SMTET documentation
Contents
1. file D Georgia Tech Javajars Savannah 0 99 jar lt jar file gt Note Java uses forward slash instead of backslash characters to specify the path b treatment o startTime Start time for the simulation Not used in the current version o TimeUnit Time units for the simulation This should be kept in seconds as for the time being some model parameters are included in the code in this time unit warmupPeriod System statistics will be discarded at the end of this time period runLength Duration of a run numberofreplication This parameter is a DSOL parameter not used in our project properties Properties are variables that are fixed for each treatment experiment These variables have the following sequence of elements Only one of these fields is currently read by the Port of Savannah model Key simulation totalreplications oo0 0 Value The number of replications to be run Key simulation tracetofile Value True or false This parameter defines whether a log file is created or not Key simulation enhanced animation Value True or false This parameter defines whether enhanced animation can be played from the animation menus Key simulation enhanced animation class Value gatech ece savannah Animation3D The user should not change this value since it specifies the name and location of the Java class file that will run the enhanced animation Key simulation map file Value Name of the XML file that defi2. timeUnit default UNIT maxOccurs 1 minOccurs 1 type smtet timeUnit gt lt xsd element name warmupPeriod maxOccurs 1 minOccurs 1 type smtet period gt lt xsd element name runLength maxOccurs 1 minOccurs 1 type smtet period gt lt xsd element name recordlog maxOccurs 1 minOccurs 0 type xsd boolean gt lt xsd element name numberofreplications maxOccurs 1 minOccurs 1 type xsd positiveInteger gt lt xsd element name properties maxOccurs 1 minOccurs 0 type smtet properties gt lt xsd sequence gt lt xsd complexType gt lt xsd complexType name properties gt lt xsd sequence gt lt xsd element name property maxOccurs unbounded minOccurs lt xsd sequence gt lt xsd complexType gt lt xsd complexType name property gt lt xsd complexContent gt lt xsd restriction base xsd anyType gt lt xsd attribute name key type xsd string gt lt xsd attribute name value type xsd string gt lt xsd restriction gt lt xsd complexContent gt lt xsd complexType gt lt xsd complexType name period gt lt xsd simpleContent gt lt xsd extension base xsd double gt lt xsd attribute name unit type smtet timeUnit use required gt lt xsd extension gt lt xsd simpleContent gt lt xsd complexType gt lt xsd complexType name model gt lt xsd sequence gt lt xsd element name model class maxOccurs lt xsd element name class path maxOccurs lt xsd seque
3. to interpret because of the use of realistic 3D imagery On the backend the 3D animation system interfaces with the DSOL system using the API to exchange detailed information about the state of the simulation It keeps track of all state changes allowing for playback at different speeds and times see Section 2 Scaled models of all mobile entities are loaded from the commonly used LightWave OBJ format and allow for detailed customization as necessary DXF or other formats can be converted The test bed enables various groups such as emergency management agencies military logistics planners and Port management to a create and test alternative operational strategies sensor technologies e g RFID and various other applicable technologies and b utilize the SMTET to conduct disaster preparedness or other emergency situation training that is neither possible nor practical to do at the facility itself Further the test bed allows users to train or upgrade employees on new software for Port operations One example of this is the need to train new employees on the terminal operating software in use at the Port The employees could attend an in depth training class at the Logistics Innovation Center LINC and get real world situational training for the same system they will be using at work The current model was built based using a map and dimensional measurements taken from Google Earth The files are currently located at the site http www isy
