler can t make an error by unintentionally leaving out a needed value to define fully the initial value problem of a ModelWorks simulation run During simulations two additional windows the graph and the table window are automatically opened or brought to the front for the display of the simulation results Simulation results may also be written to a so called stash file for future references by ModelWorks or other application software such as a spreadsheet or a statistics program Simulations and interactive changes of parameter values or settings can be done freely in any order To issue a command to ModelWorks the simulationist has several options First the omni present menu ba
203. he standard version and runs on 512KBytes machines like the Macintosh Reflex Mac 512KE The following restrictions apply no graph printing except screen dumps no clip board support and no dumping of graphs onto the stash file Colors are available on color screens and on printer systems which support color screen dumps However the simulation environment mode restore graph with colors is not available 3 The IBM PC version V1 1 PC It is also a reduced subset from the standard Macintosh version Besides the same restrictions which apply to the Reflex Macintosh version this version can not support colors This is because of the MS DOS memory limitation of 640 KBytes Furthermore this version requires static linking 4 The Macintosh II version V2 0 II It is functionally identical with the standard Macintosh version but takes full advantage of the Motorola 68020 resp 68030 32 Bit CPU and the arithmetic coprocessor Motorola 68881 resp 68882 It is faster however it runs on Macintosh II family computers only The usage of the software for noncommercial purposes is free and unrestricted as long as the authorship of the used software is stated clearly on any redistributed model or other program i e any product descriptions or labels must state in writing that the Interactive ModelWorks Simulation Software by A Fischlin er al from the Swiss Federal Institute of Technology Z rich ETHZ has been used to develop the product All cop
204. heckedRadioButton RadioBut PROCEDURE GetCheckBox ei EditItem VAR boxChecked BOOLEAN PROCEDURE GetScrollBar ei EditItem VAR r REAL PROCEDURE InstallEditHandler u Window eh EditHandler PROCEDURE GetEditHandler u Window VAR eh EditHandler PROCEDURE SelectField ei EditItem PROCEDURE ClearFieldSelection u Window PROCEDURE EnableItem ei EditItem PROCEDURE DisableItem ei EditItem PROCEDURE IsEnabled ei EditItem BOOLE PROCEDURE EditItemExists ei EditItem BOOLEAN PROCEDURE GetEditItemType ei EditItem VAR it ItemType PROCEDURE RemoveEditItem VAR ei EditItem PROCEDURE RemoveAllEditItems u Window PROCEDURE AttachEditFieldObject ei EditItem obj ADDRESS PROCEDURE EditFieldObject ei EditItem ADDRESS LE E TEE ER IER RE IE KN REINE DMEntryForms KKK KEKE KR KKK KEKE KKK KKK TH TH TH TH TH HH N HH A HH HH a a a a VAR FieldInstalled BOOLEAN TYPE FormFrame RECORD x y INTEGER lines columns CARDINAL END DefltUse useAsDeflt noDeflt RadioButtonID PROCEDURE WriteLabel line col CARDINAL text ARRAY OF CHAR PROCEDURE CharField line col CARDINAL VAR ch CHAR du DefltUse charset ARRAY OF CHAR PROCEDURE StringField line col CARDINAL fw CARDINAL VAR string ARRAY OF CHAR du DefltUse PROCEDURE CardField line col CARDINAL fw CARDINAL VAR card CARDINAL du DefltUse minCard maxCard CARDINAL PROCEDURE LongCardField line col CARDINAL fw CARDINAL VAR longCard LONGCARD du DefltUse minLC
205. ible to draw only scatter grams instead of line charts The color currently in use for a particular monitorable variable is not only shown in the legend and used to draw curves in the graph but also displayed in the column Monitoring of the IO window Monitorable variables while displaying the current monitoring settings F T or Y In order to toggle between automatic definition and its overriding the simulationist must click into one of the line style radio buttons or into the radio button Automatic definition In particular note that in the current version of ModelWorks it is not sufficient to select just another color stain to turn automatic definition off without selecting also a new line style This is because automatic definition can not be set individually for the various curve attributes but hold for all attributes of a monitorable variable together either the curve attributes of a particular monitorable variable are all defined automatically or all are user defined If a multiple selection of monitorable variables has been made not only one but a series of entry forms will be offered one form for each monitorable variable This sequence can be terminated by pressing the push button Cancel Note however that in the latter case all changes which have been made to variables before can no longer be reversed Only the eventual changes made to the monitorable variable currently in display will be discarded This function causes sooner or
206. ible 0C SetCurveAttrForMV obsMod ymaxDashLn turquoise invisible 0C END IF ELSE SetCurveAttrForMV obsMod yminDash turquoise invisible 0C SetCurveAttrForMV obsMod ymeanDash turquoise invisible 0C SetCurveAttrForMV obsMod ymaxDash turquoise invisible 0C SetCurveAttrForMV obsMod yminDashLn turquoise invisible 0C SetCurveAttrForMV obsMod ymeanDashLn turquoise invisible 0C SetCurveAttrForMV obsMod ymaxDashLn turquoise invisible 0C yminDash MIN REAL stands for undefined ymeanDash MIN REAL ymaxDash MIN REAL T F 17 17 yminDashLn Ln negLogDelta ymeanDashLn Ln negLogDelta ymaxDashLn Ln negLogDelta END IF test whether Output is called just before legend drawing then make an exception and draw legend with curve attributes used for display when observations exist GetGlobSimPars t0 tend h er c hm IF TRUNC t0 0 1 k THEN SetCurveAttrsForLegend END END Output PROCEDURE Terminate BEGIN SetCurveAttrsForLegend END Terminate PROCEDURE ModelObjects BEGIN DeclMV yminDash yLL yUL Minimum larval density per site Ymin larvae kg branches notOnFile notInTable notInGraph SetDefltCurveAttrForMV obsMod yminDash turquoise spotted 0C DeclMV ymeanDash yLL yUL Average larval density in valley Y larvae kg branches notOnFile writeInTable notInGraph SetDefltCurveAttrForMV obsMod ymeanDash turquoise dashSpotted 0C DeclMV ymaxDash yLL
207. ic pseudo random number generator producing uniformly distributed variates within interval 0 1 The generator is based on a combination of three multiplicative linear congruential random number generators Remarks The generator is highly portable and produces very long cycle random number sequences They exceed the usual period length of MAX INTEGER given by the machine dependent word length Thus the generator produces satisfactory results even on a personal computer with a small word length e g 16 Bit machines and it is efficient since it does not require double precision arithmetics On 32 Bit machines like IBM main frames or the Apple Macintosh PC this means that the slow 64 Bit multiplication and division can be avoided The cycle length of the generator is estimated to be gt 2 78 E13 so that the sequence will not repeat for over 220 years in case that 1000 variates were calculated per second Wichmann amp Hill 1987 References Wichmann B A amp Hill I D 1982 An efficient and portable pseudo random number generator Algorithm AS 183 Applied Statistics 31 2 188 190 Wichmann B amp Hill D 1987 Building a random number generator A Pascal routine for very long cycle random number sequences Byte 1987 March 127 28 Programming Design A Fischlin 21 Dez 88 Implementation A Fischlin O Roth 21 Dez 88 Swiss Federal Institute of Technology Zurich Systems Ecology Department of Environmen
208. ile during simulations for Z Models O State variables O Model parameters EJ Monitorable variables O Graph Fig R7 Entry form Project description D Footer By default the footer contains the date the time and the simulation run number but it may also be used to store any other information If the flag Automatic data amp time update in footer is turned on ModelWorks will update this information at the begin of each simulation run If the menu command Print graph is chosen this footer string will always be printed in a small font R79 ModelWorks 2 0 Reference size below the graph However the graph transferred into the clipboard will never contain this string Record data on the stash file during simulations for With these recording flags the simulationist may control which information and values are written onto the stash file Check the appropriate boxes for models model parameters state variables monitorable variables and the graph if you wish to have them written onto the stash file at the begin for all except the graph and at the end graph only of simulation runs Note that the flags Models Model parameters State variable and Monitorable variables mean that information about these objects is written to the stash file They are the descriptor the identifier the unit unless a model and the object specific current values Except for the monitorable variables information about all objects will be
209. ime submodels jE Z on the output of the discrete time submodels j A is resolved by using the last defined values sample of all variables of A while mapping time t to k using Eq 12 hold ujs t hL ur WRI yE Yo5Os Yp 1 K Yn K INT t 7c HUK ROL u 8 t ur lk y18 t y2 Yp 1 kK EO t INTW k EA The dynamic equations for the calculation of the derivatives for or the new values for A of the state variables of the submodels iF resp 1 dx fe Kl WH pre t G Z dynamics of submodels 15 i 8c SKEI fig xK uK pps K EA The output of the submodels i resp i are calculated of the states and the parameter set of the particular submodel 15 resp 1 An output variable must not depend directly on an input variable no direct output input coupling avoids unresolvable circularity Yield Sie Ke Pg t E outputs of submodels i5 id 9c Yi Zis LK Pe k A Some of these outputs are not connected to other submodels but are global outputs These elements from y t resp y k form for each submodel the global output vectors y t resp yE The global output y of the structured model consists of two vectors one for the global continuous time output y and the other for the global discrete time output y gt Each is again composed from the global output vectors y t of the continuous time submodels i5 Z resp the global output vectors
210. imulation parameters a model s inte gration method the initial values of state variables model parameters or monitoring settings two copies One is the default value the other is the current value All simulations use only the current values and unless accessed via the client interface the defaults cannot be changed by the simulationist from within the standard simulation environment Fig T22 The defaults are defined by two mechanisms First ModelWorks assigns to every global simu lation parameter the project description and the stash file name a predefined default value Tab T1 Then the model definition program may overwrite these values with the defaults as specified by the modeler E g does Model Works use a predefined default simulation start and simulation stop time of ty 0 0 respectively teng 100 0 If the modeler wishes to use a dif ferent default simulation time range he calls the procedure SetDefltGlobSimPars from module SimBase to define it e g with the statement SetDefltGlobSimPars 1989 0 2000 0 While starting a model definition program ModelWorks always assigns automatically all de faults either provided by ModelWorks or overwritten by the modeler to the current values Fig T22 The modeler is forced by the client interface to specify defaults for all models and model objects They are the values passed to ModelWorks while declaring the particular ob jects E g the value 0 1 is the default of the model
211. ing NoModelObjects and to defer all model object declarations to a later time note that this requires proper programming of such a feature since it is not available in the standard simulation environment The modeler does not need to install any one of his her implementation for the procedure parameters initialize input output dynamic terminate or installModelObjects for a convenient use the following procedures from module SimBase with an empty body can be used as actual arguments when calling DeclM PROCEDURE Nolnitialize PROCEDURE NoInput PROCEDURE NoOutput PROCEDURE NoDynamic PROCEDURE NoTerminate PROCEDURE NoModelObjects PROCEDURE NoAbout PROCEDURE DoNothing See to it that calculations which should be performed only once per time step are included in the procedure input or output only not in the procedure dynamic which may called more than once per time step For further information on the correct use of these procedures see part II Theory chapter Simulations of this manual s a Fig T16 T17 and T18 R 101 ModelWorks 2 0 Reference Finally the last formal procedure parameters are used to identify and describe a model so that the simulationist may recognize it during simulation sessions descriptor String containing a long description of the sub model m identifier Short string identifying the sub model m Although there is no limit to the actual size of this string it is advisable to keep it as sho
212. ing terms and concepts program text or source code compilation and the execution of programs 3 2 2 MODEL DEFINITION PROGRAM FOR THE NEW MODEL The ModelWorks distribution diskettes contain an editor it is the shareware desk accessory MockWrite a simple straight forward text editor To program the new model choose MockWrite in the Apple menu and open the file Logistic MOD in the folder Sample Models We create a copy of this file with the same content by choosing the menu command MockWrite Save As Give the new file the name GrassAphids MOD5 and save it in the folder Work This is the copy we shall use to develop the new model thus avoiding to destroy the program text of the logistic grass growth sample model First change the module name It is recommended to use the same name for the module name as you have used to name the file containing the module i e GrassAphids Throughout the following explanations the affected program text is shown together with its context The text portions actually having been altered or added are shown underlined At the beginning of the file MODULE GrassAphids My name today s date ES 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 AAA Lotka Volterra Grass and Aphids EEE 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 20 and at the end of the file RunSimMaster ModelDefinitions END GrassAphids There is no need to change the import list All objects requir
213. inimum and maximum values used for the scaling of the curve to the ordinate while drawing values of the monitorable variable mv in the graph The default minimum and maximum of the ordinate scale is re assigned to the current scale minimum and scale maximum by ModelWorks during the initialization phase of the simulation session or after every reset of the scaling During a simulation session the current scale minimum and scale maximum may be changed by the simulationist using an IO window or by the modeler via procedure SetMV There apply no restrictions to the values of these variables During interactive changes ModelWorks will use the range boundaries MIN REAL and MAX REAL descriptor String containing a long description of the monitorable variable mv This string may have any length but might not be visible till its end when it is too long to fit into the IO window column where it is displayed during a simulation session see also identifier Example Density of alga Scenedesmus obliquus identifier Short string identifying the monitorable variable mv This string should be kept as small as possible in order to ensure full visibility for the display in small IO windows during a simulation session In particular on small screens IO windows become small in the tiled window position see menu command Tile windows and they will display only this identifier to denote the monitorable variable mv Example U xa unit String containing
214. inition program Each monitorable variable is associated with a clipping range used for the graphical display of the simulation results Fig T2 All values specified by the modeler are remembered by ModelWorks as the so called default values The values currently in use by the simulation environment are called the current values While starting the model definition program ModelWorks assigns the default values to the current values This is called a reset Any time the simulationist wishes to do so he may execute a further reset of a specific class of values so that their current values are overwritten with their defaults This mechanism is most useful if the simulationist wants to resume a well defined state before continuing with his work especially after having made many and complex interactive changes From the definitions x y and x y follows x i e the variable substitutions y gt x y gt x y x rearrange resulting two equations to make the derivatives explicit T12 ModelWorks 2 0 Tutorial ModelWorks Client s interface Import list Model defini tion program ModelWorks simulation program Client s interface Model defini tion program data exchange Cotrolled data exchange during program execution Fig T3 Organization of ModelWorks ModelWorks is the constant part common to any simulation program forming the simulation environment The variable part the model definition pr
215. ion by the extension APP Now type GEM and start the program FIRST APP If you had problems compiling or linking check the PATH statement type PATH at the DOS prompt and make sure that the TopSpeed directoy is included in your path If you have named the directories other than TOPSPEED and DM you must edit the file M2 RED in the compiler directory to reflect the different names Also make sure that the DM directory has been copied completely including the subdirectories DM DEF DM OBJ and DM EXAMPLES If you have the 360K disks you might not have copied the extra files from disks 2 and 4 into the ap propriate places D 1 4 Installation of ModelWorks Copy the whole directory MW onto your hard disk If you you have the 360K disks you must copy the rest of the ModelWorks files from disk 2 of 2 manually into the directory C MW Copy the file M2 RED into the compiler directory TOPSPEED To test ModelWorks go to the directory MW SAMPLES start the compiler don t forget to set the compiler option Runtime checks to ON and use the Make facility ALT M to create the LOGISTIC program Exit and rename LOGISTIC EXE to LOGISTIC APP run GEM and see how it runs Once within LOGISTIC check the free memory status by calling System Info from the leftmost menu the one with the For further information about GEM or the compiler consult the manuals For more information about the Dialog Machine see this appendix the
216. iption D Displays the entry form to edit a global project description and control the recording of data on the stash file Project title String which can be freely used to describe the on going project i e for instance a title for the current simulation session If the menu command Print graph is chosen this Project title string will always be printed in bold above the graph However the graph displayed and transferred into the clip board will contain this string only if the flag Use in Graph has been checked In the graph window this string will be displayed in the middle and at the top of the data panel Remarks String which can be freely used to add some remarks on the on going project For instance it may be used as a sub title similar to the project title string or it may contain some information on specific model parameter settings used in the simulations If the menu command Print graph is chosen this Remarks string will always be printed just below the title in a smaller font and with style plain However the graph displayed and transferred into the clip board will contain this string only if the flag Use in Graph has been checked In the graph window this string will be displayed to the right of the legend at the bottom of the window Project title EJ Use in Graph Project Description Remarks J Use in Graph Footer EJ Automatic date amp time update in footer 23 Mai 1990 14 47 Runi Record data on the stash f
217. is is much more convenient as if the simulationist would always have to assign the desired values interactively at the begin of each simulation session If a simulation study has advanced to a later stage it is often desirable to be able to run multiple simulation runs in a systematic well defined way The interactive control of the simulations becomes then rather an obstacle than a help On the other hand if much effort has been invested in the development of a complex model it is desirable to be able to run structured simulations under program control using the same model implementation To support the modeler in such tasks is exactly the purpose of most of the additional objects exported by the client interface The principles behind the usage of the client interface have been described elsewhere Manual Part II Theory in the chapter Modeling Please consult also the listings of the client interface definition modules SimMaster and SimBase plus the sample model definition programs in the appendix while reading the following explanations of the Modula 2 objects exported by the client interface The following two modules belong to the optional part of ModelWorks and are only briefly described herel SimIntegrate Provides means to integrate autonomous models without any monitoring and without affecting the global simulation time of the simulation environment SimGraphUtils provides utilities to make output to the graph window and the graph suc
218. itel in your system or you copy your favorite text editor onto the disk 2 2 In case you intend to use MockWrite regularly please don t forget to pay the shareware fee Please quickly test your installation by starting first the MacMETH shell RMSMacMETH with a doubleclick and by choosing the menu command Execute If the standard file selection dialog box appears select the sample model Logistic OBM contained in the folder Work on the first diskette 1 2 If you see a screen like the one in Fig A3 the installation is complete File Edit Settings Windows Simulation State variables Model parameters Monitorable variables n onitorable variable names Ident Unit 3 0 10 0 Minimum Maximum 0 000 1000 000 Fig A3 Screen which should appear while testing the installation of the ModelWorks software and activating the already compiled model definition pro gram Logistic OBM If you are not yet familiar with ModelWorks please quit the program now by choosing the menu command Quit under the menu File Then consult Part I Tutorial of this text C 2 HOW TO WORK WITH MACMETH Modeling with ModelWorks Macintosh versions V2 0 V2 0 Reflex V2 0 H is done by de veloping Modula 2 model definition programs using the MacMETH A Fast Modula 2 Language System Wirth et al 1988 The sequence of steps to follow is shown in Fig 23 in part II Theory section The model development cycle MacMETH consists of an application plus several subp
219. its content and restarts tabulating from the top again called a page up In the current version of ModelWorks any erased values are lost and the simulation has to be repeated to display them again 1949 8 1954 8 1959 8 1964 6 1969 6 1974 8 1979 48 Ber ime Curves Minimum Hax imum Unit Lrt s 4 505 6 297 Tbmkg Lacy 3 4 6605 6 397 larvae Ka Branches Fig T19 The graph window of ModelWorks ModelWorks can display in the graph window one graph only The graph has a linear abscissa x axis with time or any monitorable variable as independent variable allowing for state space curves and a linear ordinate y axis with a fixed scaling from 0 1 According to the currently set minimum and maximum values for the range of interest an arbitrary number of dependent variables range shown in the legend can be plotted simultaneously in the graph An unlimited number of simulation runs can be recorded in one graph The graph will be automatically cleared after changes of the graph definition the global simulation parameters e g if the start or stop time has been changed and time is the abscissa or if the window is resized after a 6 Actual number of rows erased depends on the currently set preferences or simulation environment modes The number specified as Common rows between page ups in table defines what happens during a page up First it specifies how many rows at the bottom are not erased but copied to the top of the nex
220. ixel based raster graphics Graph restoration becomes necessary whenever the simulationist moves rearranges or closes windows and the graph window is involved Note that with this option active graphs may be restored slower and more memory may be needed Otherwise graphs are restored in black and white only Vector graphics contain data about particular objects and the coordinates defining them e g a line is stored as a line object together with the coordinates of its begin and end point Hence vector graphics are usually of a higher quality than raster graphics However the printing of a vector graph may require too much time if draft quality of a graph would be sufficient Note that with this option active complicated graphs particularly if drawn during large experiments may use up tremendous amounts of memory On black and white monitors activate this mode if you wish to use color printers or transfer the graph via the clipboard to other color devices Not available in Reflex and PC version Quit Q Terminates the program and goes back to the program level from which the model has been started Usually this is either the MacMETH programming environment or the Finder 6 1 3 MENU EDIT Enin Fig R4 Menu Edit The whole menu is not available in the Reflex and PC version Undo Z not available present only for compatibility with the user interface guide lines Cut X Clears the graph and copies it into the clipbo