4. Simulation Model and Technology Evaluation Test bed SMTET User s Manual The scope of SMTET is the development of a state of the art high fidelity simulation model for the Port of Savannah with a 3 dimensional true to scale animation Off the shelf simulation products incorporate impressive animation capabilities but often place severe constraints on implementing various modeling aspects These limitations along with the desire for flexibility and ownership of the test bed motivated the development of an engine based on a publicly available web enabled programming language with object oriented capabilities such as Java Amongst the various Java based simulation packages we chose the Distributed Simulation Object Library DSOL ver 2 0 as the backbone because it has been used to develop large scale distributed simulation models for port operations e g the Port of Rotterdam in The Netherlands DSOL is a full featured discrete event simulation environment developed at Delft University of Technology The Netherlands and is available at http sourceforge net projects dsol The animation is an extension to the existing DSOL based system It is likewise implemented using Java and utilizes the cross platform Java 3D and Java Advanced Imaging API Application Programmer Interface for graphics It presents a 3D view of the same information presented by the existing 2D visualization However it is improved with better usability and is easier
5. e Jockey Trucks the Rubber Tired Gantry RTG cranes the Toplifters and Cranes are similar Controt Experiment Description Trucks Trains Vessets Jockey Trucks RIGS Topamers Cranes Gate Gates Yards Customs Output About Speed miles tu Loading Time per Container Triangular D striberion in mintes Lower Mose Unloading Time per Container Triangadar Distribution in minados Lower Mode Figure 1 Data for Trucks 2 Page 2 3 Trains Tab This tab is used for entering information related to the schedule of arriving trains their contents and the unloading loading times per container Figure 2 The latter times are modeled using the triangular distribution The day is divided into hourly slots The number of arrivals in each slot should be 0 or 1 Clicking in a time slot activates a dialog where the user can enter the speed of the train the number of rail sections and the number of containers in each section CO SMTET Main Window Fie Tools Window Help Controt Experiment Description Trucks Trains Vessets Jockey Trucks RIGS Topitters Cranes Gates Gaes Yards Customs Output About Loading Time per Container Triangular Distribution in minses Lower Mode Upper Untoading Time per Container Triangutar Distribution in minutes Lower Mode Upper r J1 a1 teamber of Arrivals in Hourty Time Slots Figure 2 Rail Data 2 4 Vessels Tab This tab is used for enteri
6. e gatech edu ywu60 smtet but will be moved to a powerful Linux server soon The site contains the Java Web Start file smtet jnip and the scenario file currently SavannahXmlMap xml The application can also be launched directly from the site Under Mac OS X Java Web Start installs an application app file in the Applications directory 1 Page 1 Requirements e Operating System Windows XP Vista or 7 Macintosh OS X 10 5 or higher Linux and Unix e Software Requirements o Version of Java JRE 1 5 o OpenGL ver 1 2 or later e Hardware Requirements o 3 0 GB of RAM o Dedicated graphics card with at least 256 MB of memory 2 Running the Application The input parameters for the simulation run can be assigned by creating an xml file with structure similar to SavannahXmIMap xml of by using the tabs in the application s Main Window as illustrated in the following headings The scenario file is selected from the File gt Open File menu in the Main Window Or you could select the embedded scenario file from the Help gt Tutorial gt SavannaXmIMap xml menu in the Main Window 2 1 Control Tab This tab allows the user to input the run length the number of independent replications the time step and the animation delay 2 2 Trucks Tab This tab allows the user to input the speed and number of trucks and parameters for the triangular distributions for container loading and unloading times see Figure 1 The tabs for th