221. kind of all models all objects of a particular kind of a model individual object of a particular kind Since model objects belong to models the selection of a model in the models IO window also implies the selection of all its objects Hence the scopes in the models IO window are individual model respectively all objects of a model all models respectively all objects of all models model model object model object model object model model Fig T21 Scopes used for the selection of operands in the IO windows Selecting a model implies the selection of all model s objects Note that the operands actually affected by an operator are determined also by the operator it self For instance The selection of an individual model does not only allow to change an at tribute such as the integration method of this model but also to reset the values of all its objects such as the resetting of all initial values of the model s state variables to their defaults T59 ModelWorks 2 0 Theory In an inactive IO window no objects can be selected The simulationist can recognize this status if clicking within the list field does not invert any lines Second the button field 2 on the left side contains a palette of adjacent square buttons Each button has a separate function operator which can be activated by clicking on the little button picture with the mouse There are two kinds of functions basic window functions such as wi
222. l Sample Models su ee eigenes 152 H 1 The Sample Model Logistic Grass Growth Logistic MOD nesses 152 H 2 The New Model GrassAphids MOD see 153 H 3 Sample Model Using Table Functions UseTabFunc MOD sosse 154 H 4 A Mixed Continuous and Discrete Time Sample Model Combined MOD au ei eaten sane ariii iiia 155 H 5 Research Sample Models nannte a dena nee 157 H 5 1 Third order finite Markov chain Markov MOD ccenennnen 157 H 5 2 Population dynamics of larch bud moth LRM sss 164 Di Conic ke Reterenees ze TT conan a a eis 173 1 1 Dialog ea TT 173 1 2 ModelWorks Client Interface and Optional Modules sss sese sese 179 INDEX T 183 ModelWorks 2 0 About ModelWorks and this Text ModelWorks is a simulation environment to solve dynamic systems as they are used in biology physics chemistry environmental and engineering sciences to model various processes It is also particularly well suited to be used by university students during a modeling course ModelWorks can be used for simple didactic models as well as for very complex research models ModelWorks allows to work with an arbitrary number of dynamic models described by differential or difference equation systems A global model can be separated into possibly hierarchically organized submodels which exist as independent units communicating via output input coupling Modular and hierarchical modeling is supported which is particularly useful if for instance one wi
223. l monitorable variables of all models The stash file name and directory as defined with the menu command Select stash file is not affected Not available as a menu command in Reflex and PC version Reset All model s tabulation Resets the tabulation settings writeInTable notInTable of all monitorable variables of all models Not available as a menu command in Reflex and PC version Reset All model s graphing Resets the graph settings isX isY notInGraph of all monitorable variables of all models Not available as a menu command in Reflex and PC version Reset All model s scaling Resets the minimum and maximum values used for the scaling of all monitorable variables of all models on the ordinate These scaling extremes define the range of interest Fig T2 part Tutorial and are used during the drawing of values of the monitorable variables in the graph Not available as a menu command in Reflex and PC version Reset All model s curve attributes Resets the curve attributes of all monitorable variables of all models to their default values Not available as a menu command in Reflex and PC version Initialize session Calls the procedure again which was installed by the modeler with DeclInitSimSession from module SimMaster Note that this procedure has already been called at least once during the start up of the simulation session Be warned that it could contain erroneously calls to procedures which must not be called repeate
224. lWorks applies the so called sample and hold technique see also below under Simulation environment of the next chapter ModelWorks Functions Finally a simple example shall illustrate the whole concepts discussed in this chapter The model is a system consisting of two ordinary nonlinear first order differential equations First it shall be modeled simply and secondly it shall be modeled as a structured model built from submodels Ex The following model equations shall be modeled first within a single model Eq 13 X1 ax bx cxx2 Model M 13 X2 c X1X2 dx2 This system consists of two ordinary but coupled differential equations formulated according to Eq 8a with neither input nor output completely autonomous system See Fig T13a for the relational diagram of this model system Secondly the two differential equations shall be distributed into two separated submodels 14 respectively 15 which are coupled with each other Fig T13b ul y2 input according Eq 7a x ax bx cxw dynamic according Eq 8a Submodel u 14 y X output according Eq 9a T 43 ModelWorks 2 0 Theory respectively u2 y input according Eq 7a X1 ax bx cx dynamic according Eq 8a Submodel u2 15 y2 X2 output according Eq 9a Model M a Structured Model System Submodel p1 u1 a l Y2 Submodel H2 b Fig T13 Relational diagrams of a model once a formulated as a single elementary
225. lar linking loading and debugging will function without problems and more conveniently if you obey these conventions and restrictions C 3 CONFIGURING MODELWORKS The ModelWorks software can be configured by file renaming any time in the following ways e SANE Standard Apple Numeric Environment usage by procedures from module MathLib File MathLib OBM in the folder M2MiniLib on diskette ModelWorks 1 2 comes in two versions a MathLib OBM NO SANE B MathLib OBM SANE These files are different implementations of module MathLib for exported procedures see e g quick reference listing of the Dialog Machine one uses SANE the other not Implementation a does not use SANE but instead much more efficient but less precise 32 bits floating point arithmetics which are not available in SANE Implementation uses SANE which is much slower but results are computed with maximum precision 80 bits floating point arithmetics even exceeding requirements of the IEEE standard 754 for binary floating point arithmetic CoDY 1981 IEEE STD 754 1985 1985 In general it is recommended to use implementation a if you work with single reals type REAL uses 32 bits and standard ModelWorks 2 0 The loss in accuracy stems mostly from the fact that you store in particular the intermediate results only as single precision reals type REAL According to our research the latter effect is so dominant on the precision of the final results that it is rarely wo
226. lation Start run Define a graph where the predator is plotted versus the prey state space curve Select the prey and toggle its curve definition by clicking on the button Click the button to define a plot which uses the x axis abscissa to plot the prey values Start a new simulation run The resulting curve shows nicely how the grass and the aphids reach an equilibrium point Set ca 0 0 no self inhibition of the prey population This results in a different stability behavior of the system The oscillations of the population are no longer damped but persist in a marginally stable limit cycle In the state space you may observe closed trajectories each corresponding to such a limit cycle You should have obtained a graph similar to the one shown in Fig T6 Marginally stable limit cycles can be easily perturbed verify this by changing the integration method to Euler or while using the method Heun by increasing the integration step and the monitoring interval up to 0 5 How accurate is the numerical integration algorithm Congratulations You have reached the end of the introductory tour through ModelWorks You should have learned to develop and simulate simple models using ModelWorks In case of difficulties with Modula 2 please refer to WIRTH N 1988 Programming with Modula 2 Springer Verlag Heidelberg New York 4th corrected ed or the MacMETH documentation WIRTH et al 1988 for details on the specific MacMETH M
227. lation environment for the following set tings and parameter values Global parameters and settings e start and stop time for simulation runs e integration step respectively maximum integration step plus maximum relative local error e discrete time step coincidence interval e monitoring interval e project description consisting of a title remark and footer string plus parameters which control the display of strings in the graph respectively the recording of information on models and model objects together with their current values and settings on the stash file recording flags e stash file name Model specific attributes e integration method Model objects specific attributes e initial values of state variables e values of model parameters e kind of monitoring and scaling for monitorable variables In addition the simulation environment of ModelWorks offers a versatile reset mechanism which allows to reset any parameter or setting which may have been changed interactively during the simulation session by the simulationist or via the client interface by the model definition 1 only if at least one continuous time sub model is present 2 only if at least one continuous time sub model with a variable step length integration method is present 3 only if at least one discrete time sub model is present T 46 ModelWorks 2 0 Theory program The values or settings to which ModelWorks resets are the so calle
228. layTime showAtInit showAtTerm noAutoShow VAR timeIsIndep REAL PROCEDURE DeclDispData mDepVar Model VAR mvDepVar REAL mIndepVar Model VAR mvIndepVar REAL X V vLo vUp ARRAY OF REAL n INTEGER withErrBars BOOLEAN dispTime DisplayTime In order to display a data series e g validation data f ex before a simulation run the necessary data have to be declared beforehand i e normally just at the end of all other ModelWorks objects declarations The variables are as follows real arrays are called by name for efficiency only mDepVar model to which belongs the mvDepVar mvDepVar monitoring variable representing the dependent data array mIndepVar model to which belongs the mvIndepVar mvIndepVar monitoring variable representing the independent data array if mvIndepVar is specified timeIsIndep or is not a declared monitoring var then time is assumed to be the independent variable x array of independent values v array of dependent values vLo array of lower e g confidence values vUp array of upper e g confidence values n number of given data withErrBars flag if TRUE error bars will be drawn dispTime the time when the data should be displayed Note The curve attributes of the data to display can be set through the procedure SetCurvAttrForMV on the monitoring variable mvDepVar A 144 ModelWorks V2 0 Appendix and the default strategy for curve at
229. ld also require to click into the button Use immediately after having clicked the button nitial i e without any editing inbetween For actual resets use the menu command Table Functions Reset To re draw the graph with the current coordinates of the supporting values as shown at the right of the graph click into button Draw The push button Use accepts the edited coordinates of the supporting points for the new current values and redraws the graph Since it is the default button it may also be used by pressing either the key Return or Enter Note that from then on computations of the model will use the new values to evaluate function values The push button Close discards all editing and closes the window the table function will remain untouched exactly as it was before the menu command Edit has been chosen Editing of table functions is possible in the states NoSimulation and Pause It is also only available if there is currently at least one table function present which is modifiable see below section Declaration of table functions and and usage of module TabFunc in the next chapter of this part This may change even during a simulation session since the modeler may add or remove table functions or change their attributes dynamically via the client interface Reset Resets the values of an individual table function to their defaults This command displays an entry form similar to the one shown in Fig R14 Select the table f
230. lication To link an application start the MacMETH linker from within the MacMETH shell and specify the object file OBM file of the model definition program functions only by typing the name and add the linking option a You get an application which will work fine but which does not have its own icon and is still bundled with MacMETH Normally this causes only troubles very confusing icon behavior in the Finder desktop and should be avoided Follow the steps described below to create a true stand alone application conforming fully to the standard user interface of your Macintosh computer Procedure A In case you should have access to the shareware application conia 6 3 or a later commercially available version there is a simple way to unbundle the application from MacMETH and provide the new application with its its own unique icon and bundle Just start iconia and define the icon its mask the creator string 4 letters not conflicting with any already existing creator present on your disk the version information shown by the Finder via the menu command File Get Info and execute the menu command File Compile by selecting as the destination file your just linked application That s it If you use the shareware version please don t forget to pay the shareware price Procedure B First start the resource editor ResEdit and remove all resources of the following types BNDL ETHM FREF ICN MLNK SIZE This i
231. ling L Always document run on stash file Tabulation Once changed immediately redraw table Common rows between page ups in table Graphing Z Once changed immediately redraw graph E Restore graph with colors and high quality vector graphics for printing and clipboard Fig R2 Entry Form of the menu command Preferences shown with settings recommended for the beginner Filing If the simulation environment mode Always document run on stash file is active the stash file is opened at the begin of every simulation regardless of the current setting for stash filing of the monitoring variables If this mode is inactive the stash file will only be opened in case that at least one monitoring variable has the stash filing currently set F In case you rather use the stash file for run documentation purposes than for post run analysis purposes it is recommended to have this simulation environment mode active It will then force the opening of the stash file always and document every simulation The recording flags and the stash file settings F of the monitorable variables will then no longer affect the file opening but only determine which data are to be written to the stash file s a below menu command Set Project description recording flags Tabulation If the simulation environment mode Once changed immediately redraw table is active the table is redrawn immediately after each change in the tabulation settings Otherwise the
232. lled another time unless the state variable has been removed in the meantime DeclSV may not be called in the sub state Running s a part II Theory Fig T26 PROCEDURE DeclSV VAR s ds REAL initial minRange maxRange REAL descriptor identifier unit ARRAY OF CHAR The model to which the state variable will belong is normally the last model declared with a call to procedure DeclM unless the procedure SelectM has been called for another model The meanings of the formal procedure parameters of DeclSV are s Variable to be declared as state variable DeclSV does assign to s the value defaultInitial The real s can be declared anywhere in the model definition program and may be even part of any data structure However make sure that it is declared as a global real variable and does exist as long as the model definition program ds Variable to be declared as the derivative ds dt for continuous time models using time t as independent variable or the new value s k 1 for discrete time models using time k as independent variable of s For every state variable the derivative or the new value must be assigned to this variable by the procedure dynamic which is called during numerical integration Normally ds appears only on the left side of the dynamic equations in procedure dynamic DeclSV assigns to ds the value 0 0 R 102 ModelWorks 2 0 Reference defaultlnitial Default initial value for state variable s ModelWorks uses the curren
233. lobal simulation parameters of ModelWorks The first column contains the identifiers used to designate the corresponding variables in the client interface the second the symbol used to denote the corresponding variable in this manual fide Profilen remark Remark ting for __ Fooersing OoOO E wine With tite in graph fwremark With remarks in graph fautofooter Automatic update of date time and run number in footer R 107 ModelWorks 2 0 Reference Identifier _ Symbol recM Recording of data on models in stash file recSV Recording of data on state variables in stash file receP Recording of data on model parameters in stash file receMV Recording of data on monitorable variables in stash file receG Recording of graph in stash file at end of run Tab R3 Global project description of ModelWorks The first column contains the identifiers used to designate the corresponding variables in the client interface 7 2 1 a Retrieval of read only current values A user can access internal variables Tab R1 of ModelWorks by means of special procedures This guarantees undisturbed data consistency For instance the procedure PROCEDURE CurrentStep INTEGER exported by module SimBase returns the current simulation step i e the current value of discrete time must not be confounded with the integration step used by numerical integration for continuous time models Note
234. lso if you want to alter an allready existing curve A142 ModelWorks V2 0 Appendix PROCEDURE RemoveCurve VAR c Curve This procedure removes a curve definition This procedure sets c to nonexistent PROCEDURE DrawLegend c Curve x y INTEGER comment ARRAY OF CHAR Draws a portion of curve c with the current attributes at position x and y and writes the comment to the right of c After this procedure the pen location is just to the right of the string comment so it s possibe to add for example values of parameters by calling DMWindowIOo procedures WriteReal etc just after this procedure PROCEDURE Plot c Curve newX newY REAL You can plot draw a curve from the last saved position to the point specified by the new coordinates newX and newY Note ModelWorks resets the pen position when clearing the graph Errors If the point specified by newX and newY lies outside the integer pixel range DM2DGraphsDone will be set to FALSE PROCEDURE Move c Curve newX newY REAL moves the pen to postion x y Typically used to draw several curves with the same attributes to reset the pen position after having drawn a curve Errors If the point specified by x and y lies outside the integer pixel range DM2DGraphsDone will be set to FALSE PROCEDURE PlotSym x y REAL sym CHAR draws the symbol sym at the position x y May be used as an alternate method to
235. ly drawn curves are not erased unless demanded by a change of the graph settings since the last simulation a curve added or removed scaling changed In the upper right corner the current run number k and the current simulation time t are displayed in a small window k t This command lasts as long as the simulation runs It may be terminated by the simulationist menu command Stop Kill run or by the model i e if the simulation time reaches the stop time or if the installed termination condition returns true Note that the menu command Halt run Pause does not really terminate this command Halt run Pause H or Resume run R Temporarily halts or pauses a simulation run if the current program state is Simulating The new program state entered is Pause If the current program state is Pause this menu command will resume the interrupted simulation run where it has been left i e reenter the state Simulating and continue with the integration monitoring etc A pause can be used to study a curve or a tabulated result in more detail or can be used to change model parameters in the middle of a simulation Note however that current values of model parameters can only be modified if the flag rtc run time change has been set for that particular parameter Keyboard equivalent for Resume run is not available in PC version Stop Kill run K This command terminates the simulation before the simulation time reaches the stop time teng K
236. ly following runs are skipped by ModelWorks ModelWorks accomplishes this by emptying the body of the procedure SimRun from module T51 ModelWorks 2 0 Theory SimMaster However ModelWorks is unable to actually stop the experiment procedure which will be otherwise executed till it ends as programmed by the modeler Simulation run Internal initialization Initialization of state variables Client initialization procedures Initialization section false Start consistency check lt Client Output procedures Client Input procedures Client Dynamic procedures Initialization of standard monitoring or Inte gr ation lo op standard monitoring procedures Initialization of client monitoring or client monitoring procedures Update of state variables yes e check Simulation killed or Stop time reached or Termination condition true no Termination section Fig T17 Simplified flow chart of an elementary simulation run as performed by ModelWorks A run consists of the three basic steps initialization integration and termination T52 ModelWorks 2 0 Theory 5 1 2 c _ Elementary simulation run This level can always be accessed by the simulationist directly without going through the exper iment level see above by choosing the menu command Start run under menu Simulation Otherwise this level is the next level below the level of the structured
237. m It contains the Modula 2 translation of the mathematical equations describing the model s dynamics here Eq 1 for a proper functioning of ModelWorks it is very important that this procedure computes the exact values of the derivatives or new values of all state variables as required by the given equation s PROCEDURE Dynamic BEGIN grassDot cl grass c2 grass grass END Dynamic The procedure Objects contains the declarations of all model objects For each of the four model objects models state variables parameters and monitorable variables there exists a special declaration procedure Once such a procedure has been called ModelWorks knows the variable defaults plus ranges corresponding to the model object and can access it to maintain its values or can show it in a window or use it to display its current value in a graph It is mandatory to declare a model if you wish to declare model objects see below The declaration of model objects i e state variables model parameters or monitorable variables is optional and depends only on the current needs The procedure DeclSV declares the state variable grass DeclSV grass grassDot 1 0 0 0 10000 0 Grass G g dry weight m 2 The actual parameters are the two real variables for the state variable itself grass and its derivative grassDot Next there are three real constants the default initial value and the upper and lower limit of the range of initial val
238. make scatter grans Errors If the point specified by x and y lies outside the integer pixel range DM2DGraphsDone will be set to FALSE PROCEDURE PlotCurve c Curve nrOfPoints CARDINAL x y ARRAY OF REAL Plots an entier sequence of nrOfPoints coordinate pairs contained within the two vectors x and y May also be useful to implement an update mechanism Errors If the point specified by x and y lies outside the integer pixel range DM2DGraphsDone will be set to FALSE If the maximum number of elements of x or y is less than nrOfPoints then only the lower number of elements of either x or y will be plotted PROCEDURE GraphToWindowPoint xReal yReal REAL VAR xInt yInt INTEGER Calculates the pixel coordinates xInt and yInt of the graph s window see WindowIO from the specified graph coordinates xReal and yReal Note that the vertical axis of the ModelWorks graph is transformed to yMin 0 0 and yMax 1 0 Errors If the point specified by xReal and yReal lies outside the integer pixel range DM2DGraphsDone will be set to FALSE and xInt and yInt is set to MIN INTEGER or MAX INTEGER respectively PROCEDURE WindowToGraphPoint xInt yInt INTEGER VAR xReal yReal REAL Calculates graph coordinates xReal and yReal from the specified pixel coordinates xInt and yInt of the graph s window see WindowIO Note that the vertical axis of the ModelWorks graph is transformed to yMin 0 0 and
239. mbol will be inheritated by the later associated data arrays So if the monitoring variable s graphing variable is set isY the data are selected to be displayed 2 Since the data arrays symbol CHAR line style LineSTyle and color Stain will be taken from the master monitoring variable call SetCurveAttrForMV and ev SetDefltCurveAttrForMV 3 Declare the associated data arrays with the master monitoring variable the independent monitoring variable and all the data arrays with a call to DeclDispData 4 To enable the display mechanism the monitoring variable mvDepVar must be isY and mvIndepVar must be isX If another monitoring variable represents the current x axis then nothing can be displayed 5 ModelWorks will display automatically all declared data in the normal graph of the Graph window at the specified moment i e typically at InitMonitoring or at TermMonitoring To allow for a general control of the moment of display the procedure DisplayDataNow and DisplayAllDataNow are also exported Caution Be sure to follow the steps given above in the correct order or no data can be declared and displayed Do not assign any values to the master monitoring variable to avoid conflicts with the data declaration Setting writeInTable or writeOnFile of the Master monitoring variable is not prohibited but makes no sence since a dummy value NAN 017 and not the data series will be displayed TYPE Disp
240. meter vector composed of the elements of pp and Pg Eig T8 Schematic representation of a dynamic a continuous time or b discrete time system These systems constitute the basic types used in ModelWorks to describe elementary and structured models Note that the output y defined by Eq 4 2 resp 5 2 does not depend on the input u This restriction guarantees the correct calculation of structured models Structured models are explained below T37 ModelWorks 2 0 Theory In the context of ModelWorks the term output is used in a different than the usual meaning It is reserved to the output produced by a submodel to be connected with the input of another submodel coupling of submodels It should not be confounded with the display of simulation results for the simulationist The latter is called monitoring 4 2 Structured Models Coupling of Submodels Any number of elementary models in this context called submodels may be coupled to form complex structured models Any number of hierarchical levels may be introduced Elementary models are defined exactly the same way as described in the previous chapter The coupling is realized by connecting a submodel s output to another submodel s input There are three cases to be distinguished A all submodels are continuous time systems only B all submodels are discrete time systems only C and there are some continuous as well as discrete time submodels Fig T9 Example of a
241. mic includes the sections Output Input plus Dynamic s str and Terminate The dynamic section is executed according to the chosen time step and simulation time an arbitrary number of times i Unless de faults were changed a simulation session can be fully reset to the original start up conditions Reset all largument passed in call to procedure RunSimMaster ModelWorks 2 0 Theory 5 1 2 a Simulation session The simulation session Fig T16 consists of all activities done by means of the ModelWorks simulation environment during the execution of a typical model definition program It repre sents the topmost level of all simulation tasks and corresponds to the execution of the procedure RunSimMaster from module SimMaster It may be repeated with the same set of models as many times the simulationist wishes but otherwise there are no relationships to other simulation tasks on the same level In particular ModelWorks does not support any communication of data from a simulation session to another one except for the simulation results contained in the stash file Neither does the current version of ModelWorks support the direct reading of the stash file However it is possible to construct a particular model which reads a stash file and declares the models and model objects needed for a post run analysis ModelWorks writes data onto the stash file according to a syntax particularly designed for this purpose When ModelWorks enters a sim
242. modeler ModelWorks assigns the listed default values to the various parameters during start up of a model definition program For all other defaults i e the defaults of models and model objects the modeler is forced to specify them while declaring the models and the model objects in the model defini tion program Predefined values can not be overwritten by the modeler 7The abbreviations stand for dd current day e g 01 for the first day of a month mon current month e g Jan for January yyyy current year e g 1989 hh current hour e g 22 for 10 pm mm current minute e g 04 in 10 04 pm 8On the IBM PC MODELWOR DAT Will be created in the folder where the application resides which has started the model definition program respectively on the IBM PC in the current working directory 9Order of activation of monitorable variables The first variable has the valuei 0 T 62 ModelWorks 2 0 Theory Resets may affect only a single model object all objects of a single model or all objects of all models s a Fig T21 Furthermore resets can be executed for a particular class of model objects only e g only model parameters or only the curve attributes of monitorable variables Unless curve attributes are assigned to the monitorable variables either interactively by changing the current curve attributes in the monitorable variable window or via the client interface by cal ling the procedures SetCurveAttributesForMV or SetD
243. models structured simulations are particularly important The model definition program contains an experiment which asks first the simulationist how many runs the experiment shall encompass suppresses all stash filing and then writes the re sults for further statistical analysis onto a separate data file default name Markov DAT The latter serves to reestimate the coefficients of the Markov matrix and can be directly opened by A 157 ModelWorks V2 0 Appendix many applications for a subsequent statistical data analysis For instance StatView could suc cessfully be used to enter the simulation results by choosing menu command mport and se lecting file Markov DAT plus to compute and analyze the estimates MODULE Markov FREER ERR 2 2 2 2 2 5 2 2 2 2 2 2 2 2 2 202 2 5 2 2 2 2 2 5 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 Modelworks model Markov Copyright 1989 by Andreas Fischlin and Swiss Federal Institute of Technology Zurich ETHZ Department of Environmental Sciences Systems Ecology Group ETH Zentrum CH 8092 Zurich Switzerland Version written for Dialog Machine V2 0 User interface MacMETH V2 6 1 Pass Modula 2 implementation ModelWorks V2 0 Modeling amp Simulation Purpose Simulates a stochastic process defined by a given Markov matrix in order to estimate the matrix from the statistics collected during the simulations Implementation and Revisions Author Date Descriptio
244. n af 18 10 89 First implementation DM 2 0 MacMETH 2 6 ModelWorks 1 3 af 21 10 89 Extended to interactive renaming of states af 03 05 90 Refining of experiment for direct import by StatView 512 statistics program KE RENEE RE EEE RRR ERA EKER KE RREE EERE EERE RE ER ERE EERE RE RERER ERE FROM DateAndTime IMPORT GetDateAndTime DateAndTimeRec FROM WriteDatTim IMPORT DateFormat TimeFormat WriteDate FROM SimBase IMPORT Model IntegrationMethod DeclM DeclSV RTCType DeclP StashFiling Tabulation Graphing DeclMV CurrentStep CurrentTime SetSimTime SetMonInterval SetIntegrationStep TerminateConditionProcedure InstallTerminateCondition InstallStartConsistency StashFileName SetMV GetMV SetP SetDefltP SetDefltMV GetDefltMV LineStyle GetCurveAttrForMV SetCurveAttrForMV ClearGraph GetDefltCurveAttrForMV SetDefltCurveAttrForMV Stain StackWindows TileWindows SetWindowPlace SetDefltWindowPlace MWWindow UseCurWSettingsAsDefault NoInput NoOutput NoTerminate NoAbout FROM SimMaster IMPORT RunSimMaster SimRun DeclInitSimSession ExperimentAborted DeclExperiment WriteTime FROM SYSTEM IMPORT VAL only for inverse of ORD FROM DMSystem IMPORT MenuBarHeight FROM DMStrings IMPORT Concat ConcatCh AssignString FROM DMConversions IMPORT IntToString FROM DMWindowIO IMPORT BackgroundHeight BackgroundWidth FROM DMMenus IMPORT Menu Command AccessStatus Marking InstallMenu InstallCommand In
245. n language since a model definition program is written as a plain Modula 2 program text As a consequence ModelWorks can not automatically sort the statements which compute derivatives Compared with other si mulation software e g ACSL this may be considered to be a draw back However experience shows that automatic sorting of statements is error prone if one models 1 ACSL is a proprietary simulation software program that is leased with restricted rights according to license agreement and terms and conditions by Mitchell and Gauthier Associates Inc USA Concord MA respectively by Rapid Data Ltd Europe Worthing Sussex UK T 14 ModelWorks 2 0 Tutorial complex and ill defined systems Moreover the greater flexibility offered by the host language Modula 2 a modern powerful and formally defined programming language often outweighs the lack of automatic sorting which is mostly not much more than a little inconvenience if the model definition has been carefully worked out before its im plementation Most models maintain tight relationships among their objects such as state variables parameters and auxiliary variables etc The modeler may keep logically connected objects close together by defining related objects local to the model boundary The latter normally coincides with the boundary of the scope of a Modula 2 module Moreover the modeler is free to use any Modula 2 feature he wishes For instance model object
246. n requires an IBM PC or 100 IBM compatible computer either a fast XT Turbo an AT or a PS 2 model equipped with a hard disk and a mouse almost any type of mouse will do The computer should be equipped with 640 KBytes of RAM and should be operated under the MS DOS operating system The resolution of the monitor should be better than 512 x 342 pixels EGA VGA or Hercules are OK CGA and MGA are not For ModelWorks the resolution should not be higher than 600 x 800 640 K limit of DOS leaves not enough memory to store the window bitmaps 1 Upward compatible CPUs are Motorola 68030 or 68040 upward compatible numerical coprocessors are 68882 2 Order ModelWorks Macintosh V2 0 from the following address Projekt Zentrum IDA re ModelWorks Swiss Federal Institute of Technology ETHZ ETH Zentrum CH 8092 Ziirich Switzerland V2 0 Reflex and V2 0 II from the address Systems Ecology re ModelWorks Department of Environmental Sciences ETHZ ETH Zentrum CH 8092 Ziirich Switzerland 20T course the use of the distributed system software is optional and the user may use instead any other system he she wishes given the system version is equal or newer to that mentioned above 4512KBytes might just work with the Dialog Machine alone but not with ModelWorks A 126 ModelWorks V2 0 Appendix Together with a GEM supported mouse the GEM desktop software must be installed in order to execute existing ModelWorks model definition programs To
247. n spring individuals eggs lbm notOnFile notInTable notInGraph DeclMV yt yLL yUL Larval density larvae kg branches Y lbm kg notOnFile writeInTable notInGraph DeclMV ytLn Ln negLogDelta Ln negLogDelta tyUL Ln of larval density larvae kg branches Ln Y lbm kg notOnFile notInTable isY DeclMV def 0 0 1 0 Defoliation def notOnFile notInTable notInGraph DeclP nrt 511147 0 511147 0 511147 0 noRtc nrt number of trees trees trees DeclP cl 0 5728 0 4841 0 6538 noRtc el egg winter mortality cl lbm DeclP c2 0 05112 0 016 0 087 noRtc c2 slope of small larvae mortality vs rf c2 DeclP c3 0 17932 0 565 0 206 noRtc c3 y intercept of small larvae mortality vs rf c3 Decl1P c4 2 25933 nrt 2 4129 nrt 2 1057 nrt noRtc c4 slope of needle biomass vs rf ca Dec1lP c5 67 38939 nrt 62 8076 nrt 71 9712 nrt noRtc c5 y intercept of needle biomass vs rf c5 DeclP c6 0 005472 0 0027 0 0106 noRtc c6 food demand of a large larvae c6 kg lbm DeclP c7 0 124017 0 1070 0 1410 noRtc A 166 ModelWorks V2 0 Appendix c7 slope of large larvae mortality vs rf c7 DeclP c8 1 435284 1 685 1 1855 noRtc c8 y intercept of large larvae mortality vs rf c8 Dec1P c9 0 44 0 363 0 517 noRtc c9 sex ratio c9 DeclP c10 18
248. naeh 88 6 2 2 TO window State variables sss sees 90 6 2 3 JO window Model Parameterssanannnn anne 91 6 2 4 IO window Monitorable variables sss eee eee 93 TA CLIEN Te INTERBA O EEE E E AASE EE E EE 98 7 1 Declaring Models and Model Objects sss sese 99 7 1 1 Running a simulation session nr Bike 99 7 1 2 Detlaration OF models rn ae ea 100 7 1 3 Declaration of state variables case 102 7 1 4 Declaration of model parameters sss sees eee eee 103 7 1 5 Declaration of monitorable variables sese eee eee 104 7 1 6 Declaration of table functions sss sees eee eee 105 7 2 Accessing Defaults and Current Values eee eee 106 7 2 1 Global simulation parameters and project description 107 7 2 1 a Retrieval of read only current Values sss sese eee eee eee 108 7 2 1 b Modification of defaults sss sese 108 7 2 1 c Modification of current values sse eee eee 109 7 2 2 Installed models and model objects an aa 110 7 2 2 a Modification of defaults na a ra 110 7 2 2 b Modification of current values aa ae gen 111 ModelWorks 2 0 7 2 2 c Inter and extrapolation with table functions sss sees eee eee eee 112 7 3 Removing Models and Model Objects sss sss 113 7 4 Simulation Control and Structured Simulation Runs sss sees sees eee eee 113 W Display and Monitorme 2 BEN BIER 115 PD W NOW operati ON era 115 722 General TOM WO IN reinlesen 117 239 Stash Hinspiel ren 118 7 5 4 Graphical monitoring sss sees eee eee eee eee 118 7 5 5 Simulation environm
249. nce the following information should be easy to interpret also for a non standard ModelWorks version Reading Hint If there is a functional difference to the standard version the fact will be stated in a phrase within brackets with a font similar to the following Not available in Reflex and PC version 6 1 Menus and Menu Commands This section explains all menu commands in detail Many and often used menu commands can also be invoked by using the keyboard instead of the mouse For easier remembering the keys to be used for the keyboard shortcuts are shown to the right of the command text Fig R1 Keyboard shortcuts or so called keyboard equivalents are entered by pressing the command key clover leaf key simultaneously with another key X Reading Hint Throughout this reference manual such keyboard equivalents are abbreviated as X The following keyboard commands are globally available in the simulation environment In all entry forms the simulationist may press the key Return or Enter instead of clicking into the push button OK Pressing the keys or Escape is equivalent to the clicking into the push button Cancel The latter two keyboard equivalents may also be used to stop a simulation run Stop Kill run Pressing the key tab in an entry form allows to move to the next edit field plus to fully select its content In some edit fields the key combination Shift tab is available to move backwards from field to field e g in
250. ndGen SetPars mu stdDev REAL PROCEDURE GetPars VAR mu stdDev PROCEDURE N REAL PROCEDURE ResetN do always call defaults U REAL N u stdDev call after SetSeeds for full reset of N 0 stdDev 1 TN RR RRR e e e e IRR IR IIR I Kk ReadData FROM DMStrings IMPORT String FROM DMFiles IMPORT TextFile VAR dataF TextFile PROCEDURE OpenADataFile VAR fn ARRAY OF CHAR VAR ok BOOLEAN PROCEDURE OpenDataFile VAR fn ARRAY OF CHAR VAR ok BOOLEAN PROCEDURE ReReadDataFile performs a reset PROCEDURE CloseDataFile PROCEDURE SkipHeaderLine PROCEDURE ReadHeaderLine VAR labels ARRAY OF String VAR nrVars INTEGER assign NIL to labels before first use PROCEDURE ReadLn VAR txt ARRAY OF CHAR PROCEDURE GetChars VAR str ARRAY OF CHAR PROCEDURE GetStr VAR str String PROCEDURE SkipGapOrComment skips lt and PROCEDURE ReadCharsUnlessAComment VAR string ARRAY OF CHAR Used for error messages only desc describes the item to be read loc is location e g a line where error was encountered PROCEDURE GetInt desc ARRAY OF CHAR loc INTEGER VAR x INTEGER min max INTEGER PROCEDURE GetReal desc ARRAY OF CHAR loc INTEGER VAR x REAL min max REAL values used if measurement is missing PROCEDURE SetMissingValCode missingValCode CHAR default N PROCEDURE GetMissingValCode VAR missingValCode CHAR
251. ndow set up and functions operators on the selected elements operands Two buttons are common to all windows The button activates an entry form where the columns to be displayed in the object field can be selected The button S serves to select all objects of the particular kind shown in the IO window e g all parameters of all models Scope All in Fig T21 While ModelWorks is executing a button function the button picture will be shown inverted In an inactive IO window no button functions can be activated The simulationist can recognize this status if clicking on buttons does not invert them Third on the right side there are the scrollers to scroll lines individually or whole pages of the object list field up and down in case that the window is too small to show all objects at once 3 During the actual scrolling the button picture will be shown inverted In an inactive IO window no scrolling can be done The simulationist can recognize this status if clicking on scrollers does not invert them The last group of elements are not specific to ModelWorks but are general and may be present in any window the title bar to move the window 4 the close box to close it 5 the zoom box to enlarge it to the size of the screen or back 6 and the grow box to change the size of the window to any shape 7 5 1 5 PREDEFINITIONS DEFAULTS CURRENT VALUES AND RESETTING ModelWorks maintains for many values such as global s
252. ndows subdivided into three fields In the middle the object list 1 on the upper left corner the palette of button commands 2 and on the upper right corner the scrollers to scroll the items in lists too large to show all items at once 3 s a text All IO windows have a common structure The content area of any IO window is subdivided into three fields Fig T20 First the field in the middle of the window contains a list of ModelWorks objects 1 Its title line 1a displays the headers of the columns currently in use which describe display and designate ModelWorks objects and their values Below there is the T 58 ModelWorks 2 0 Theory actual list of the objects e g the parameters which have been installed in ModelWorks by the model definition program The order follows the declaration order in the model definition program and objects belonging to the same model appear together under the bold title of the corresponding model 1b An object in the list can be selected as an operand by a mouse click on the corresponding line which is confirmed to the simulationist by inverting the line 1c The selection of the bold model title is interpreted as the selection of all objects belonging to this model From this follow scopes and scope rules for the selection of operands For the O windows containing the state variables model parameters and monitorable variables exist the following scopes all objects of a particular
253. nly at every coincidence point Fig T10 By definition the coincidence interval c is a positive integer and must have a size of at least 1 In the context of simulations it is meaningful to match the values of the two time variables thus at every coincidence point it must hold t k This requirement is guaranteed if the following conditions are satisfied to k and tend Kf where ko kp E 3 11 T39 ModelWorks 2 0 Theory Any structured model mixed of continuous and discrete time submodels can be subdivided into two portions the first is the continuous time portion Z consisting of the set of all continuous time submodels with their related continuous time inputs and outputs plus the continuous time global input and output the second is the discrete time portion A consisting of the set of all discrete time submodels with their related discrete time inputs and outputs plus the discrete time global input and output At every coincidence point the system is fully defined and all submodels are fully coupled System E A At these points the real time t is mapped to the discrete time k as follows k INT t where INT yields the integral part of its argument 12 Between coincidence points the dynamics of the system collapse or degenerate to the continuous time portion amp the other portion of the system A remains constant but is still accessible to This corresponds to a sample and hold technique sample at coincidence points h
254. nvironment and secondly general aspects of the model development process Any serious modeling with ModelWorks requires to read at least this part of the manual and the section on the client interface of the manual Part III Reference Reading Hint For easier orientation the pages figures and tables of Part II Theory are prefixed with the letter T Within this part the numbers of figures and tables follow those used in Part I Tutorial T35 ModelWorks 2 0 Theory 4 Model Formalisms This chapter deals with theoretical aspects of modeling which are used in addition to the standard knowledge when developing models with ModelWorks Please refer to a textbook for a general introduction to the modeling and simulation of dynamic systems In particular this chapter explains the mathematical formalisms in which the modeler should describe ModelWorks models ModelWorks distinguishes between two model types Elementary models and structured models Structured models are composed of several coupled elementary submodels 4 1 Elementary Models The elementary models which are used in ModelWorks are discrete or continuous time dynamic systems They are formally described by a set of ordinary first order differential or difference equations Generally model parameters are considered to be time invariant but ModelWorks supports time variant parameters too However it is recommended to treat them either as auxiliary variables becoming part of the diffe
255. o the user by the so called system behavior monitoring concept Values of any variable may be written onto a file for future reference written into a table or displayed as curves in line charts All data can be reset to a given default value Further the model s data structure are all stored dynamically This allows the user to install an unlimited number of models of an arbitrary size with an arbitrary number of variables each up to the limits of the hardware ModelWorks simulation environment is based on the Dialog Machine guaranteeing a consistent user interface and has originally been implemented using MacMETH3 a fast and efficient Modula 2 language system for the Apple Macintosh computer WIRTH et al 1988 ModelWorks simulation environment runs on any machine on which the Dialog Machine is available If this is the case an efficient and smooth port of ModelWorks in a few days work is possible Currently ModelWorks is available for Macintosh computers with at least 512 KBytes of memory RAM plus at least two floppy drives and IBM PCs which run under MS DOS and have 640 KBytes of memory RAM plus a hard disk For more details on particular implementations and hard plus software requirements for specific versions see the Appendix This text serves as a manual for the ModelWorks software Since all versions are very similar and differences are the exceptions there exists only this one text Differences between versions are minor and
256. odelWorks rejects any attempt to enter a value out of the range as defined by the modeler in the model definition program If a multiple selection of state variables has been made not only one but a series of entry forms will be offered one form for each state variable This sequence can be terminated by pressing the push button Cancel Note however that in the latter case all changes which have been made to state variables before can no longer be reversed Only eventual changes made to the state variable currently in display will be discarded E Init Reset initial value Resets the initial values of the currently selected state variable s to their default values 6 2 3 IO WINDOW MODEL PARAMETERS The IO window Model parameters displays information name identifier unit value change at run time enabled disabled about the installed model parameters of all models Fig R20 Furthermore it offers a mechanism to select model parameters and to execute functions which operate on the current values of the selected model parameters by clicking on the buttons in the window R91 ModelWorks 2 0 Reference Model Parameters Parameter names Ident Unit Value Potato model kGrowth gdwplant 50 000 kLeat 1 000 kStern 6 000 kRoot 1 000 kTuber 6 000 Weather input module Weather input source Inp Fig R20 IO window Model parameters showing the list of all model parameters and offering a palette of button functions operating on the
257. odula 2 implementation 2 On the IBM PC you have to start first the GEM desktop and then to start the application GRASSAPH APP T32 ModelWorks 2 0 Tutorial In addition to the basic techniques you have just learned ModelWorks features many more advanced modeling and simulation techniques Among the more important features are modular hierarchical modeling including the coupling of several models and the mixing of discrete time with continuous time models With ModelWorks it is easy to analyze results of complex simulation studies by means of a sensitivity analysis or a parameter identification Moreover thanks to its architecture open for extensions it allows for an unlimited number of possibilities For a complete full and detailed description of all of ModelWorks features please refer to the parts Theory and Refe rence of this text 2608 4 4608 4 baba 2006 H 106002 n Curves Minimum Has imum Unit A 4 488 1584 HEE g dry weight m 2 Fig T6 Graph of the simulation results produced with ModelWorks simulating the new developed sample model The graph shows a state space representation of a Lotka Volterra like grass aphids model system In case you would like to continue with the introductory example here some suggestions how you could possibly explore it further on your own introduce an auxiliary variable for the total biomass b t b t G t A t 3 Declare b t as a monitorable variable and compute its v