7. ely use the time slide to move to the desired time o Time Scale This slide is used to control the speed of the animation independently of the speed of simulation e Location Controls The animation perspective can be changed by means of the following controls o Moving the mouse pointer o Zoomin or out Use mouse wheel 5 Page Move forward W Move backward S Move left A Move right D Orbit left clockwise Q Orbit right anti clockwise E Orbit down grazing angle Z Orbit up X O 0 0 0 0 0 O 0 Note Holding the shift key will increase the speed of movement rotation Figure 6 below depicts a vessel along with the assigned cranes a nearly container yard and several Rubber Tired Gantry RTG cranes Figure 6 A Vessel and a Nearby Container Yard 3 Output Statistics Tab Output contains point estimates and 95 confidence intervals for various means estimated by means of discrete and continuous time statistics Figures 7 8 Confidence intervals are generated only when the number of replications exceeds one but they practically require several replications to be meaningful 6 Page Conson Experiment Description Tracks Trams Vessets Jockey Trucks RTGS Toplimers Cranes Gate3 Gates Yards Customs Output Aho Descrote Statestics 94 Confidence Interval Tine to maw a Container from train to yard Tine to move a container trom yard to traire Tir to move a container fram v
8. essel to yard Time to mow a Container from yard to vasset Tine a truck spent at checkin quote Time a truck spert at check ine Tine a truck spent at chech out queue Time a truck spont at check out Time to unload trai Time to toad a train Time to unload a vessel Time to baad a voest Time for customs inspection DIG db Time o nat Repth Troatit Controt Experimemt Description Tucks Irans Vessets Jockey Trucks RTGS Topaners Cranes Gate3 Gates Yards Customs Output Abou Discrete Continuous Continuous SICS Confidence Interval j Gmod Gatos RIG Crane 1 Average Number of Tracks at Chack ts Qum 0 000 Average Number of Trucks at Check out Queue 0 000 i Average umber of Trucks at Check t Quove 0 000 Average Number of Trucks at Check out Queue 0 000 1 Average Number of Trucks at Check te Queue 0 000 Avet age Humber of Tracks at Check out Queue 0 000 Average Number of Tracks at Check in Queue 0 000 Avago Number of Trucks at Chock out Queun 0 000 i Average Number of Tracks at Chock Quoue 0 000 Average Number of Trucks at Chock out Queue 0 000 1 Average Humber of Trucks af Check in Queue 0 000 Average Number of Trucks at Check out Gueue 0000 1 Average Number of Trucks at Check in Queur 0 000 Average Number of Trucks at Check out Queue 0000 i Average Number of Trucks at Check in Queue 0000 Average Number af Tracks at Check out Queue 0000 Average Humber of Tr
9. nce gt lt xsd complexType gt lt xsd complexType name classPath gt lt xsd sequence gt lt xsd element name jar file maxOccurs unbounded minOccurs 1 type xsd string gt lt xsd sequence gt lt xsd complexType gt lt xsd complexType name replication gt lt xsd sequence gt lt xsd element name stream maxOccurs unbounded minOccurs 1 type smtet stream gt lt xsd sequence gt lt xsd attribute name description type xsd string use required gt lt xsd complexType gt lt xsd complexType name stream gt lt xsd attribute name lt xsd attribute name lt xsd complexType gt lt xsd simpleType name timeUnit gt lt xsd restriction base xsd string gt MINUTE gt HOUR gt lt xsd enumeration value DAY gt lt xsd enumeration value WEEK gt lt xsd enumeration value YEAR gt lt xsd enumeration value UNIT gt lt xsd restriction gt lt xsd simpleType gt lt xsd schema gt 1 minOccur 1 minOccur 1 name model type smtet model gt 1 name treatment type smtet treatment gt type smtet property gt minOccurs minOccurs type xsd string gt type smtet classPath gt xsd string use required gt type xsd positiveInteger use required gt lt xsd enumeration value lt xsd enumeration value 9 Page
10. nes the port lay out 8 Page 4 2 XSD file The following is the schema verification xsd file that is used to check the parameter file for consistency lt xml version 1 0 encoding UTF 8 gt lt xsd schema xmlns xsd http www w3 org 2001 XMLSchema xmlns smtet http www isye gatech edu targetNamespace http www isye gatech edu gt lt l PHEPHHPPEPETHEEHPAE HAE R EEEH HHH DEFINITION OF DATABASE tttttt ttt tet te tate ttt t gt lt xsd element name simulationparametersdatabase gt lt xsd complexType mixed true gt lt xsd sequence gt lt xsd element maxOccurs unbounded minOccurs 1 name experimentDef type smtet experimentalFrame gt lt xsd sequence gt lt xsd complexType gt lt xsd element gt lt l PARERE EREEREER REEERE REEERE TYPE DEFINITIONS Ht EER REHEAT HEE HEE ETHER HEHEHE ATA RE HRT HEE E EE gt lt xsd complexType name experimentalFrame mixed true gt lt xsd sequence gt lt xsd element maxOccurs unbounded minOccurs 1 name experiment type smtet experiment gt lt xsd sequence gt lt xsd complexType gt lt xsd complexType name experiment mixed true gt lt xsd sequence gt lt xsd element maxOccurs lt xsd element maxOccurs lt xsd sequence gt lt xsd complexType gt lt xsd complexType name treatment mixed true gt lt xsd sequence gt lt xsd element name startTime maxOccurs 1 minOccurs 0 type xsd dateTime gt lt xsd element name