258. of an IO window This control is only available to the modeler but not to the simulationist PROCEDURE DisableWindow w MWWindow DisableWindow disables the IO window w for any usage i e neither the opening editing nor the closing of the IO window by the simulationist is any more possible Only supported for the IO Windows In case the IO window w should be currently open it is closed The corresponding menu commands are disabled dimmed PROCEDURE EnableWindow w MWWindow EnableWindow reverses the effect of DisableWindow and enables the subsequent usage of the IO window w by the simulationist to its normal and full functionality PROCEDURE UseCurWSettingsAsDefault Copies all the current settings of the windows to their defaults A typical usage may be TileWindows SetWindowPlace MIOW SetWindowPlace SVIOW SetwindowPlace GraphW UseCurWSettingsAsDefault The calls to SetWindowPlace customize particular locations of windows The final call to UseCurWSettingsAsDefault saves all the current settings as the new defaults to be used if the simulationist performs a reset of the windows This solution is more convenient than having to specify first the current values and then the defaults again with exactly the same values 7 5 2 GENERAL MONITORING After having called SuppressMonitoring all subsequent monitoring will be suppressed PROCEDURE SuppressMonitoring The next procedure PROCEDURE ResumeMonitoring R 117
259. of the dynamics of the system This is different from conventions in systems theory but this solution has been favored over other possible designs due to its flexi bility and convenience Since continuous time measurement would result in an exorbitant amount of data actual moni toring is possible only at discrete points in time the monitoring times They are equidistant and the time interval between monitoring times is constant and global i e the same for all models and all kinds of monitoring the so called monitoring interval ha ModelWorks computes val ues exactly at the time points ta for which monitoring is requested that is tm t hy to simulation start time i 0 1 2 T55 ModelWorks 2 0 Theory Predefined standard monitoring of ModelWorks is available in one or any combination of the following three kinds Values typically simulation results may be written and stored on a so called stash file for later usage tabulated as numbers in a table or shown as curves in a graph The stash file can store an arbitrary amount of information on models model objects and their values and it is usually produced for further numerical e g statistical analysis of the simulation results or for future report generation to document simulation runs in all details The size to which the stash file may grow is limited only by the available disk space The file is written in a formally defined syntax and contains several types of
260. of the model definition program and actually knows almost nothing about anything the modeler does in his program The only objects ModelWorks cares about are models state variables to be integrated numerically model parameters to be changed interactively during a simulation session and monito rable variables for the monitoring of the simulation results Hence they are the only objects which have to be made known i e declared to ModelWorks The link of models and their model objects to ModelWorks is achieved via the client interface In its essence it consists of two library modules SimBase and SimMaster These modules provide all Modula 2 objects types and procedures needed to describe a model in the model definition program Executing a ModelWorks model definition program means to start first the simulation environment Running the simulation environment is called a simulation session At the begin of a simulation session ModelWorks initializes the simulation environment and normally executes all model and model object declarations as programmed by the modeler It then performs a reset of all current values using all the defaults specified during the declarations Subsequently ModelWorks is ready to execute commands en tered by the simulationist such as a simulation run the execution of a simulation expe riment or the editing of the current values e g of a model parameter or an initial value ModelWorks is not just another simulatio
261. of their installation order see also chapter Model formalisms The calculation order which meets all these requirements is shown in Fig T18 First the Output procedures of all models then all Jnput procedures are called Thereafter the numerical integration can be done Depending on the integration algorithm the procedure Dynamic will be called once or several times for the evaluation of the derivatives Every sub model is integrated as an independent unit Therefore it is possible to integrate different sub models with different integration algorithms This can be of interest if some models are numer 4This level could also be understood as a structured simulation experiment with k 1 see above T53 ModelWorks 2 0 Theory ically less stable than others However the same integration step length is used for all models to guarantee a coordinated data transfer between submodels T 54 At begin of time step Input Output Stat var Aux var tia tia ti UH Calculation of output ta toren Calculation of input Output Stat var Aux var ti ti tiji Integration tum o tio Tt S tum Stiti Update Fig T18 Calculation of time input output state and auxiliary variables during an integration step The calculation order guarantees that all calculations are based on valid values which have been calculated in a previou
262. ogel Wak 2 by calling SetGlobSimPars 2000 0 2100 0 or S by Simulationist by menu command Set global simulation parameters under M by Modeller menu Setting Eig T22 Relationship between default and current values and the reset mecha nism of ModelWorks For most values ModelWorks maintains two copies One is the default the second is the current value which is actually used for simulations During a reset also executed at program start up the default values are copied assigned to the current values Interactive modifications editing of values during the simulation session affect only the current values Via the client interface it is possible to change the defaults as well as the current values Note that as long as the simulationist remains within the standard simulation environment of ModelWorks a reset resumes the initial program state which existed at the very begin of the simulation session This is because in contrast to the client interface it is not possible to access and modify defaults via the user interface However if the modeler by using the client interface has programmed extensions which also allow to change interactively the defaults the reset mechanisms provided by ModelWorks will no longer allow the simulationist to reset this initial start up condition Instead the state as defined by the last default specifications is resumed However if the client has defined a simulation session initialization procedure Mod
263. ogram describes the actual model to be simulated Both units form together the final simulation program They are linked by procedures provided in the client interface and by mutual data exchange Ranges for initial values of state variables or model parameters are defined solely by the modeler They become effective only in the simulation environment while the simulationist edits the current values of these model objects ModelWorks guarantees that the simulationist assigns only values to an initial value of a state variable or a model parameter which lie within these ranges Hence the modeler can use this mechanism to enforce limits within which the model equations are still valid in order to reduce the danger that the simulationist runs a meaningless simulation experiment or encounters a fatal error condition However the clipping ranges for monitorable T 13 ModelWorks 2 0 Tutorial variables behave differently and should not be confounded with range limits The simulationist can change clipping ranges interactively anytime ModelWorks has been designed to make modeling as easy as possible yet as powerful and flexible as possible Hence for ModelWorks a model is a variable not predefined portion of a simulation program which has been left out so that the modeler may define it at a later time Fig T3 A user of ModelWorks wishes to define freely this open portion according to his current needs for instance by specifying a new set of c
264. old between Fig T10 Discrete system x Continuous system Cont time scale Disc time scalk k oO l 5 e Coincidence time points coincidence interval c Fig T10 Coupling of discrete and continuous time submodels The figure shows the results of a simulation of a structured model system composed of one discrete and one continuous time submodel A communication between the two submodels occurs at every coincidence time point when the output of the discrete submodel determines the rate of change of the continuous time submodel No data exchange takes place during the coincidence interval during which the rate of change of the discrete time submodel remains constant sample and hold The global input consists of two vectors one for the global continuous time input u and the other for the global discrete time input u um Il u t E global inputs 6c l u k EA At the coincidence points the inputs of all submodels i depend on the continuous as well as the discrete time global input u resp u gt and on the output of the continuous time submodels j E Z as well as the discrete time submodels EA i E P E 1 2 n Between coincidence points the inputs to the continuous submodel i Z depend continuously on the continuous time global input u and on the output of the continuous time submodels it Any T 40 ModelWorks 2 0 Theory dependence of the continuous t
265. ommand ch CHAR PROCEDURE InstallSeparator m Menu s Separator PROCEDURE RemoveSeparator m Menu s CARDINAL PROCEDURE InstallQuitCommand s ARRAY OF CHAR p QuitProc aliasChar CHAR PROCEDURE UseMenu m Menu PROCEDURE UseMenuBar PROCEDURE RemoveMenu VAR m Menu PROCEDURE RemoveMenuBar PROCEDURE EnableDeskAccessories PROCEDURE DisableDeskAccessories PROCEDURE EnableMenu m Menu PROCEDURE DisableMenu m Menu PROCEDURE EnableCommand m Menu c Command PROCEDURE DisableCommand m Menu c Command PROCEDURE CheckCommand m Menu c Command PROCEDURE UncheckCommand m Menu c Command PROCEDURE SetCheckSym ch CHAR PROCEDURE ChangeCommandText m Menu c Command newCmdText ARRAY OF CHAR PROCEDURE ChangeAliasChar m Menu c Command newCh CHAR PROCEDURE ExecuteCommand m Menu c Command VEREINE I ER ARN HR EIR ERK amp DMStorage PR MIR IB Re MER Re IER K TR eK MII BR He IK HHH KR TR GN RAE PROCEDURE Allocate VAR p ADDRESS size LONGINT PROCEDURE AllocateOnLevel VAR adr ADDRESS size LONGINT level INTEGER PROCEDURE Deallocate VAR p ADDRESS PROCEDURE ALLOCATE VAR p ADDRESS size CARDINAL IBM PC DM compatibility PROCEDURE DEALLOCATE VAR p ADDRESS size CARDINAL IBM PC DM compatibility TSR ee ee IE E E AA E LE SeC IE E He ee SESE HEE DMStrings ee BeBe Fe He He RHE MN TIER RA IER eK SE He RR IER RIK MRK Key TYPE String PROCEDURE NewString VAR s ARRAY OF CHAR String PROC
266. on diskettes in directory MW SAMPLES Depending on the actual ModelWorks version you are using minor changes to the listings given here might be neces sary This is in particular the case for the PC version and the following sample models Markov MOD DM PC 1 5 provides a random number generator in DMMathLib H 1 THE SAMPLE MODEL LOGISTIC GRASS GROWTH LOGISTIC MOD The following program code contains the sample model described in the manual Part I Tutorial in the section Getting started with the simulation environment especially the subsection Simulating the sample model MODULE Logistic mu 9 4 88 Version used in ModelWorks manual F He 2 2 5 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 Logistic grass growth Frk kk kkk 2 2 2 2 2 2 2 2 2 2 2 22 2 2 2 22 FROM SimBase IMPORT Model IntegrationMethod DeclM DeclSV DeclP RTCType StashFiling Tabulation Graphing DeclMV SetSimTime NoInitialize NoInput NoOutput NoTerminate NoAbout FROM SimMaster IMPORT RunSimMaster VAR m Model grass grassDot cl c2 REAL PROCEDURE Dynamic BEGIN grassDot cl grass c2 grass grass END Dynamic PROCEDURE Objects BEGIN DeclSV grass grassDot 1 0 0 0 10000 0 Grass G g dry weight m 2 DeclMV grass 0 0 1000 0 Grass G g dry weight m 2 notOnFile writeInTable isY DeclMV grassDot 0 0 500 0 Grass derivative dG dt g dry weight m 2 time notOnFile notInTable notInGraph Decl
267. on of standard world time or Greenwich time There is a modified Julian Date M J D in use today much used in space travel which starts at 17 Nov 1858 00h00 00 24 00 000 5 J D Programming Design A Fischlin 24 09 89 Implementation A Fischlin 24 09 89 Swiss Federal Institute of Technology Zurich ETHZ A 147 ModelWorks V2 0 Appendix Department of Environmental Sciences Systems Ecology Group ETH Zentrum CH 8092 Zurich Switzerland Last revision of definition 24 09 89 af He ke Fe Fe Fe He He Fe e Fe Fe Fe e 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 5 2 2 5 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 CONST Jan 1 Feb 2 Mar 3 Apr 4 Mai 5 Jun 6 Jul 7 Aug 8 Sep 9 Oct 10 Nov 11 Dec 12 Sun 1 Mon 2 Tue 3 Wed 4 Thur 5 Fri 6 Sat 7 PROCEDURE DateToJulDay day month year INTEGER LONGINT PROCEDURE JulDayToDate julday LONGINT VAR day month year weekday INTEGER Convert between a julian day and an ordinary calendar date Note Only valid for dates between 1 Jan 1949 and 31 Dec 2036 PROCEDURE LeapYear yr INTEGER BOOLEAN PROCEDURE SetCalendarRange firstYear lastYear firstSunday INTEGER This procedure allows to set the calendar range for which the algorithms of this module shall work They work correctly from the date 15 0ct 1582 onwards for the next 3333 years and given the following restrictions are satisfied The first y
268. onSetCol BOOLEAN fw dec INTEGER END RECORD END CASE END RECORD The booleans determine whether a column is to be displayed or not they correspond to the check boxes which may be set by the simulationist in the entry form which is activated when the IO window button Ser Up is clicked The integers specify the format in which to display real numbers where fw is the field width and dec is the number of decimal digits R 116 ModelWorks 2 0 Reference PROCEDURE SetIOWColDisplay mww MWWindow wd IOWColsDisplay SetIOWColIDisplay allows to set a new setup of the columns and new display formats in the IO window mww The predefined default of the simulation environment is 3 decimal digits to display or parameter values this procedure allows to alter this format to any other value PROCEDURE GetIOWColDisplay mww MWWindow VAR wd IOWColsDisplay GetIOWColDisplay returns the setup of the columns and the display formats currently in use by the IO window mww Instead of the current values the following two procedures affect the default values otherwise they function the same way as the previous two procedures PROCEDURE SetDefltIOWColDisplay mww MWWindow wd IOWColsDisplay PROCEDURE GetDefltIOWColDisplay mww MWWindow VAR wd IOWColsDisplay The following procedures allow the modeler to customize the simulation environment even a step further he she may even completely disallow or allow any usage
269. oncat descr istr Concat descr istr ident f IntToString ORD is 1 istr 0 Concat ident istr ConcatCh ident IntToString ORD js 1 istr 0 ConcatCh ident SetDefltMV m f is js 0 0 1 0 descr ident writeOnFile writeInTable isY END FOR END FOR END AssignStatesNames Concat ident istr PROCEDURE DefineMarkov CONST lem 3 tabl 25 tab2 35 tab3 45 VAR ef FormFrame ok BOOLEAN cl INTEGER BEGIN cl 2 WriteLabel cl lem Define Markov Process WriteLabel cl lem State WriteLabel cl tabl Transition probabilities INC cl StringField cl lem 15 nameStateOne useAsDeflt RealField cl tab1 7 p one one useAsDeflt 0 0 1 RealField cl tab2 7 p one two useAsDeflt 0 0 1 RealField cl tab3 7 p one three useAsDeflt 0 0 INC cl StringField cl lem 15 nameStateTwo useAsDeflt INC cl A 161 ModelWorks V2 0 Appendix RealField cl tabl 7 p two one useAsDeflt 0 0 1 0 RealField cl tab2 7 p two two useAsDeflt 0 0 1 0 RealField cl tab3 7 p two three useAsDeflt 0 0 1 INC cl StringField cl lem 15 nameStateThree useAsDeflt RealField cl tabl 7 p three one useAsDeflt 0 0 RealField cl tab2 7 p three two useAsDeflt 0 0 RealField cl tab3 7 p three three useAsDeflt 0 INC cl ef x 0 ef y 1 display entry form in middle of screen ef lines cl 1 ef columns 55 UseEntryForm ef ok IF ok THEN AssignStatesNames nameStat
270. or statistical analysis or just to document a simulation run At a time ModelWorks uses just one stash file only The model definition program of the sample model declares a second monitorable variable beside the state variable Grass This is the derivative of grass listed in the IO window Monitorable variables with the name Grass derivative However the defaults specified by the modeler for this variable are such that it is not displayed unless the simulationist activates it for actual monitoring To see what the curve of the derivative looks like bring the window for monitorable variables to the front select Grass derivative and click on the button Z Toggle function In the column Monitoring appears a Y in the row for the monitorable variable Grass derivative and the legend of the graph is accordingly updated The values of the variable Grass derivative will be drawn as another curve in the line chart of the window Graph during the next simulation run Running another simulation displays the two curves Grass respectively Grass derivative You may generate also other graphs e g Grass derivative versus Grass Select the monitorable variable Grass in the window for the monitorable variables window and click onto the buttons iX and then Z In the column Monitoring disappears first the Y and then appears a X in the row for the monitorable variable Grass This means that the values of the variable Grass will no longer be shown on the y axi
271. osed of two subsystems left onto a Modula 2 model definition program right The outputs are exported by the definition modules DEF and imported by the implementation modules MOD of the other submodel The program module ModelMaster links both submodels by importing and executing the submodel declarations All modules together form the model definition program 5 2 3 MODEL INSTALLATION ModelWorks allows to install any number of models and model objects The actual limitations are not inherent in the software but are only given by the available computer resources i e the currently available heap space and the computing power needed to numerically solve the models The body of the main module from the model definition program calls the procedure RunSimMaster from module SimMaster Typically the actual definition of the models and model objects is indirectly executed by RunSimMaster It takes place in a procedure which is not called by the model definition program itself but instead is passed to ModelWorks as an actual parameter while calling procedure RunSimMaster Executing this procedure will result in the loading and linking of the model together with the eventual set of submodels in the ModelWorks simulation environment Tutorial Fig T3 From within the standard simulation environment once installed neither the number of models nor the number of model objects may be changed without going through a full development cycle Fig T16 T
272. ould have changed already any settings up to this point reset it first with the menu command Settings Reset All above before continuing with this guided tour 2 2 1 DEFAULT SIMULATION You can immediately start a simulation experiment run because ModelWorks ensures that any valid model definition program contains all necessary data for the so called default simulation run Choose the menu command Simulation Start run to actually start the simulation The graph and the table windows are automatically opened and a small time display window appears in the upper right corner Now ModelWorks integrates the differential equation and displays simultaneously the results in the graph and table windows In the graph window you see how the grass grows at the beginning exponentially and how it reaches finally its equilibrium density Fig T5 T18 ModelWorks 2 0 Tutorial Ed Settiines Windows RII EUT Models State variables Model parameters Monitorable variables Monitorable variable names Ident Unit Monitoring ES al x RE BEHE Logistic grass growth model Y F l kee 4 16 25000000 16 50000000 28 8 25 48 Minimum Hax imum Unit 9 900 10606 900 g dry weight m 2 Fig T5 ModelWorks simulation environment during a simulation run of the logistic grass growth sample model In addition to the IO windows the table plus graph windows are currently open The current time is displayed in the upper right corner The graph wind
273. oupled differential equations The modeler does it by writing simple Modula 2 statements which are to be filled in and linked to the remaining constant parts of ModelWorks This is similar to a key which fits into a matching hole of a lock only the two together rendering the lock into a fully functional unit With the model definition program the modeler provides the missing key The key must conform to certain rules in order to fit into the hole However in all other aspects this analogy breaks down since a key is not constructed before each use anew or must not be extended or has not his own particular functionality the latter are all typical properties of ModelWorks model definition programs The remaining parts of the simulation environment i e the actual ModelWorks can not be modified and constitute the preprogrammed ModelWorks software They are general and hence common to any simulation program and resemble the lock with a hole for the key When the simulationist starts a model definition program containing a model definition the latter is inserted automatically into the hole of ModelWorks and what results is a fully functional simulation program Fig T3 Technically a model definition program is a simple Modula 2 program module Its main purpose is to define declare your model and its model objects thus preparing the data exchange needed for simulation sessions ModelWorks does not care how the modeler organizes the structure
274. ourth order one step integration method with constant integration step Runge Kutta 4 5th order variable step length 4th Sth order variable step length Runge Kutta Fehlberg method ATKINSON amp HARLEY 1983 ENGELN M LLGES amp REUTTER 1988 The local error is estimated comparing the 5th order with the 4th order result Depending on the error estimate the step length is increased or reduced to obtain optimal results in terms of efficiency and accuracy This method is most useful for solving models with time constants which strongly vary during the course of a simulation Not available in Reflex and PC version The number of times the procedure Dynamic is called is equal to the order of the integration method Be aware that despite their higher computing load high order methods are often more efficient than those of lower order This is because higher order methods allow for larger integration steps while retaining the same accuracy The latter may result in a reduction of the total number of times the procedure Dynamic has to be computed However since the actual R 89 ModelWorks 2 0 Reference performance depends also on the characteristics of the model the best method for a particular model has often to be identified by numerical experiments E Reset integration method Resets the integration method of the currently selected model s to their default E Init Reset initial values Resets all initial values of the state vari
275. oveOriginTo u Window x0 y0 INTEGER PROCEDURE SelectForOutput u Window PROCEDURE CurrentOutputWindow Window TYPE PaintMode replace paint invert erase Hue 0 359 GreyContent light lightGrey grey darkGrey dark Saturation 0 100 Color RECORD hue Hue greyContent GreyContent saturation Saturation END PatLine BYTE Pattern ARRAY 0 7 OF PatLine VAR pat ARRAY light dark OF Pattern black white red green blue cyan magenta yellow Color PROCEDURE SetMode mode PaintMode PROCEDURE GetMode VAR mode PaintMode PROCEDURE SetBackground c Color pat Pattern PROCEDURE GetBackground VAR c Color VAR pat Pattern PROCEDURE SetColor c Color PROCEDURE GetColor VAR c Color PROCEDURE SetPattern p Pattern PROCEDURE GetPattern VAR p Pattern PROCEDURE IdentifyPos x y INTEGER VAR line col CARDINAL PROCEDURE IdentifyPoint line col CARDINAL VAR x y INTEGER PROCEDURE MaxCol CARDINAL PROCEDURE MaxLn CARDINAL PROCEDURE CellWidth INTEGER PROCEDURE CellHeight INTEGER PROCEDURE BackgroundWidth INTEGER PROCEDURE BackgroundHeight INTEGER PROCEDURE GetEOWAction u Window VAR action PROC PROCEDURE SetEOWAction u Window action PROC PROCEDURE RedrawContent PROCEDURE EraseContent PROCEDURE SetClipping cr RectArea PROCEDURE GetClipping VAR cr RectArea PROCEDURE RemoveClipping TYPE WindowFont Chicago Monaco Geneva NewYork FontStyles
276. ow shows the growth curve of the grass g m 2 2 2 CHANGING INITIAL VALUES Initial values can be changed in the window for state variables bring the state variable window to the front click on it or choose the command Windows State variables and select the state variable Grass Click on the button and change in the appearing entry form the initial value to 4 0 This means that the grass starts growing at a higher density Verify this in another simulation run the maximal density remains the same Use other initial values to explore the model s behavior e g 200 0 1 If you want to enter a initial value out of the allowed range 0 10 000 the program will refuse to accept it For instance try to enter 10001 or 1 and see what happens After your explorations reset the initial value with the menu command Settings Reset All model s initial values or with the button eal 2 2 3 CHANGING PARAMETERS Model parameter values can be changed in the window for model parameters in the same manner as described for initial values Clear the graph with the menu command Windows Clear graph and perform a simulation run for reference purposes What will happen if you increase the growth rate c of the grass Faster growth or higher maximal density Increase the growth rate from 0 7 to 1 2 and perform a simulation run Now the population grows faster and reaches a higher equilibrium value In the table output you can see the maximum value the grass
277. pWatch OBM no checks These files are different implementations of module DMHeapWatch one watches the heap management the other does not The module is actually just used internally of the Dialog Machine but depending on the implementation you use the software will behave slightly different Implementation a tests whether allocations of memory blocks in the heap are exactly matched by deallocations In case a mismatch is de tected a warning message is displayed at the end of the program Implementation ignores the heap management Normally it is recommended to use implementation but in case you suspect heap problems you may gain insights using implementation a Configure your installation by deleting the already existing file DMHeapWatch OBM it is actually a copy of file a in the folder and renaming the copy of any of these files to DMHeapWatch OBM will result in the corresponding heap watching behavior of any Dialog Machine program e g ModelWorks e Bundling of text files produced by all ModelWorks software compiler module DMF iles etc with a particular editor application File FileSystem OBM in the folder M2MiniLib on diskette ModelWorks 1 2 comes in two versions a FileSystem OBM MEdit B FileSystem OBM orig 2 6 If you double click on a text document produced with ModelWorks e g ModelWorks DAT the bundled text editor will be opened For file a this is MEdit Macro editor an excellent shareware editor
278. parameter c This requires that the modeler declares the parameter c with the following call DeclP c1 0 1 ModelWorks will keep a copy of the object s default values in order to be able to reassign them to the current values if a reset is requested by the simulationist Executing a ModelWorks command such as Reset all model s parameters in a simulation session while running above example will then assign 0 1 to the variable c Fig T22 regardless of what the current value of c might be T 60 ModelWorks 2 0 Theory Interactive modifications of values from within the standard simulation environment by using entry forms or the IO windows affect always only the current values not the defaults Calling a reset function from ModelWorks will then reassign the default values to the current values All current values affected by the reset e g all initial values of a particular model will then be set to their defaults as have been defined by the modeler via the client interface Fig T22 Assignments E g to tend MW Defaults from Model Works Defaults to tend 0 0 100 0 a L 7 R Defaults from a S AA cK S S E RSIS odel Definition 1 1989 0 2000 0 MW S M Reset 1989 0 2000 0 Assigns Defaults to current values Current values MW S M Edit or Set EARE 2000 0 2100 0 Modify Current values MW M Simulations 1 by calling SetDefltGlobSimPars 1989 0 2000 0 Mw by M
279. potted emerald pink C7 spotted turquoise gold invisible purge O symbol Fig R25 Entry form opened by the button in the IO window Monitorable variables The following curve attributes are available A color with which curves are drawn on color screens printed on color printers or exposed on slide recorders Line styles affect the style with which connecting lines between points are drawn Points can be emphasized by drawing plotting symbols The colors are coal black snow white ruby red emerald green sapphire blue turquoise cyan pink magenta and gold yellow The graph monitoring procedure produces line charts i e points are connected with lines which can be drawn with one of the following styles unbroken broken dashSpotted erer spotted gt s invisible no drawing at all may be used to stop drawing of a particular curve while others are still drawn purge used to erase already drawn curves R 96 ModelWorks 2 0 Reference autoDefStyle line style will be determined by ModelWorks according to the automatic definition mechanism of curve attributes Any character can be used as a plotting symbol for instance using the plotting symbol results in curves like Note that using a blank will result in the drawing of no plotting symbol at all If automatic definition of curve attributes is active for one or several monitorable variables ModelWorks follows a
280. press the space bar to clear the symbol and hit return Now click on Grass and activate it with button Rerun the simulation and note that you can no longer separate the two curves on a printer or a monochrome screen Of course it is also possible to switch back to the automatic assignment strategy Select Grass and click the button Zl in the appearing entry form click into the top most radio button automatic definition of curve attributes and close it by pressing the enter key or alternatively reset the curve attributes of variable Grass with the button or with the command Reset All model s curve attributes under menu Settings Finally you can learn how to monitor the values of the variable Grass derivative in tabular form Bring the window for monitorable variables to the front select Grass derivative and click on the button Toggle function In the column Monitoring appears a T in the row for the monitorable variable Grass derivative This means that the values of the variable Grass derivative will be written into a column of the table in the window Table during the next simulation run Bring the table window to the front enlarge it till you see all columns and rerun the simulation Before continuing reset this time the table and graph monitoring plus all curve attributes to their defaults with the following method Bring first the window Models to the front select the row containing the model title Logistic grass growth model and
281. procedures used to declare the models and their objects The types Model IntegrationMethod StashFiling Tabulation Graphing RTCType are needed in order to declare the model objects All these objects i e models state variables model parameters monitorable variables auxiliary variables and all associated variables like derivatives initial values and default values are typically declared locally to the program module boundaries VAR m Model grass grassDot cl c2 REAL m is a variable of the opaque type Model It allows to reference the whole model The logistic growth model has one state variable and two parameters ModelWorks requires that state variables monitorable variables and model parameters are variables of the elementary Modula 2 data type REAL For every state variable of a ModelWorks model we also have to define an associated second variable of type REAL It either corresponds to the derivative x t dx dt continuous time or the 1 On the IBM PC you have to use the JPI TopSpeed Modula 2 development environment For more in formation on how to operate it consult the Appendix section installation or your JPI TopSpeed Modula 2 documentation T 25 ModelWorks 2 0 Tutorial new value x k 1 discrete time of the state variable x t respectively x k For the sample model which is continuous time this is the variable grassDot The procedure Dynamic is the heart of a ModelWorks model definition progra
282. ptor identifier unit ARRAY OF CHAR defaultSF StashFiling defaultT Tabulation defaultG Graphing PROCEDURE GetM VAR m Model VAR curMethod IntegrationMethod PROCEDURE SetM VAR m Model curMethod IntegrationMethod PROCEDURE GetSV m Model VAR s REAL VAR curInit REAL PROCEDURE SetSV m Model VAR s REAL curInit REAL PROCEDURE GetP m Model VAR p REAL VAR curVal REAL PROCEDURE SetP m Model VAR p REAL curVal REAL PROCEDURE GetMV m Model VAR mv REAL VAR curScaleMin curScaleMax REAL VAR curSF StashFiling VAR curT Tabulation VAR curG Graphing PROCEDURE SetMV m Model VAR mv REAL curScaleMin curScaleMax REAL curSF StashFiling curT Tabulation curG Graphing Removing of models and model objects aa Sit ito lg ag tg akg gl i a Sa a a TA pl ih a PROCEDURE RemoveM VAR m Model PROCEDURE RemoveSV m Model VAR s REAL PROCEDURE RemoveMV m Model VAR mv REAL PROCEDURE RemoveP m Model VAR p REAL PROCEDURE RemoveAllModels Global simulation parameters and project description e PROCEDURE CurrentStep INTEGER PROCEDURE CurrentTime REAL PROCEDURE SetDefltGlobSimPars t0 tend h er c hm REAL PROCEDURE GetDefltGlobSimPars VAR t0 tend h er c hm REAL PROCEDURE SetDefltProjDescrs title remark footer ARRAY OF CHAR wtitle wremark autofooter A 179 ModelWorks V2 0 Appendix recM recSV recP PROCEDURE Ge
283. r offers a set of pull down pop up or tear off menu commands second some menu commands their menu text is followed by will open so called entry forms offering from one to several editable fields to enter numbers or change settings via check boxes etc Thirdly the so called O windows allow to select particular models or model objects for modifi cations or to issue further commands The use of menus menu commands and entry forms are intuitively appealing easy to understand and follow the user interface of the Dialog Machine described elsewhere see Appendix and Literature There hold certain relationships among the tasks which are performed by ModelWorks during a simulation session Fig T16 Tasks are nested and each belongs to a particular hierarchical le vel T 49 ModelWorks 2 0 Theory T 50 Execution of procedure md 1 Simulation session Reset All Reset all Initialize session Initialize simulation session Structured simulation Experiment Initialize experiment Initialize run i a Output Input Dynamic Terminate experiment Fig T16 Relationships among the nested ModelWorks tasks performed during a simulation session The simulationist may execute an arbitrary number n of struc tured or elementary simulation runs A structured simulation experiment consists of an arbitrary number k of elementary runs Every elementary simulation run consists of the sections Initialize dyna
284. rameters PSAN aaa T SAAN NEE ANAN Project description Table Stash file name Clear table Windows Graph All model s integration methods Clear graph ModelWorks 2 0 Reference Start run R Halt run Pause H fot Stop Kill run K 5 Execute Experiment E All model s initial values All model s parameters All model s stash filing All model s tabulation All model s graphing All model s scaling All model s curve attributes Initialize Session All above Fig R1 All ModelWorks standard menus Fig R1 shows an overview of all standard menus and menu commands of the ModelWorks simulation environment A detailed explanation is given below If the modeler imports from module TabFunc an additional menu will appear Fig R11 In the PC version any of the menu commands starting with the phrase All model s are missing but note that these functions are also available in the IO windows 6 1 2 MENU FILE Page setup Print graph Preferences Fig R2 Menu File Page setup Usual page set up dialog box used for the printing currently chosen printer Not available in Reflex and PC version of the graph on the Print graph Prints the graph on the currently chosen printer Not available in Reflex and PC version Preferences Allows to set the modes of the simulation environment R 73 ModelWorks 2 0 Reference Preferences for simulation environment modes Fi
285. rential or difference equations or to treat them as an input A graphical representation is given in Fig T7 A more detailed representation is given in Fig T8 Input State Output State Output Fig T7 Graphical representation of a dynamic system a continuous time b discrete time These systems constitute the basic types used in ModelWorks to describe models A continuous time system is given by the following equations Dynamic equations dx t f x t u t p t t te tg teng tE R 4 1 Output equations y t g x t Po t t tE totend tE R 4 2 Initial conditions X t Xo 4 3 Input function u t tE totend tE N 4 4 Parameter set p t tE totend te N 4 5 1 E g LUENBERGER D G 1979 Introduction to dynamic systems Theory models and applications Wiley New York 446pp T 36 ModelWorks 2 0 Theory A discrete time system is described by the following equations Dynamic equations x k 1 f x k u k pr k k Keka ss KEL KE S 5 1 Output equations y k g x k Do k k Keka kp kE RJ 5 2 Initial conditions SKA Xo 5 3 Input sequence u k u ko u ky ulkp Keka kp kE x 5 4 Parameter set p k p k p ky DLK k ko kf ke I 5 5 where t Continuous time k Discrete time R Set of real numbers S Set of integer numbers x State vector dx Derivative vector for continuous time systems only u Input vector y Output vector p Para
286. rking with MacMETH T31 ModelWorks 2 0 Tutorial position the error has been detected E g if your model definition program is missing the declaration of the parameter c3 then your file may look similar to this PROCEDURE Dynamic BEGIN grassDot cl grass c2 grass grass c3 grass aphids identifier not declared or not t visible incompatible operand types T aphidsDot c3 c4 grass aphids c5 aphids Insert your corrections and recompile your program Execute the Merge command also after a successful compilation in order to have the old error messages automatically removed Merge always removes old error messages before it inserts any new ones If needed repeat the edit compile merge cycle until the compiler finds no more errors 3 2 3 SIMULATION OF THE NEW MODEL Choose from within the MacMETH shell the menu command File Execute A dialog box is displayed where you can select and open the just compiled object file GrassAphids OBM Note the name of the OBM file is defined by the name of the module and not by the name of the file containing the source program To avoid confusion it is recommended to use always the same names for files and modules After a while you see the initial start up screen of the ModelWorks simulation environment with the new variables displayed in the I O windows Execute the following steps to explore the new model Runa simulation with the default settings Choose Simu
287. rograms the first is the development shell RMSMacMETH and the subpro lUse the Font DA Mover a system utility which you obtained from Apple as part of the system software together with your Macintosh computer 2Although small and plain vanilla MockWrite has proven to be a very reliable text editor and the author probably suffers already too much from happy but nevertheless non paying customers A 129 ModelWorks V2 0 Appendix grams are the compiler the linker the debugger and several other utilities The application RMSMacMETH is an ordinary Macintosh application and can be treated as such i e start it with a double click Most of the time you will only work with the MacMETH shell It fully supports compilation and execution of Modula 2 program modules as sub programs e g the compiler or your model all run as subprograms under the shell Basics very brief Edit source file e g model definition program with any editor MacWrite and save as text only MockWrite Edit or MEdit etc To compile choose the menu command File Compile C and type the name of the module followed by a and a return extension MOD will be added automatically gt compilation starts If you encounter errors repeat se quence File Merge M edit source file recompile till you obtain no more compiler errors For a last time choose File Merge M to remove remaining comments with error messages from last erroneous compilation Choose the menu command
288. rrent values Fig 22 part II Theory The PC version will not offer any of the menu commands starting with the phrase All model s are available in O windows Set Set Global simulation parameters 1 Displays an entry form Fig R5a c to set the global simulation parameters such as the integration step Note that all of these parameters are valid globally i e they determine time and integration parameters for all present sub models together Start time for simulation t k The next simulation run will start with this time If only discrete time models are present case B Fig R5b k is an integer otherwise t is a real case A Fig R5a Stop time for simulation tgyq ks The next simulation run will stop with this time If only discrete time models are present case B Fig R5b k is an integer otherwise tenq is a real case A Fig R5a The following one or two fields vary depending on the kind of models which are present In the case A only continuous time sub models Fig R5a in case B only discrete time Fig R5b and in case C continuous time as well as discrete time submodels i e a mixed time structured model Fig R5c are present Integration step h h If at least one continuous time sub model is present h is the fixed time step for the numerical integration of the differential equations Moreover if a variable step length integration method is in use by at least one of the continuous time sub
289. rt as possible about Procedure writing additional information about the sub model e g by using routines such as WriteString WriteLn etc from the Dialog Machine module DMWindowIO into the help window Once a model has been declared it is ready for the declaration and the attaching of model objects to it Typically model object declaration procedures are called immediately following the call to DeclM However the following procedure can be used to change this behavior so that model objects can be attached to models in any sequence PROCEDURE SelectM m Model VAR done BOOLEAN The latter is particularly important if models and their model objects are declared dynamically during a simulation session This feature is not supported by the standard simulation environment but such an extension can be easily programmed via the client interface by the modeler He She will then use procedure SelectM to attach model objects to the proper model 7 1 3 DECLARATION OF STATE VARIABLES State variables are declared as real variables in the model definition program They may be declared anywhere in the program and may be part of a structured data type E g the following variables x and z may be used as state variables respectively state vector VAR x xDot REAL z zDot ARRAY 1 n OF REAL In order to declare a state variable s to ModelWorks or install it in the simulation environment the procedure DeclSV must be called It may not be ca
290. rth using SANE in expressions of type REAL However the situation is different if you use reals of type LONGREAL then the MacMETH compiler uses by default SANE and for math func tions you should use module LongMathLib which also uses SANE Note also that if your machine has an arithmetic coprocessor in particular if you work with version II V2 0 II SANE should be avoided in order to gain maximum perfor mance This can be achieved by importing math functions from module SYSTEM in stead of using any MathLib and compiling the code with the V2 0 II compiler com pile20 Then you bypass SANE completely and can take full advantage of the speed of the coprocessor without much loss in precision regardless of the real types you are using always maximum coprocessor precision 1 e 80 bits arithmetics Although a bit less precise as when using SANE the results still conform to the precision re quirements of the IEEE standard 754 for binary floating point arithmetic A 132 ModelWorks V2 0 Appendix Configure your installation by deleting the already existing file MathLib OBM it is actually a copy of file a in the folder and renaming the copy of any of these files to MathLib OBM will result in the corresponding SANE usage by the exported math functions e Supervision of heap usage by Dialog Machine and ModelWorks File DMHeapWatch OBM in the folder DMLib on diskette ModelWorks 1 2 comes in two versions a DMHeapWatch OBM check B DMHea
291. s For more specific information on how to program with the Dialog Machine please refer to the Dialog Machine software description and documentation F Bug Report Form If you encounter an error in ModelWorks please use a copy of the following bug report form fill it in and mail it to the address listed at the bottom See next page Order the Dialog Machine from the following address Projekt Zentrum IDA re Dialog Machine Swiss Federal Institute of Technology ETHZ ETH Zentrum CH 8092 Ziirich Switzerland A 139 ModelWorks V2 0 Appendix Eidgen ssische Ecole polytechnique f d rale de Z rich Technische Hochschule Politecnico federale di Zurigo Z rich Swiss Federal Institute of Technology Zurich Fachgruppe System kologie Systems Ecology Group Bug Report Form Use a copy of this form to report ONE software problem or documentation error please use addi tional forms to report multiple problems Thanks Your Name Your Address City State Country Your Phone Date Name of Product Version Date of Product Error Category check where appropriate Software Error Documentation Error Software Enhancement use reverse side Other please specify Problem Description Please answer check the applicable questions using the space below or the back of this form Software Error Encountered Bomb ID Number_____ System Freeze Recoverable Error Modul
292. s ordinate toggle function but will be used as x values on the abscissa The values of the variable Grass derivative should still be displayed as y values check the Y in the column Monitoring and the legends for the curves and the abscissa in the graph window Run another simulation run and you should see a dome shaped curve of Grass derivative vs Grass Before you proceed please select the command Reset All model s graphing under menu Settings This may depend on the currently set preferences i e the immediate update of the graph takes place only if the option Once changed immediately redraw graph available under menu command File Preferences is currently checked otherwise the redrawing of the graph will be deferred till the begin of the next simulation run T 20 ModelWorks 2 0 Tutorial During the steps described in the previous two paragraphs you may have noticed that ModelWorks uses different colors and line patterns if you have activated several monitoring variables at once For instance the Y in the window Monitoring variables is drawn in the same color as the corresponding curve and curves are drawn using different patterns important on monochrome screens and laser printers in order to assist you in telling the curves apart These characteristics of a monitoring variable are called curve attributes and they consist of first the line style LineStyle the pattern with which a line connecting two points is dra
293. s already sufficient to unbundle your application from MacMETH and avoids any user interface problems In case you wish to add to the new application its own unique icon instead of the default appli cation icon provided by the Macintosh system perform also the following two steps First edit a new desktop icon using the resource editor resource of type ICN Second start the application IconSwitcher3 You simply have to use the option install and make sure that the option Update Desktop is active before quitting this tool That s it T Iconia 6 3 shareware 10 or Iconia 7 0 are available fromHenrik Floberg of PHP Innovation M taregr nden 6 S 222 47 Lund Sweden Phone 046 12 69 14 or in case of Iconia 6 3 from a user group 2 The development tool ResEdit may be obtained from your Apple dealer from a software developer or from a user group 3 This tool is often available from the same sources as you can obtain the resource editor ResEdit see above A 134 ModelWorks V2 0 Appendix Problems with the linker Linking stand alone applications may not be possible because the linker cannot find all modules he needs This may happen even if the same program could be run successfully because contrary to the linking loader the linker may have troubles to find prelinked modules This is often the case with the file SimMaster OBM which contains almost the whole Dialog Machine and all ModelWorks object files The search strategy of the
294. s may be part of a complex data structure or the model definition may be spread over any number of modules thus supporting modular modeling This extensibility is one of the strongest features of ModelWorks Even if one is not familiar with the programming language Modula 2 but knows Pascal it is feasible to use ModelWorks On the other hand ModelWorks is powerful and flexible enough to allow also the advanced modeler to develop sophisticated models Note that with ModelWorks the modeler has not only full access to all features of Modula 2 but also to those of the Dialog Machine The Dialog Machine is a generally applicable software layer between an application program such as Model Works and the system software respectively hardware In this situation the user interacts via the the latter mouse keyboard screen only indirectly with the application the Dialog Machine intercepts all user interaction and filters it according to a simple user interface The Dialog Machine substantially facilitates the writing of interactive programs Not only does it simplify the programming of sophisticated dialogs but also does it ensure automatically a consistent man machine interface Hence it allows the modeler to extend the standard predefined ModelWorks simulation environment easily efficiently and without forcing him or her first to become a computer scientist yet it supports an easy programming of windows menus bit mapped graphics plus mouse inpu
295. s step The arrows indicate which values have become valid At the begin of the simulation run only time and the state variables are available for t These values are used to calculate the output variables for t Next the input variables can be calculated since they depend on the previously calculated outputs Next the numerical integration respectively the new state variables for time tj 1 are computed and if this step has been completed all state variables are assigned update the new values for tj Not defined the first time the integration loop is entered Value for t is calculated but not yet assigned to state variable field Only calculated if an integration method of higher order used fE 0 1 e g f 0 5 for Runge Kutta 4th order ModelWorks 2 0 Theory Once all dynamic client procedures i e Output Input and Dynamic have at least been called once all model variables i e input output state plus auxiliary variables are defined and have a correct value valid at the point t This is the moment ModelWorks does the monitoring i e the current value for any monitorable value is written onto the stash file tabulated in the table or drawn into the graph if the corresponding kind of monitoring is activated for the particular variable The monitoring is followed by additional calls now only of the client procedure Dynamic in case of higher order integration methods used to solve continuous time models Discrete