11. nfunctional 0 2 SMTET Main Window om File Tools Window Help Control Experiment Description Trucks Trains Vessels Jockey Trucks RIGS Topliters Cranes Gate3 Gated Yards Customs Output About Check in Time per Truck Triangular Dtstriution in minutes Lower Mose upper Check out Time per Truck Triangular Distribution in minutes Lower Mose Upper Number of Arrivats in Hourty Time Stots Hour 12 8 WS 6 p w hom 12 13 ha hs he hr he hs 20 21 22 23 ae Rate 50 Jeo 40 60 60 70 8 z0 60 40 40 50 6 fa ie TEB BTI Gmo Schodute 0 off 1 in out Figure 5 Tab for Gate 3 2 7 Customs Tab This tab is used for setting the fraction of containers undergoing customs inspection and the parameters of the triangular distribution for the inspection time per container This inspection is not displayed in the current animation it is modeled only as delay 2 8 Model Execution To run the model hit the Play button at the bottom left corner of the application window The 3D animation window is activated by the Window gt 3D Animation menu item or the F5 key The animation is onset by clicking on the Play button on the 3D Animation window The following controls are available to the user e Time Controls o Play button Start the animation o Time Step While the Play button is in Pause mode type the desired time past or future not to exceed the simulation time in the time indicator Alternativ
12. ng information related to the schedule of arriving vessels their contents and the unloading loading times per container Figure 3 The latter times are modeled using the triangular distribution The day is divided into hourly slots Clicking in a time slot activates a dialog where the user can enter the berth for each vessel the number of containers and the numbers of cranes and jockey trucks assigned to the vessel 3 Page Loading Time per Container Triangular Distribution in minutes G R 7 03 10 05 Unloading Time per Container Triangutar Distribution in manates Got 02 Figure 3 Vessel Data 2 5 Yards Tab This tab allows the user to input the initial contents of each container yard by row and block as well as the yard at the Intermodal Container Transfer Facility ICTF See Figure 4 bnttial Contents of CTF Yard Figure 4 Initial Yard Contents 2 6 Gate Tabs The gate tabs allow the user to input the operational parameters for Gates 3 and 4 The check in and check out times per truck follow triangular distributions The arrival processes for trucks at each gate 4 Page follow the nonstationary Poisson model with rates that can vary on an hourly basis Under this model the number of arrivals during a given time window follow the Poisson distribution with mean equal to the area under the rate function during that time window Finally the user can designate each lane as ingoing 1 outgoing 2 or no
13. ucks at Checkin Gueue 0000 Average Number of Tracks at Check out Queue 0000 Lane 10 Average Number of Tracks at Check tn Queue 0 000 oem age Number of Trucks at Check out Queue 0 000 Lane 11 Average Number of Trucks at Check tn Quaut 0 000 vetage Numbes of Trucks at Check out Queue 1 000 Lane 12 Average Number of Trucks at Check tn Queue 0 000 Average Number of Trucks at Check out Queue 0 000 Lane 13 Average Number of Trucks at Chock te Quaus 0 000 Average Humber of Trucks at Check out Quau 0 000 Lame 14 Average Number of Tracks at Chock ln Queue 0 000 Average Number of Trucks at Check out Quoue 0 000 gt DAG Dh Time o null Figure 8 Continuous time Averages and Confidence Intervals Additional statistics and graphs can be viewed from the Window gt Statistics menu item F2 key This action opens a new window with a customizable number of rows and columns Each entry of the window can be filled with graphs of estimates by dragging contents from the left single column to the window 7 Page 4 Appendix 4 1 Parameters a experiment o name Name of the experiment o Model Location of the Main class of the model This parameter should not be changed since it is the location of the Main class in the jar file The value is gatech isye sim smtet scenario Scenario o classPath jar file Location of the jar file This parameter shall have the location of the model jar file e g lt jar file gt
Download Pdf Manuals
Related Search