296. s the standard monitoring The exact sequence observed is that the client monitoring comes last so that it can also be used to customize or extend the just having made standard monitoring e g by drawing tangents along a solution of a differential equation 5 1 4 IO WINDOWS INPUT OUTPUT WINDOWS IO windows are available during a simulation session and have two functions First they dis play all models and all model objects plus their current values and settings Output Second they allow to modify interactively the current values and settings of these objects Input For instance the value of an individual model parameter can be changed or reset or the kind of monitoring for a particular variable during simulations can be defined There are four IO win dows The first IO window with the title Models lists all models known to ModelWorks i e which have been declared by the modeler via the client interface The second IO window State variables lists all declared state variables the third Parameters all parameters and the fourth Monitorable variables all monitorable variables al pyd 6 Model Parameters m Parameter names il Ident Unit Value il Potato model 6 0 hy lt 5 kGrowth kGrowth d dw plant 20 000 fel a 1 000 T kStem 5 000 kRoot E kRoot 1 000 kTuber k Tuber 6 000 I gt maxFlantF Age maxPlantP Phys age unit 675 000 maxLeafF Age maxLeatF Phys age unit 425 000 Der Fig T20 Basic structure of IO wi
297. scaling O Maximum scaling EJ Monitoring Fig R23 Entry form opened by the E button in the IO window Monitorable variables Selects all monitorable variables All subsequent button functions will operate on the scope All Fig T21 part II Theory i e on all monitorable variables of all models Set delete stash filing F writeOnFile notOnFile Adds the selected monitorable variable s to the list of variables which are to be written onto the stash file The function toggles actually the setting i e if the current setting is on F it is disabled otherwise enabled In the monitoring column of the IO window the monitorable variables to be written onto the stash file are marked with an F Fig R22 If a multiple selection is active the function adds or removes all currently selected variables to or from the list reversing the current setting of the first variable in the selection E Reset stash filing Resets the stash filing of the currently selected monitorable variable s to their defaults Set delete tabulation T writelnTable notInTable Adds the selected monitorable variable s to the list of variables which are to be tabulated in the table window The function toggles actually the setting i e if the current setting is on T it is disabled otherwise enabled In the monitoring column of the IO window the monitorable variables to be tabulated are marked with an T Fig R22 If a multiple selection is activ
298. section below How to work with the Dialog Machine D 2 HOW TO DEVELOP MODELS Modeling with ModelWorks PC version V1 1 PC is done by developing Modula 2 model definition programs using the JPI TopSpeed Modula 2 The sequence of steps to follow are shown in Fig 23 in part II Theory The model development cycle Once all installations as de scribed above have been made the model development cycle is exactly the same as that for any other Modula 2 program Therefore please consult the instructions as given in the JPI TopSpeed Modula 2 documentation Caution Do not invoke the TopSpeed linker from the DOS command line with M2 L this will cause linkage errors However it works correctly when you start the linker from within the editor with ALT L or ALT M A 138 ModelWorks V2 0 Appendix E How to Work With the Dialog Machine ModelWorks has been built as a Dialog Machine program s a ModelWorks module structure Fig 25 part II Theory This allows to use the Dialog Machine directly while programming models in two ways Either the model developer uses any of the Dialog Machine objects e g an entry form to edit values of some variables or an additional window to display simulation results in a fashion not provided by ModelWorks or to install an additional menu without activating the Dialog Machine itself E g the installation of an extra menu with its own customized menu commands can be easily made in the procedure passed as the
299. set ARRAY OF CHAR PROCEDURE StringField u Window VAR ei EditItem x y INTEGER fw CARDINAL string ARRAY OF CHAR PROCEDURE TextField u Window VAR ei EditItem x y INTEGER fw lines CARDINAL string ARRAY OF CHAR PROCEDURE CardField u Window VAR ei EditItem x y INTEGER fw CARDINAL card CARDINAL minCard maxCard CARDINAL PROCEDURE LongCardField u Window VAR ei EditItem x y INTEGER fw CARDINAL card LONGCARD minCard maxCard LONGCARD PROCEDURE IntField u Window VAR ei EditItem x y INTEGER fw CARDINAL int INTEGER minInt maxInt INTEGER PROCEDURE LongIntField u Window VAR ei EditItem x y INTEGER fw CARDINAL int LONGINT minInt maxInt LONGINT PROCEDURE RealField u Window VAR ei EditItem x y INTEGER fw CARDINAL real REAL minReal maxReal REAL PROCEDURE LongRealField u Window VAR ei EditItem x y INTEGER fw CARDINAL real LONGREAL minReal maxReal LONGREAL PROCEDURE PushButton u Window VAR ei EditItem x y INTEGER buttonWidth CARDINAL buttonText ARRAY OF CHAR pushButtonAction PROC PROCEDURE UseAsDefaultButton pushButton EditItem PROCEDURE BeginRadioButtonSet u Window VAR ei EditItem PROCEDURE RadioButton VAR radButt RadioBut x y INTEGER text ARRAY OF CHAR PROCEDURE EndRadioButtonSet checkedRadioButton RadioBut PROCEDURE CheckBox u Window VAR ei EditItem x y INTEGER text ARRAY OF CHAR boxChecked BOOLEAN PROCEDURE ScrollBar u Window
300. sh filing setting is activated F writeOnFile ModelWorks will never open a stash file regardless of the current settings of the recording flags see above Reset The reset menu commands Fig R5 assign to the selected element s the default value s s a section on Resetting Fig T22 part I Theory The latter have been defined by the modeler in the model definition program or have been predefined by ModelWorks All commands in this menu operate on the scope of all models respectively all objects of all models Fig T21 part II Theory other scopes are available in the IO windows only Reset Global simulation parameters Resets all global simulation parameters Reset Project description Resets all strings and flags to their defaults Reset Stash file name Resets the stash file name i e defines that the stash file used during the next simulation run will be the default stash file Reset All model s integration methods Resets the integration methods of all models Not available as a menu command in Reflex and PC version Reset All model s initial values Resets the initial values of all state variables of all models Not available as a menu command in Reflex and PC version R81 ModelWorks 2 0 Reference Reset All model s parameters Resets all parameters of all models Not available as a menu command in Reflex and PC version Reset All model s stash filing Resets the stash file setting writeOnFile notOnFile of al
301. shes to keep experimental results clearly separated from a theoretical mathematical model by formulating them as a parallel model or to enhance model clarity or to build model libraries Discrete and continuous models can be combined in one global model with correct data exchange controlled by the simulation environment Simple mathematical models can be built with only minor programming knowledge whereas programming experts have full access to a powerful programming language and may expand into any realm of sophisticated calculations still profiting from the si mulation environment and numerical algorithms provided by ModelWorks Hence in contrast to most existing simulation software ModelWorks fully supports the researcher during a model development process which often starts with a first crude model and ends with the most sophisticated in every detail refined research model ModelWorks is based on a high level programming language which has been selected considering the following criteria It has been formally defined it is general and powerful enough to support not only numerical computations but also a window based graphical user interface on the other hand it is also simple enough to be comprehended and mastered by the non computer scientist having learned programming in a basic computer science course such as for instance taught in Pascal programming courses on the other hand it also offers support for the development of large and
302. simulation Experiment Fig T16 An elementary simulation run is provided by ModelWorks and has not to be pro grammed by the modeler It is exactly the same as calling the procedure SimMaster SimRun via the client interface ModelWorks supports this level by requiring the modeler to specify for each model an nitialize and Terminate procedure The organization of an elementary simulation run is shown in more details in Fig T17 ModelWorks will execute for every model in the sequence of their declaration the nitialize pro cedures once at the begin of the simulation run and the Terminate procedures once at the end of the simulation run Note that the execution of the Initialize procedures happens at a moment when ModelWorks has already assigned the initial values to all state variables This is impor tant to know since this design makes it possible to overwrite the values assigned by ModelWorks with other values for instance during a structured simulation which is used to draw a phase portrait Typical use of Initialize and Terminate procedures is also the opening and closing of a file at the begin respectively the end of a simulation run in order to write simu lation results onto a file different from the stash file Note also that in contrast to the procedure md passed to ModelWorks as actual argument in the call to procedure RunSimMaster it is called only once per simulation session and typically de clares the models and model objects th
303. splay A model is always of a particular type i e either continuous time or discrete time This type is given by the kind of equations which belong to the model In the case of continuous time the model equations are ordinary differential equations in the case of discrete time they are ordinary difference equations Note however that aModelWorks model definition program may be structured i e it consist of several models which may be of a differing type i e some models may be continuous time other discrete time In the latter case results a so called mixed continuous and discrete time model definition T11 ModelWorks 2 0 Tutorial A model may consist of any number of model equations However they must be given as explicit either first order differential equations or first order difference equations E g the following differential equation describing the van der Pol oscillator y u y Dyt y 0 is not in the proper form since it is neither explicit nor is it first order On the other hand the same equation reformulated as a system of explicit coupled first order differential equations x X2 x u 1 XxT x2 X1 is now suitable to be used directly as a set of ModelWorks model equations The second form is called the state variable form Most differential or difference equations can be formulated in this form Usually each model uses a number of state variables Each state variable must be associated with a second variable
304. stallAliasChar Separator InstallSeparator DisableCommand EnableCommand ChangeCommandText InstallQuitCommand FROM RandGen IMPORT U ResetSeeds Randomize FROM DMFiles IMPORT TextFile CreateNewFile Close WriteEOL PutReal Response WriteChar WriteChars PutInteger FROM DMEntryForms IMPORT FormFrame WriteLabel DefltUse IStatView 512 is an interactive statistics amp graphics package from Abacus Concepts Inc published by Brainpower Inc 24009 Ventura Blvd Suite 250 Calabasas CA 91302 A 158 ModelWorks V2 0 Appendix CharField StringField CardField IntField RealField PushButton RadioButtonID DefineRadioButtonSet RadioButton CheckBox UseEntryForm CONST firstIndiv 1 lastIndiv 100 TYPE Individuals firstIndiv lastIndiv TYPE State one two three CONST firstState MIN State lastState MAX State VAR m Model x xDash ARRAY firstIndiv lastIndiv OF State pseudo state vars i e not declared to ModelWorks p ARRAY firstState lastState firstState lastState OF REAL Markov matrix pacc ARRAY firstState lastState firstState lastState OF REAL Matrix containing accumulated transition probabilities initP ARRAY firstState lastState OF REAL Probabilities used to compute initial states c ARRAY firstState lastState firstState lastState OF INTEGER counting of transitions n ARRAY firstState lastState OF INTEGER number of
305. t Moreover the resulting program will be user friendly Thanks to the Dialog Machine s dialog capabilities the simulationist will be able to enjoy the use of a simulation program which automatically conforms to a robust man machine interface This offers the advanced modeler to concentrate on the modeling process instead of being distracted by the cumbersome and complex implementation details of user interface problems The easy access to the Dialog Machine is another strength of ModelWorks For instance the modeler may wish to extend the simulation environment by programming his own graphical monitoring in an additional separate window or by adding further customized functions to the simulation environment i e by installing more menus offering additional menu commands To give an example ModelWorks and the Dialog Machine have been successfully used to program an interactive modeling environment which allows to enter differential equations and model objects at run time without having to resort to any programming at all Despite the many features ModelWorks offers typical model definition programs are written in a simple standard format Hence as long as one develops models without any sophisticated extras even the beginning programmer can quickly learn to use ModelWorks successfully Finally as a simulationist only there is no need to know anything about the more advanced features of ModelWorks since ModelWorks itself has been implem
306. t dots hidden wipeout Range RECORD min max REAL END AxisType RECORD range Range scale ScalingType dec CARDINAL tickD REAL label LabelString END GraphProc PROCEDURE Graph VAR DM2DGraphsDone BOOLEAN PROCEDURE DefineGraph VAR g Graph u Window r RectArea xAxis yAxis AxisType grid GridFlag PROCEDURE SetNegLogMin nlm REAL PROCEDURE DefineCurve g Graph VAR c Curve col Color style PlottingStyle sym CHAR PROCEDURE RedefineGraph g Graph r RectArea xAxis yAxis AxisType grid GridFlag PROCEDURE RedefineCurve c Curve col Color style PlottingStyle sym CHAR PROCEDURE ClearGraph g Graph PROCEDURE DrawGraph g Graph PROCEDURE DoForAllGraphs u Window gp GraphProc PROCEDURE DrawLegend c Curve x y INTEGER comment ARRAY OF CHAR PROCEDURE RemoveGraph VAR g Graph PROCEDURE RemoveAllGraphs u Window PROCEDURE RemoveCurve VAR c Curve PROCEDURE Plot curve Curve newX newY REAL PROCEDURE Move c Curve x y REAL PROCEDURE PlotSym g Graph x y REAL sym CHAR PROCEDURE ConvertToPoint g Graph xReal yReal REAL VAR xInt yInt INTEGER PROCEDURE PlotCurve c Curve nrOfPoints CARDINAL x y ARRAY OF REAL PROCEDURE WindowToGraphPoint g Graph xInt yInt INTEGER VAR xReal yReal REAL PROCEDURE GraphExists g Graph BOOLEAN PROCEDURE CurveExists g Graph c Curve BOOLEAN UE II IE SEI RE IE NE IE RR IRN EEE DMAlerts SHR SN AAR E E AR RS
307. t initial value at the beginning of each simulation run to initialize s The default initial value is re assigned to the current initial value by ModelWorks during the initialization phase of the simulation session or after every reset of the state variables During a simulation session the current initial value may be changed by the simulationist using an IO window or by the modeler via the procedure SetSV The modeler could also overwrite the value of s with another value within procedure Initial see procedure DeclM since ModelWorks has already assigned the current initial value to the state variable s Note however that in the latter case inconsistencies might occur between the display of the current value in the IO window with the current values actually used in the simulations Avoid this method and use the procedure SetSV instead minCurlnit maxCurlnit Lower and upper bounds for the current initial value Attempts by the simulationist to assign values out of this range are not accepted descriptor String containing a long description of the state variable s This string may have any length but might not be visible till its end when it is too long to fit into the IO window column where it is displayed during a simulation session see also identifier Example Density of alga Scenedesmus obliquus identifier Short string identifying the state variable s This string should be kept as small as possible in order to ensure full visibility
308. t any further computations by the ModelWorks run time system The procedures ExperimentRunning and ExperimentAborted from module SimMaster allows to determine if an experiment has been started respectively aborted by the simulationist This allows the modeler to program structured simulations with maximum flexibility 5 2 7 PROGRAMMING STRUCTURED SIMULATIONS EXPERIMENTS Many functions of ModelWorks are also available via the client interface For instance it is possible to start an elementary simulation run by calling procedure SimRun from module SimMaster A structured simulation or experiment consists typically of a sequence of calls to procedure SimRun The following example illustrates a situation in which four initial state vectors x y 1 1 2 2 1 1 and 2 2 for a second order system of differential equations are to be used to produce a phase portrait Each combination will be used in a simu lation run PROCEDURE MyExperiment BEGIN SetSV m x 1 0 SetSV m y 1 0 SimRun SetSV m x 2 0 SetSV m y 2 0 SimRun SetSV m x 1 0 SetSV m y 1 0 SimRun SetSV m x 2 0 SetSV m y 2 0 SimRun END MyExperiment T 68 ModelWorks 2 0 Theory With the exception of the declaration procedures DeclM DeclSV DeclP and DecIMV it is ba sically possible to use any procedure from the client interface to program a structured simulation experiment However since some procedures affect values currently in use such as
309. t least of the modules SimBase and SimMaster mandatory part The model definition program may ac tually consist of just one program module up to any number of modules The in ternal ModelWorks modules are linked together in SimMaster OBM mandatory imports optional imports The optional client interface TabFunc is useful if the modeler uses nonlinear functions which are defined by a table of supporting points so called table functions During simulations ModelWorks will then linearly interpolate or extrapolate needed values SimIntegrate can be used to integrate a model anytime without actually running a simulation The global simulation time of the simulation environment will remain unaffected by such an integration SimGraph Utils can be used to draw into the standard graph window This feature can be used to draw measurements with error bars into the graph or to customize the graph in any desired way by using routines from the Dialog Machine module DMWindowIO In the current ModelWorks version the auxiliary library consists among others of the modules ReadData JulianDays DateAndTime WriteDatTim RandGen and RandNormal ReadData facilitates the reading of data from text files for instance when the simulationist wishes to com pare simulation results with measurements JulianDays provides functions useful for the map ping of the simulation time to calendar dates and vice versa DateAndTime and WriteDatTim al low to access the b
310. t page Second only the space below these now top rows will be used to add new rows Hence this number specifies how many rows are common to two consecutive pages T57 ModelWorks 2 0 Theory simulation However this behavior may differ depending on the currently set mode of the simulation environment preferences An example graph is shown in Fig T19 The graph s size is automatically fit to the window s size The actual graph is drawn as large as possible which depends on the number of curves to be listed in the legend at the bottom of the window However if there are too many curves requested so that the legend would become too big and there would not be left a minimal space for the panel of the graph ModelWorks will not be able to list all curves in the legend Only the first ones will be visible the remaining ones at the bottom of the list will be missing If an other than the standard monitoring of ModelWorks is required the modeler can program and install it by calling the procedure DeclClientMonitoring Any kind of monitoring will then be possible e g the writing of simulation results onto a file or the drawing of animated graphical objects which move within a window according to computed positions etc Typically in order to accomplish such tasks the modeler uses the Dialog Machine to which he she has full access Concerning values of state and other variables the client monitoring will be done as often and at the same time a
311. tDefltProjDescrs VAR title remark footer recM recSV recP PROCEDURE SetDefltIndepVarIdent descr ident unit ARRAY OF CHAR PROCEDURE SetMonInterval hm REAL only for upward compatibility PROCEDURE SetIntegrationStep h REAL only for upward compatibility PROCEDURE SetSimTime t0 tend REAL only for upward compatibility PROCEDURE SetGlobSimPars t0 tend h er c hm REAL PROCEDURE GetGlobSimPars VAR t0 tend h er c hm REAL PROCEDURE SetProjDescrs title remark footer ARRAY OF CHAR recMV recG BOOLEAN ARRAY OF CHAR VAR wtitle wremark autofooter recMV recG BOOLEAN wtitle wremark autofooter recM recSV recP recMV PROCEDURE GetProjDescrs VAR title remark footer ARRAY VAR wtitle wremark autofooter recM recSV recP recMV recG BOOLEAN OF CHAR recG BOOLEAN PROCEDURE SetIndepVarIdent descr ident unit ARRAY OF CHAR Control of simulation run conditions Je a 5 i i Sr m u Md nt st ct E TYPE TerminateConditionProcedure PROCEDURE BOOLEAN StartConsistencyProcedure PROCEDURE BOOLEAN PROCEDURE InstallStartConsistency sc StartConsistencyProcedure PROCEDURE InstallTerminateCondition tc TerminateConditionProcedure Control of Display and Monitoring PROCEDURE TileWindows PROCEDURE StackWindows TYPE MWWindow MIOW SVIOW PIOW MVIOW TableW GraphW AboutMW PROCEDURE SetWindowPlace mww MWWindow Y w h INTEGER PROCEDUR
312. tal Sciences ETH Zentrum CH 8092 Zurich Switzerland Last revision 31 Jan 89 A F Fe ke de Fe Fe e He He e e Fe Fe He Fe Fe He RE ERE ER ERE ERE REE EER He He He ERE He Fe He He He He e He He He He He He e He He He e He Me Me He e He e PROCEDURE SetSeeds z0 zl z2 INTEGER defaults z0 1 z1 10000 z2 3000 PROCEDURE GetSeeds VAR z0 zl z2 INTEGER PROCEDURE Randomize set seeds using seed values depending on a particular unique and non repeatable event in real time e g date and time of the clock Implies a call to SetSeeds PROCEDURE ResetSeeds reset seeds to values defined by last call to SetSeeds PROCEDURE U REAL returns within 0 1 uniformly distributed variates Based on a combination of three multiplicative linear congruential random number generators of the form z k 1 A z k MOD M with a prime modulus and a primitive root multiplier gt individual generator full length period The multipliers A are 171 172 and 170 the modulus M are A 150 ModelWorks V2 0 Appendix 30269 30307 and 30323 b END RandGen G 2 6 RandNormal DEFINITION MODULE RandNormal BER e Fe He Fe He Fe Fe 22 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 5 2 2 2 2 2 2 2 2 2 5 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 Module RandNormal Version 1 0 Copyright 1987 by Andreas Fischlin and CELTIA Swiss Federal Institute of Technology Z rich ETHZ Version for MacMETH V2 6 1 Pass
313. ter value p by ModelWorks during the initialization phase of the simulation session or after every reset of the model parameters During a simulation session the current parameter value p may be changed by the simulationist using an IO window or by the modeler via overwriting the value of p with another value e g within procedure initialize see procedure DeclM by calling procedure SetP R 103 ModelWorks 2 0 Reference minVal maxVal Lower and upper value bounds for p Attempts by the simulationist to assign values out of this range are not accepted runTimeChange rtc run time change interactive changing of values of model parameter D during a simulation run in the program state Pause is enabled noRtc no run time change disallows completely any changing of values of the model parameter p during a simulation run even in the program state Pause descriptor String containing a long description of the model parameter p This string may have any length but might not be visible till its end when it is too long to fit into the IO window column where it is displayed during a simulation session see also identifier Example Half saturation constant for algal growth identifier Short string identifying the model parameter p This string should be kept as small as possible in order to ensure full visibility for the display in small IO windows during a simulation session In particular on small screens IO windows become small in the til
314. tes No simulation the state in which the simulationist may change values or settings e g simulation time initial values of state variables or values of model parameters During a simulation run or a structured simulation ex periment ModelWorks is in the state Simulating In this state the simulationist may only temporarily pause or stop kill the running simulation The state Pause al lows to change model parameters with the attribute RTC Run Time Change set or to resume respectively abort the simulation For every state the status of the menu commands is symbolized as follows A black line signifies an active a grey an inactive or unavailable menu command The availability of the button commands of a particular O window is indicated by a black object selection possible all buttons active or grey disabled selection inactive buttons window title bar T 48 ModelWorks 2 0 Theory For instance changing a parameter value while a simulation is running is prohibited i e may not be issued in the state Simulating in order to avoid the display of incorrect simulation results or the command to stop a simulation is meaningless if there is no simulation run in progress i e this command is available only in the states Simulating or Pause In particular this design also ensures that the simulationist can not start a second simulation run on top of an already running one unless the first one has been terminated The current states can be det
315. tes the typical use of the ran dom number generators in the auxiliary library and serves as an example for stochastic simula tions The program models a finite 3rd order Markov chain process Note that the program also directly accesses frequently the Dialog Machine for instance it installs the custom menu Markov which allows the simulationist to alter the meaning of the 3 states and to set the coef ficients of the Markov matrix in an especially suitable form This program also demonstrates the use of the start consistency testing mechanism of ModelWorks see function procedure TestConsistency The coefficients of any probability vector or of a row of a Markov matrix must add up to the sum 1 otherwise the simulationist has defined an illegal probability vector or Markov matrix Thus any simulation using a Markov matrix or an initial probability vector not satisfying these conditions would produce meaningless results The here used start consistency testing mechanism of ModelWorks automatically prevents this This program also demonstrates the use of state events Since the default process contains an absorbing state further computations beyond the state all dead i e probability vector healthy ill dead becomes 0 0 1 are superfluous The terminate condition testing mechanism of ModelWorks see function procedure AllDead allows to stop a running simulation anytime this state event is encountered As it holds in general for stochastic
316. th initial value during declaration write overwrite with initial value at begin of run write integration continuous time only read and write update with new value obtained via integration write x x k 1 Derivative continuous time new value discrete time integration read Xo Initial value 9 overwrite with default during declaration write overwrite with default while resetting initial values write editing of value via IO window read and write X State variable Model parameters Model parameter overwrite with default during declaration write overwrite with default while resetting parameters write editing of value via IO window read and write Monitorable variables Monitorable variable overwrite with 0 0 during declaration write monitoring read Stash filing Tabulation or Graphing attribute overwrite with default during declaration write overwrite with default while resetting parameters write editing of value via IO window read and write Tab T2 Actions of ModelWorks performed on installed model objects 10 variable belongs to ModelWorks not to the client s model definition program 11 see previous footnote T 66 ModelWorks 2 0 Theory Models are always calculated in parallel regardless of the presence of any coupling among them i e the calculation order of the client procedures is first all Ouzput of all submodels sec ond all nput of all submodels etc The actual sequence of the computations of a par
317. the User Profile Experience shows that many user problems with MacMETH are only due to inconsistencies between the names in the User Profile and actual disk or folder names The ModelWorks distribution disk already contains a User Profile which allows you to work directly with the disks given you do not change any folder names Path names are a list of names separated by colons and always end with a colon There are relative and absolute path names An absolute path starts with the name of the disk as its first element and lists all folders as they reside within each other E g the absolute path name ModelWorks 1 2 RAMSESLib listed in the User Profile will result in file searches e g by the compiler within the folder named RAMSESLib on the disk named ModelWorks 1 2 Any path name starting with the special character is interpreted as a relative path name All searches start relative to the location i e the folder where the MacMETH shell RMSMacMETH resides E g the relative path name Work ProjectLib will cause the compiler or linking loader to search for files in the folder ProjectLib contained within the folder Work Note that this example would function properly only if the MacMETH shell is in the same folder or on the same disk as the folder Work From this follows that files residing on other disks than the MacMETH shell can only be accessed by absolute paths Absolute paths are more error prone than relative paths try to avoid them
318. the point G t i e solves an initial An exact description on how to install ModelWorks is given in the appendix Please follow these instructions exactly otherwise you may have difficulties while executing the described steps T 16 ModelWorks 2 0 Tutorial value problem The sample model with the differential equation 1 has already been programmed compiled and is ready for execution 2 2 Simulating the Sample Model To run the sample model you have first to start the MacMETH Modula 2 development shell Start it with a double click on its icon as you start any other Macintosh application By the way although there are other methods possible it is generally recommended to develop and simulate ModelWorks models only by working with the MacMETH shell in the here described way To simulate the logistic grass model choose the menu command Execute under the menu File A dialog box is displayed where you can select and open the object file of the model Logistic OBM contained in the folder Work or the folder Sample Models on one of the ModelWorks diskettes Upon opening this object file you will start the ModelWorks simulation environment linked together with the sample model Technically speaking this simulation program is an ordinary MacMETH Modula 2 program running as a subprogram under the MacMETH shell Once fully started you see the initial screen of the ModelWorks simulation environment with its menu bar and the four
319. ticular kind of client procedures e g the sequence with which ModelWorks calls the procedures Dynamic is given by the sequence of their declarations in the model definition program In the current version of ModelWorks input output and auxiliary variables do not appear expli Program states and transition commands of client interface No simulation Last run executed or Stop kill run Execute Experiment or Start run Simulating Stop kill run Stop time SimR PauseRun reached or Er iZ Termination Haltrun condition true or Start Pause Stop kill run Resume run Fig T26 State transition diagram of the client interface of ModelWorks In contrast to the program states viewed by the simulationist Fig T15 a simplified projection of this diagram the modeler has a different view While programming structured simulations Experiment the modeler needs to consider two sub states in the state Simulating No run and Running The state No run allows the mode ler to modify current values e g initial values parameters or the integration step In the state Running no changes may be made to global simulation parameters Note that some transitions are only available to the simulationist bold text of menu commands e g Start run or Execute experiment some are also under the control of the modeler plain procedure identifiers e g PauseRun Note also that some commands which are available to the simulationist e g Start run
320. ting is quite normal and frequent and causes usually no harm Hence the display of an alert would be too cumbersome but the quiet overwriting can become dangerous if the modeler programs the stash file name erroneously see procedure SwitchStashFile from module SimBase If the monitoring settings for tabulation or graphing are activated the corresponding windows are automatically opened or brought to the front If already any of the windows Table or Graph are the frontmost window ModelWorks will not automatically open the other window or bring it to the front This allows the simulationist to suppress the automatic opening or bringing to the front of either the table or graph window by bringing first the other one to the front before he starts or resumes a simulation If there are no monitorable variables which have an active tabulation setting T writeInTable the table window may not remain open and will be automatically closed in case it should be already open at the begin of the simulation If there are no monitorable variables which have an active graphing setting Y isY the graph window may not remain open and will be automatically closed in case it should be already open at the begin of the simulation In all other cases ModelWorks leaves the windows where they are 6 1 7 OPTIONAL MENU TABLE FUNCTIONS In addition to model parameters ModelWorks offers the optional possibility to use non linear functions as parts of equations which are not
321. tings For further explanations see mode redraw graph always above PROCEDURE SetCommonPageUpRows rows CARDINAL PROCEDURE GetCommonPageUpRows VAR rows CARDINAL This mode controls the number of common rows between page ups in the table window A page up occurrs when the table window is full but more rows should be written then ModelWorks attempts to erase most of the table and restarts tabulating from the top again The number rows specifies how many rows at the bottom are not erased but scrolled to the top of the next page The rest of the table window is then used to add the rows of the new page Thus rows specifies how many rows are common to two consecutive pages PROCEDURE SetRedrawGraphAlwaysMode rga BOOLEAN PROCEDURE GetRedrawGraphAlwaysMode VAR rga BOOLEAN If the mode RedrawGraphAlways is activated each modification of the graphing settings will be displayed immediately not only at the begin of the next simulation run This implies an immediate loss of all simulation results eventually currently visible in the graph as soon the simulationist edits any graphing settings If this mode is not active the current graph will not be touched unless the user starts another simulation at its begin the whole graph will be redrawn R 120 ModelWorks 2 0 Reference PROCEDURE SetColorVectorGraphSaveMode crg BOOLEAN PROCEDURE GetColorVectorGraphSaveMode VAR crg BOOLEAN Above procedures allow to control the mode of gr
322. to operate the computer and its software and that you have ModelWorks installed and ready in order to actually perform the described steps on your computer while reading this chapter 3 1 The Model Definition Program of the Sample Model In the last chapter you worked with the grass growth model in the ModelWorks simulation environment as a simulationist Now we shall have a closer look at the used simulation program as a modeler This program defines declares a logistic growth model and is called a model definition program Step by step we shall now go through this sample program contained in the file Logistic MOD and have a closer look at all its elements The complete listings of the program Logistic and of the definition modules SimMaster and SimBase which form the client interface are given in the appendix Many ModelWorks model definition programs have the same structure as Logistic MOD The import list contains all the items types constants variables and procedures used within the program module They are exported by the modules which form the client interface of ModelWorks FROM SimBase IMPORT Model IntegrationMethod DeclM DeclSV DeclP RTCType StashFiling Tabulation Graphing DeclMV SetSimTime NoInitialize NoInput NoOutput NoTerminate NoAbout FROM SimMaster IMPORT RunSimMaster RunSimMaster is the procedure which will start the simulation environment of ModelWorks DeclM DeclSV DecIMV and DeclP are the
323. tosh Internal report Project Centre IDA Swiss Federal Institute of Technology Z rich ETHZ Switzerland FISCHLIN A amp ULRICH M 1987 Interaktive Simulation schlecht definierter Systeme auf modernen Arbeitsplatzrechnern die Modula 2 Simulationssoftware ModelWorks Proceedings Treffen des GVASIM Arbeitskreises 4 5 2 1 Simulation in Biologie und Medizin February 27 28 1987 Vieweg Braunschweig 1 9 FORRESTER J R 1970 Principles of systems Addison Wesley N Y IEEE STD 754 1985 1985 IEEE standard for binary floating point arithmetic New York IEEE Inc or IEEE TASK P754 1981 A proposed standard for binary floating point arithmetic Draft 8 IEEE Computer 14 3 51 62 KORN G A amp WAIT J V 1978 Digital continuous system simulation Prentice Hall Englewood Cliffs N J 212pp LOTKA A J 1925 Elements of physical biology Baltimore Williams and Wilkins LUENBERGER D G 1979 Introduction to dynamic systems Theory models and applications Wiley New York 446pp R 123 ModelWorks 2 0 Literature ROBINSON S B 1986 STELLA Modeling and simulation software for use with the Macintosh Byte 277 278 ULRICH M 1987 ModelWorks An interactive Modula 2 simulation environment Post graduate thesis Project Centre IDA Swiss Federal Institute of Technology Z rich ETHZ Switzerland 53pp VOLTERRA V 1926 Variazione e fluttuazini del numero d individui in specie animali conviventi M
324. tributes assignments are the same as for ordinary monitoring variables for color and symbol but not for the lineStyle the default line style is hidden which means that the connections from x v point to x v point are not drawn In that case and if withErrBars is set true then the error bars are displayd solidly All other line styles are applied to the connections from point to point as well as to the error bars themselves This procedure allows also redeclare such data series i e to associate other data to the same mvDepVar and mvIndepVar PROCEDURE DisplayDataNow mDepVar Model VAR mvDepVar REAL This procedure allows to display a series of e g validation data before a simulation run The previously declared data are displayed in the current graph window under the following conditions the data have been declared properly and are valid the associated monitoring variable is selected to be displayed isY the declared indepVar is the currently active independent monitoring variable isX the declared indepVar is either not a monitoring variable for example timeIsIndep what implies that time is meant and time is the selected independent var the data fall into the declared scaling range x PROCEDURE DisplayAllDataNow Displays all declared datasets at the specified moments The same conditions apply as for DisplayDataNow X PROCEDURE RemoveDispData mDepVar Model VAR mvDepVar
325. ually used However note that this may lead to unpredictable results e g if the integration method is changed in the middle of an integration i e by calling SetM from within procedure Dynamic the results of the integration will most likely be wrong However if the modeler calls SetM from within procedure Output or Input no problems should occur e Changes programmed by calling procedures which affect default values only procedures SetDeflt StashFileName etc will not become effective until a reset of the corresponding type of values is executed The feature programming experiments offers unlimited possibilities they can t be all explained here Instead it is left to the modeler s responsibility to understand what ModelWorks is exactly doing see this chapter and in particular section on ModelWorks objects and the run time system and to program the problem at hand accordingly Structured simulations are not only most useful but a necessity if a model is to be used for a parameter identification or a sensitivity analysis etc To illustrate this point the example of a lit tle sensitivity analysis is presented Given a set of n model parameters and for each parameter a triple of values lower boundary of a confidence interval mean and upper boundary of the confidence interval a 5 The values are stored on a text file in this format min pl mean pl max pl descriptor of pl pl unit pl min p2 mean p2 max p2 descriptor of p2 p2 unit p2
326. uch as equations state variables and parameters plus the objects default values and ranges It is also the modeler who implements the model by writing a ModelWorks model definition program The simulationist runs interactive simulation experiments hereby using one or several models which have been constructed by the modeler He is restricted to use these models within certain limitations which have been specified by the modeler but within that range he may interactively define and execute with the model any kind of experiment he wishes For instance he may observe its temporal behavior sample points from particular trajectories modify parameter values within a defined range or run a Sensitivity analysis ModelWorks contains all elements and algorithms needed for computer simulations such as numerical integration algorithms the interactive changing of parameter values and the display of simulation results The only exception of course is the model itself which has to be provided by the modeler T 10 ModelWorks 2 0 Tutorial Normally a ModelWorks model definition consists of several objects which belong to various classes First there must be present at least one model but the model definition may consist of any number of models Second normally each model is associated with several objects like model equations state variables model parameters auxiliary variables and monitorable variables Such objects are called model objects
327. ues There is no such thing as negative grass hence the lower limit has been set to 0 0 the upper to a value beyond which values are no longer plausible The three strings are the name an abbreviated name and the unit of the state variable These strings and the initial value will be displayed in the IO windows for state variables The limits for the initial value will be used during interactive changes attempts by the simulationist to enter initial values out of the allowed range will be refused With this mechanism the modeler can prevent the simulationist from entering values which would result in illegal simulation experiments for which the model is not defined or which could cause some other fatal run time errors The procedure DecIMV declares the variables grass and grassDot as monitorable variables This is necessary if we want to monitor the values of these variables on the stash file in the table or ina graph Typically state variables auxiliary variables and output variables used to couple submodels are the model objects which are declared as monitorable variables Our calls of DecIMV DeclMV grass 0 0 1000 0 Grass G g dry weight m 2 notOnFile writeInTable isY Dec1MV grassDot 0 0 500 0 Grass derivative dG dt g dry weight m 2 time notOnFile notInTable notInGraph For instance you could use ModelWorks also for the plotting of a function e g a time series measured during an experiment par
328. uilt in computer clock in order to record real time events such as begin and end of a long simulation RandGen and RandNormal return uniformely within 0 1 and nor mally N u 0 distributed variates to support stochastic simulations Since this auxiliary libra T 65 ModelWorks 2 0 Theory ry is completely independent of ModelWorks the modeler is free to add any modules he she wishes 5 2 5 MODELWORKS OBJECTS AND THE RUN TIME SYSTEM ModelWorks maintains the model objects e g during numerical integration although the vari ables representing these objects are contained only in the model definition program The advan tage is that the modeler may define and access these variables in whichever way he likes e g by using state variables as part of an array or a record data structure Note however that Model Works can do so only by using the following mechanism During declarations ModelWorks stores the addresses of the variables declared as objects Later during simulations ModelWorks will access the model objects and their associated variables Tutorial Fig T2 for reading or writing Tab T2 by assuming that these objects still exist Hence the modeler must be very careful to ensure that any model object exists as long as the program environment exists i e is always declared as a global Modula 2 variable and not as a variable local to a procedure Symbol Meaning of variable Action by ModelWorks State variables overwrite wi
329. uire a reset before becoming effective This is not the case for the first two procedures which take effect immediately Colors are not available in the PC version Note that if either autoDefCol or autoDefStyle or autoDefSym is used the automatic definition mechanism for colors line styles and for symbols as provided by the simulation environment becomes active see Tab T1 part II Theory Hence if you wish to really set a curve attribute make sure that all attributes are set different from an autoDef value In particular note that the procedures affecting current values function also in the middle of a simulation This behavior may be useful for instance to make a portion of a curve for a certain time invisible A typical application is the simultaneous display of a measured time series and the monitoring of solutions of a system of differential equations if some measurements are missing there arises the need to suppress partially the monitoring i e to display nothing for the R 119 ModelWorks 2 0 Reference measurements but to monitor the behavior of the model equations Using procedure SetCurveAttrForMV with the line style invisible will allow to achieve the desired effect Note that if curve attributes are changed dynamically there may appear inconsistencies between the curve attributes used for the curves themselves and those used to draw the legend This is because the legend shows only those curve attributes which are c
330. ulation session it always first executes a full reset Fig T16 i e it ensures that all current values to be used during the session have exactly the values as de fined by the so called defaults s a chapter Predefinitions defaults current values and reset ting Defaults can stem from different sources i e they are either predefined by ModelWorks or have been specified by the modeler via the client interface Next ModelWorks initializes the simulation session Fig T16 The modeler can customize this initialization e g in order to add an additional menu or to set particular defaults etc This can be accomplished by installing an initialization procedure with a call to the procedure DeclInitSimSession from SimMaster before calling procedure RunSimMaster ModelWorks will then automatically call the installed initialization procedure at the begin of a simulation ses sion or whenever the simulationist requests this via the corresponding menu command provided in the simulation environment After the reset and the initialization the program state is always NoSimulation Fig T15 5 1 2 b Structured simulation Experiment This level can be accessed by the simulationist by choosing the menu command Execute exper iment under menu Simulation It is the next level below that of the simulation session Fig T16 A structured simulation works similar to an elementary simulation run but differs slightly in the following aspects Basically it is
331. unction to be reset by clicking in the corresponding radio button Again table functions are selected via their short identifiers RESET one of the following Table Functions w BRCMT BRFMT BRMMT gt BRPMT gt CFIFRT CIMT COR O DRCMT C DREMT CO DRMMT C DRPMT O FEMT Co FPCIT FPMT NREMT gt NRMMT gt FOLATT POLCMT O OLET OLFT gt OLMT OLPT Fig R14 Entry form to select a table function for an individual of all values of a particular table function The functions are listed with their identifiers Reset all Resets all values of all table functions to their defaults without asking the simulationist for a selection of a particular table function Show window Shows the table function editor window by bringing it to the front Note that the resetting of table functions is only available in the state NoSimulation R 87 ModelWorks 2 0 Reference 6 2 IO Windows Input Output Windows IO windows serve the entering of new data Input and the display of the current values Output of models and model objects For a description of the general operation of IO windows see also the section on IO windows and in particular Fig T20 in part II Theory 6 2 1 IO WINDOW MODELS The IO window Models Fig R15 displays information name identifier current integration method about the installed models Furthermore it offers a mechanism to select models and to execute functions which operate on the sel
332. under the control of the modeler for instance to construct a whole phase portrait by means of a single menu command or to identify a model parameter Precondition is that the simulation environment has been called procedure RunSimMaster see chapter Starting a simulation session and that it is still active PROCEDURE CurrentSimNr INTEGER Returns the current simulation run number k during structured simulations k 1 2 3 Fig T16 A typical usage of this procedure looks similar to the following statement REPEAT SimRun UNTIL CurrentSimNr maxSimNr Note however that even aborted runs are numbered To handle properly abortion of structured simulation runs see below procedure ExperimentAborted TYPE StartConsistencyProcedure PROCEDURE BOOLEAN TerminateConditionProcedure PROCEDURE BOOLEAN PROCEDURE InstallStartConsistency sc StartConsistencyProcedure Procedure sc is called at the begin of a simulation run right after the execution of the procedure initialize see procedure DeclM and after resuming a run from the state Pause If it returns FALSE the simulation will be aborted and the simulation environment immediately returns into the program state No simulation Otherwise the simulation is normally continued Typically this procedure is used to check consistency in the initial conditions e g to test relations among parameters and initial values Since the simulationist may interactively change values of R 1
333. urrently active while it is drawn Unfortunately the simulation environment draws the legend in many situations for different reasons and the modeler can not directly control this drawing However if the modeler follows the following guidelines there should result a satisfying behavior The model which changes curve attributes dynamically during the course of a simulation must set all curve attributes exactly as they should appear in the legend at the end of the procedure Output if current time t t resp k k and always at the end of the procedure Terminate for an example see the research sample model LBM module LBMObs in the Appendix PROCEDURE ClearGraph Clears the panel of the graph 7 5 5 SIMULATION ENVIRONMENT MODES The following four procedures allow to define the so called simulation environment modes They can be used to set any preference PROCEDURE SetDocumentRunAlwaysMode dra BOOLEAN PROCEDURE GetDocumentRunAlwaysMode VAR dra BOOLEAN If the mode document run always is activated every execution of a simulation run will be documented onto stash file according to the current settings of the project descriptors Note that the stash file gets rewritten with every new run PROCEDURE SetRedrawTableAlwaysMode rta BOOLEAN PROCEDURE GetRedrawTableAlwaysMode VAR rta BOOLEAN The mode redraw table always describes the behaviour of the table window in respect to modifications of the tabulation monitoring set
334. urve PROCEDURE DrawLegend c Curve x y INTEGER comment ARRAY OF CHAR PROCEDURE Plot c Curve newX newY REAL PROCEDURE Move c Curve newX newY REAL PROCEDURE PlotSym x y REAL sym CHAR PROCEDURE PlotCurve c Curve nrOfPoints CARDINAL x y ARRAY OF REAL PROCEDURE GraphToWindowPoint xReal yReal REAL VAR xInt yInt INTEGER PROCEDURE WindowToGraphPoint xInt yInt INTEGER VAR xReal yReal REAL PROCEDURE TimeIsX BOOLEAN TYPE Abscissa RECORD isMV POINTER TO REAL xMin xMax REAL END PROCEDURE CurrentAbscissa VAR a Abscissa PROCEDURE TranslStainToColor stain Stain VAR color Color PROCEDURE TranslColorToStain color Color VAR stain Stain TYPE DisplayTime showAtInit showAtTerm noAutoShow VAR timeIsIndep REAL PROCEDURE DeclDispData mDepVar Model VAR mvDepVar REAL mIndepVar Model VAR mvIndepVar REAL X V vLo vUp ARRAY OF REAL n INTEGER withErrBars BOOLEAN dispTime DisplayTime PROCEDURE DisplayDataNow mDepVar Model VAR mvDepVar REAL PROCEDURE DisplayAllDataNow PROCEDURE RemoveDispData mDepVar Model VAR mvDepVar REAL CR REE EEE EEE ERKENNE SimIntegrate AIR I IN ROR IE abala diid adad PROCEDURE Integrate m Model from till REAL Ee RR TEE EI I EA A AE RS TabFunc I E eI IEEE I IE I EA RR RR IE TYPE TabFUNC PROCEDURE DeclTabF VAR E TabFUNC XX YY ARRAY OF REAL NValPairs INTEGER modifiable
335. used as its first order derivative in the case of continuous time or its new value in the case of a discrete time model The model equations are formulated as expressions capable of defining the values of the derivative or new value The expression may be an arbitrary function of any of the other model objects such as state variables auxiliary variables or model parameters Every state variable must be associated with a particular initial value and a range within which it may be changed interactively Fig T2 Every model may have any number of model parameters each associated with a particular value and a range within which it may be changed interactively Typically model parameters are not or only rarely changed in the middle of simulation experiments Fig T2 Intermediate results from an expression may be stored in a variable which will be later used in another expression Such auxiliary variables are often used to compute com plex expressions defining the value of a derivative of a state variable In a ModelWorks model definition program the modeler may use any number of auxiliary variables However in the current version ModelWorks does neither especially recognize such variables nor does it hinder the modeler to use them in whichever way he wishes Finally models may have any number of monitorable variables They are used to monitor the current values of any variable or otherwise accessible real numbers used in the ModelWorks model def
336. variables and their unit These strings will be used by the table function editor from within the simulation environment xMin xMax yMin yMax Define the upper and lower bounds for each axis values Attempts by the simulationist to enter values outside of this range will not be accepted by the table function editor 7 2 Accessing Defaults and Current Values During simulations ModelWorks uses many internal parameters settings and other variables the so called defaults and current values Fig T22 They can be accessed by the modeler in order to control simulations in a similar way the simulationist may access them One class of procedures lets the modeler retrieve values but not change them read only values e g the simulation time or the default independent variable Another class of procedures lets the modeler get and set values e g GetGlobSimPars or GetP respectively SetGlobSimPars or SetP The accessible values are grouped into several categories First there are the global simulation parameters project description variables variables controlling the stash filing and the variables R 106 ModelWorks 2 0 Reference associated with the models and the model objects Secondly each category exists in two copies the defaults and the current values 7 2 1 GLOBAL SIMULATION PARAMETERS AND PROJECT DESCRIPTION Among the global values controlling simulations there are internal read only variables Tab R1 and the global simulation
337. w f WindowFrame PROCEDURE RedrawTitle u Window title ARRAY OF CHAR PROCEDURE DummyRestoreProc u Window PROCEDURE AutoRestoreProc u Window PROCEDURE SetRestoreProc u Window r RestoreProc PROCEDURE GetRestoreProc u Window VAR r RestoreProc PROCEDURE StartAutoRestoring u Window r RectArea PROCEDURE StopAutoRestoring u Window PROCEDURE AutoRestoring u Window BOOLEAN PROCEDURE GetHiddenBitMapSize u Window VAR r RectArea PROCEDURE UpdateWindow u Window PROCEDURE InvalidateContent u Window PROCEDURE SetCloseProc u Window cp CloseProc PROCEDURE GetCloseProc u Window VAR cp CloseProc PROCEDURE GetWindowFrame u Window VAR f WindowFrame PROCEDURE GetWFFixPoint u Window VAR loc WFFixPoint PROCEDURE DoForAllWindows action WindowProc PROCEDURE UseWindowModally u Window VAR terminateModalDialog cancelModalDialog BOOLEAN PROCEDURE PutOnTop u Window PROCEDURE FrontWindow Window PROCEDURE RemoveWindow VAR u Window PROCEDURE RemoveAllWindows PROCEDURE WindowExists u Window BOOLEAN PROCEDURE AttachWindowObject u Window obj ADDRESS PROCEDURE WindowObject u Window ADDRESS A 175 ModelWorks V2 0 Appendix OPTIONAL MODULES UE E T HR T E RAN HN DM2DGraphs MSA E E L HEFTE EEE RM FR IHRE TYPE Graph Curve LabelString ARRAY 0 255 OF CHAR GridFlag withGrid withoutGrid ScalingType lin log neglog PlottingStyle solid slash slashDo
338. w State variables is opened respectively brought to the front Model parameters P The IO window Model parameters is opened respectively brought to the front Monitorable variables O The IO window Monitorable variables is opened respectively brought to the front Table T The table window is opened respectively brought to the front If there are no monitorable variables which have an active tabulation setting T this window may not be opened in the program state Simulating Clear table This command clears the table i e erases the content of the table window if it is currently open R 83 ModelWorks 2 0 Reference Graph G The graph window is opened respectively brought to the front If there are no monitorable variables which have an active graphing setting Y isY this window may not be opened in the program state Simulating Clear graph B Clears blanks all curves in the panel of the graph if the graph window is currently open If the latter condition is true it is the same command as Edit Clear see above Menu Edit 6 1 6 MENU SIMULATION If a simulation is started by any of the menu commands available under this menu Fig R10 ModelWorks will enter the state Simulating Fig T15 part II Theory Simulation Start run Halt run Pause Stop Kill run Execute Experiment E Fig R10 Menu Simulation Start run R Starts an elementary simulation run with the current settings of all values Previous
339. we favor over all other editors known to us commercial products included because of its powerful macro capability File is bundled with the wide spread MS Edit Configure your installation by deleting the already existing file FileSystem OBM it is actually a copy of file a in the folder and renaming the copy of any of these files to FileSystem OBM will result in the corresponding bundling I MEdit can be obtained from Matthias Aebi Hirschgartnerweg 25 CH 8057 Ziirich Switzerland It can also be obtained as part of the Dialog Machine release together with Modula 2 and ModelWorks supporting Macros from Projekt Zentrum IDA re Dialog Machine Swiss Federal Institute of Technology ETHZ ETH Zentrum CH 8092 Ziirich Switzerland shareware fee has still to be paid separately to Mr Aebi A 133 ModelWorks V2 0 Appendix C 4 HOW TO MAKE A STAND ALONE APPLICATION Typically MacMETH s linking loader is used when developing and running models The only disadvantage of this technique is that models require the Modula 2 development environment to be present always Once developed and tested the modeler may wish to make the model inde pendent from this development environment i e to convert a model definition program into a normal Macintosh application This can be achieved easily by linking a ModelWorks model definition program to a stand alone application The latter may be started with a double click like any other ordinary Macintosh app
340. will perform the standard Clear command as described in the Macintosh owner s guide Keyboard equivalents of these commands are available often even when the simulation environment is in a mode which prohibits choosing menu commands e g when an entry form is currently in display The meaning of these commands is then such that textual objects and not the graph are exchanged with the clipboard For a more detailed description of these commands see the first section of this chapter 6 1 4 MENU SETTINGS Settings set Global simulation parameters Project description Global simulation parameters Project description Stash file name Windows All model s integration methods All model s initial values All model s parameters All model s stash filing All model s tabulation All model s graphing All model s scaling All model s curve attributes Initialize Session All above Fig R5 Menu Settings used to set or reset current values This menu consists basically of three parts Set Select stash file and Reset Fig R5 R 76 ModelWorks 2 0 Reference The first part lets you set the global simulation parameters such as the simulation start and stop time plus the project description The second determines which file is going to be used as the stash file The third is used to reset the current values of global parameters settings and of model objects resetting means to copy the defaults to the corresponding cu
341. windows for models state variables model parameters and monitorable variables Fig T4 The menu bar has five ModelWorks menus each with several commands File lets you print graphs set preferences and quit the program Edit allows you to access the clipboard to transfer graphs of simulation results to other programs or to desk accessories Settings offers commands to set current values of the global simulation parameters or the so called project description plus the resetting of current values to their defaults Windows opens or activates the six windows of ModelWorks and Simulation is used to execute and control simulations In the visible windows the model objects of the activated model the grass growth model are displayed Throughout this manual read instructions as e g choose menu command File Execute as choose menu command Execute under menu File The windows initially displayed serve two purposes First they are used to display current values such as initial values or parameter values and secondly they are used to enter values or settings Hence they are called IO windows input output windows Here are the common characteristics of the four IO windows All IO windows display a button field in the upper left corner a list of objects in the middle and a scroll bar on the right side Any model object can be selected by a simple mouse click All subsequent clicks on the buttons refer to the currently selected object Selection of
342. wn second the color of a curve stain and thirdly the symbol with which points of a curve are marked Unless explicitly specified ModelWorks assigns curve attributes automatically which is therefore called the automatic curve attribute definition strategy For instance following this strategy ModelWorks assigns automatically the stain coal black to the first and ruby red to the second variable being activated for monitoring This helps the user to tell curves optimally apart under many circumstances However the automatic curve attributes definition strategy has also its disadvantages in particular the attributes may change all the time For instance in one graph the Grass is black in the other it is red but Grass derivative becomes black etc The actual color will depend only on the exact chronological sequence in which a monitorable variable has been activated for monitoring with the button X To try this out click on Grass derivative in the Monitorable variables window and toggle it with button X so that it becomes activated Y Run a simulation e g this time by pressing the command key clover leaf key simultaneously with key R Note that curve Grass is drawn in black unbroken and Grass derivative in red broken Then click on Grass in the Monitoring variable window and toggle it with the button twice Again both monitoring variables are activated Y Now rerun the simulation and note that this time colors are reversed i e
343. xactly these curve attributes i e in green regardless when and with how many other curves you currently display it To see this behavior click on Grass in the Monitoring variable window and toggle it with the button twice Run a simulation and note that Grass will be drawn 1 On IBM PCs there are no colors available Sorry but the memory limitations of MS DOS have forced us to sacrifice them 2 the third becomes emerald green and the fourth sapphire blue For more details see part Reference 3 Note that specifying a line style is crucial if you should omit it automatic definition of curve attributes would still remain active regardless of the settings of stains or plotting symbols T21 ModelWorks 2 0 Tutorial in green unbroken v and Grass derivative in black unbroken then click on Grass derivative and toggle it with the button once so that it will no longer be activated for monitoring Rerun the simulation and note that this time Grass is still drawn in green unbroken v Once again there is a disadvantage to this method if all your monitoring variables adopt it you will run more often than you may first think into a situation where several curves currently in display happen to be all of the same color or line style may be important on a no color only laser printer or a publication For instance click on Grass in the Monitoring variable window and click on the button EAR then select the line style broken
344. xperiment PROCEDURE Sensitivity i CARDINAL VAR j min max BEGIN FOR j min TO max DO SetP m p i v cur p i v j IF i lt n THEN Sensitivity itl ELSE SimRun END END FOR END Sensitivity BEGIN Sensitivity 1 END MyExperiment Note that the latter form makes it less obvious how fast such a structured simulation may grow to an enormous task given n model parameters each with k values and each combination is to be tested the number of simulation runs becomes k Our most simple example has n 3 pa rameters each with k 3 values min mean max yet for a full sensitivity analysis there are already 33 27 simulation runs needed 12 n order to be able to use blanks in the middle of a descriptor and still be able to write the data onto the text file in a free format use so called hard spaces Option space bar within a descriptor 13The objects TextFile GetExistingFile GetReal SkipGap ReadChars and Close are to be imported from the Dialog Machine module DMFiles T 70 Part III Reference This reference part contains a description of the usage of every feature ModelWorks offers However it contains only little information on the elementary and typical usage or the theoretical concepts of ModelWorks In case you should not be familiar with the basic concepts of ModelWorks please read first the ModelWorks tutorial In particular you should read the first chapter of the tutorial General Description
345. y t form for each submodel the global output vectors y t The output of the global system is given by combining the global output vectors y of the n submodels 1 yi t y 7 global output 10a DRU B A structured model composed of n elementary coupled submodels each discrete time and defined according to Eq 5 x is defined as follows for an example see also Fig T9 u u k global input 6b uj k hL u k SUK Ya input of submodel i 7b xi k 1 fi x k uk pri k k dynamic equations of submodel i 8b vik gj Lak peril k output of submodel i 9b yik X l global output 10b yik C A general structured model composed of n elementary coupled submodels each either continuous or discrete time and each defined according to Eq 4 x resp Eq 5 x is defined as follows for an example see also Fig T9 There are two different independent variables the continuous time t ft and the discrete time k E 3 The constant time step 1 of the discrete time submodel s is interpreted as a real time interval the coincidence interval c on the time axis t The discrete time submodels are only known at the endpoints of these intervals the coincidence time points or coincidence points The continuous time submodel s describe continuous or faster processes which occur between the coincidence points A communication between the two submodel types occurs o
346. y be called freely Note in particular that calling of SetSV in the procedure Initialize results in the immediate use of the new initial values for the current run Any new or changed values will be displayed in the corresponding IO windows However note that the updating of some changes may require some time before they become actually visible on the screen because the Dialog Machine may need several integration steps till all updates have been completed PROCEDURE PROCEDURE PROCEDURE PROCEDURE PROCEDURE PROCEDURE PROCEDURE PROCEDURE PROCEDURE PROCEDURE GetM VAR m Model VAR curMethod IntegrationMethod SetM VAR m Model curMethod IntegrationMethod GetSV m Model VAR s REAL VAR curInit REAL SetSV m Model VAR s REAL curInit REAL GetP m Model VAR p REAL VAR curVal REAL SetP m Model VAR p REAL curVal REAL GetMV m Model VAR mv REAL VAR curScaleMin curScaleMax REAL VAR curSF StashFiling VAR curT Tabulation VAR curG Graphing SetMV m Model VAR mv REAL curScaleMin curScaleMax REAL curSF StashFiling curT Tabulation curG Graphing GetTabF t TabFUNC VAR XX YY VAR NValPairs VAR modifiable VAR tabName xName yName xUnit yUnit ARRAY OF CHAR VAR xMin xMax ARRAY OF REAL INTEGER BOOLEAN yMin yMax REAL SetTabF t TabFUNC XX YY ARRAY OF REAL NValPairs INTEGER modifiable BOOLEAN tabName xName yName xUnit yUnit
347. y be separated by their different symbols v resp none T 22 ModelWorks 2 0 Tutorial 2 2 7 CHANGING INTEGRATION METHODS To start with this section reset the program to its initial state with Settings Reset All above The numerical integration of the differential equation has to be done with special integration algorithms ModelWorks offers several different methods for numerical integration Each has its particular advantages and disadvantages The default algorithm used in this example is Euler which is shown in the model window We shall compare two integration methods and record the results on the stash file First we have to define the stash file output Bring the window for monitorable variables to the front select the variable Grass and click on Toggle function In the column Monitoring appears the letter F for stash filing this is in addition to the T and Y which signify that this variable is written already into the table and drawn in the graph Now during a simulation run the values of the variable Grass will be written also onto the stash file Note by default every new simulation will overwrite the stash file s content Differences between integration methods become more obvious with large integration step sizes this is the step which is internally used for numerical integrations Therefore we change this step to a higher value Select the menu command Settings Global simulation parameters and change in t
348. y to the program substates shown in Fig T26 part II Theory If the value noSubState is returned no experiment is currently running i e the simulationist has reached state simulating by choosing the menu command Simulation Start run s a below procedure ExperimentRunning R 114 ModelWorks 2 0 Reference PROCEDURE ExperimentRunning BOOLEAN ExperimentRunning from SimMaster returns TRUE if a structured simulation experiment is currently in execution i e if the simulationist has reached the state simulating by choosing the menu command Simulation Execute experiment s a above procedure GetMWSubState PROCEDURE ExperimentAborted BOOLEAN ExperimentAborted from SimMaster returns TRUE if the simulationist has stopped killed a running structured simulation experiment A typical use of this procedure is to skip superfluous calls to procedure SimRun E g REPEAT SimRun UNTIL CurrentSimNr maxRunNr OR ExperimentAborted 7 5 Display and Monitoring 7 5 1 WINDOW OPERATIONS The following Modula 2 objects serve to control the display on the screen e g the arrangement of windows or the monitoring PROCEDURE TileWindows PROCEDURE StackWindows The two procedures stack or tile windows on the screen Stacking is with overlapping windows similar to the ModelWorks predefined start up display Tiled windows don t overlap and fill the screen as much as possible The actual arrangement may depend on the screen in displ
349. yMax 1 0 Errors If the point specified by xReal and yReal lies outside the integer pixel range DM2DGraphsDone will be set to FALSE and xInt and yInt is set to MIN INTEGER or MAX INTEGER respectively PROCEDURE TimeIsX BOOLEAN Returns TRUE if time is the current setting of the x axis otherwise FALSE TYPE Abscissa RECORD isMV POINTER TO REAL xMin xMax REAL END PROCEDURE CurrentAbscissa VAR a Abscissa Returns a pointer isMV to the monitoring variable currently used as abscissa and its extremes xMin curScaleMin xMax curScaleMax In case that time is in use isMV will point to timeIsIndep Procedures to convert different Color Types X PROCEDURE TranslStainToColor stain Stain VAR color Color A 143 ModelWorks V2 0 Appendix PROCEDURE TranslColorToStain color Color VAR stain Stain Translates Stain from module SimBase to Color from module DMwindowIO and vice versa exception for TranslStainToColor autoDefCol is translated to black Display data series i e for validation all at once X Follow these steps to use the data display feature of that module 1 Declare an ordinary monitoring variable with the procedure DeclMV as a master monitoring variable for data arrays to be declared later see next step Several properties i e descr ident unit and curve attributes as color linestyle sy
350. yrights are reserved and are held by the authors and the Swiss Federal Institute of Technology Z rich ETHZ ModelWorks may not be sold nor included in any sold product as an incentive nor otherwise redistributed for a profit without prior written consent by the authors and the Swiss Federal Institute of Technology Z rich ETHZ INote that for heavy computations the gain in computing speed is substantial up to two magnitudes for certain functions such as Ln up to several hundred times faster this is because this implementation bypasses SANE and accesses the mathematical coprocessor directly hereby making maxium use of its power without sacrificing much in numerical precision A 125 ModelWorks V2 0 Appendix B Hard and Software Requirements B 1 MACINTOSH VERSIONS The standard and the Reflex Macintosh ModelWorks versions V2 0 resp V2 0 Reflex run on all Apple Macintosh computer models except the 128K Mac and the 512K Mac Fat Mac unless the machines should be equipped with ROM versions of 128K Bytes or later and with sufficient memory see below All three versions require at least one 800 KBytes double sided floppy disk drive and either a second 800 KBytes disk drive or a hard disk For serious model development a hard disk is recommended You have to use a Finder version 5 3 or later and a System version 3 1 or later and it is recommended to work with HFS Hierarchical File System disks only The Reflex version V2

ModelWorks - Terrestrial Systems Ecology

Contents

Download Pdf Manuals

Related Search

Related Contents