Handout - Murat Yildizoglu`s
Contents
1. Computations with agents 2 e We can also easily compute statistics and aggregate indicators about agents properties e We execute in this case operator a breed variable of plural breed name e a breed variable of plural breed name returns a list containing the value of the cor responding variable for each member of this breed Average profit of firms mean profit of firms Maximal profit of firms max profit of firms Also other operators can be used min median standard deviation variance Computations with agents 3 e We can do any computations using a breed variable of the agents of the same breed e show who of firms gt 0 3 2 1 random order e show sort who of firms gt 0 1 23 e show sort who who of firms gt 0 1 4 9 e show sort sales costs of firms gt 100 110 115 150 Computations with agents 4 e Collecting the list of agents of the same type verifying a desired condition using with e let champions firms with profit gt 1000 e Collecting the list of all ships placed on the same patch as the turtle e let friends other ships here e Creating an ordered list of agents given a criterium let increasing price firms sort by Iprice of 71 lt price of 72 firms e And execute instructions on agents selected the same order instead of arandom order foreach increasing price firms ask show price 2 The common structure of NetLogo models Main blocs of a2. typical NetLogo program 1 Declarations of global variables agent breeds and breed variables for each breed 2 setup procedure a Initialization of global variables values read from the GUI or fixed in the code b Creation of the populations of each agent type breed and initialization of their breed variables c Initialization of the ticks and of different output forms graphics CSV files etc 3 A procedure go that groups all operations that take place in each period of execution of the model and increments the period counter ticks Other common components e Complementary procedures that manage the behavior of different types of agents collect data and statistics on these behavior and on aggregate properties of the model update the outputs of the model saving the data file etc e A GUI that contains different elements that allow to the user to set the values of the parameters of the model to observe the evolution of the results from period to period e Comments for documenting the model Initialization and updating of the plots e We use the plots to observe the evolution of important variables in real time e Plots are created in the GUI e NetLogo 5 automatically updates the plots using the initialization of the plots when the commands reset ticks or setup plots are executed in procedures setup and go and the update of the plots when commands reset tick
3. Workshop ABM CEF 2015 Pre conference workshop on Agent based Models in Economics and Finance Chapter 2 Introduction to NetLogo Murat Yildizoglu http yildizoglu info Universit Bordeaux GREOA UMR CNRS 5113 CEF 2015 Conference Taipei NetLogo e A programming platform with a graphical interface GUI and a meta language dedicated to ABMs e GUI three panels Interface an interface for interaction with the model and for representing the results graphics spatial structure of the agents space networks etc Information an interface to give information to the users of the model Procedures an interface for writing the procedures of the model code for agents behaviors their interactions and system dynamics e Important stuff is in the Procedures panel while funny stuff happens in the Interface panel Main agent types A model in NetLogo may contain the following main types of agents e the World represents graphically the complete spatial environment of the model It can be in 2D or 3D e patches are the elementary spatial components of the World Agents move over the patches and can read the properties of patches the patches can have variables e turtles are individual agents in the model They have properties variables they can move in the World they can give birth to other turtles hatch an egg and they can die They can be of different breeds with different sets of properties wol
4. lor to green e Create the lambs population given their initial number fixed in the GUI lambs number e Place the lambs on random spots in the World Step 3 Go 1 Move lambs gt move lambs procedure 2 Let them eat the grass present at the arrival patches eat grass Check if some lambs are dead check deaths Living one reproduce reproduction Count the current lambs population and the current quantity of grass Grow grass on the patches grow grass ND om BR w Increment the period counter tick
5. ls globals price nbAgents As a local variable in a procedure or program or a code bloc with the instruction let let prix initial 10 As a property of a specific breed of turtle see below Variables 2 e We modify the value of an existing variable using the instruction set set initial price 0 5 set my prices 1 5 10 25 set my name Toto set price item 0 my prices set my prices replace item 1 my prices 15 show my prices 1 5 15 25 Declaration of turtle breeds and of their properties e Declaration of agent breeds breed plural name singular name breed firms firm e Declaration of individual properties of agents of a specific breed firms own profit each firm has a property called profit and it can contain a specific value for each one A variable that counts ticks e NetLogo models take place in discrete time advancing step by step NL includes a counter that memorizes the number of steps periods already executed ticks e We can read and use the value of this counter using its name let this period ticks We initialize its value at the start of the time using the instruction reset ticks executed in general at the end of the setup procedure e We increment its value by one using the instruction tick executed in general in the go procedure just after the operations corresponding to the end of a period 1 2 Procedures Declaration of specialized procedures e We regrou
6. p in specialized procedures sets of instructions that we may need in different mo ments in the program or have to execute them several times in a period e NetLogo includes two procedure types Commands are used to modify the values of global variables or variables given to the as parameters without returning any result Reporters return a result at the end of the operations they execute We need to save the returned value in a variable or directly use in other computations example replace item returns the modified list Commands e Their declaration starts with the instruction to and ends with the instruction end to my command argument1 argument2 instruction1 instruction2 end e Later in the program we can call this command my command arg1 arg2 e to execute its instructions using the values passed as arguments Reporters e Their declaration starts with the instruction to report and ends with the instruction end to report my reporter argument1 argument2 let my result 0 instruction instruction2 set my result report my result end e Later in the program we can call this reporter for setting the value of a variable using its result set my variable my reporter arg1 arg2 e my variable contains now the last computed value of my result Structure of a command my prices 1 5 10 25 set my prices replace item 1 my prices 15 r C NZ final list element tobe changed initiallis
7. s in setup tick or update plots in go are executed Help Useful resources are e The documentation of NetLogo Menu Help User manual and especially e the Dictionary of NetLogo Menu Help Dictionary e and the Programming guide in the User manual Homework e Check the tutorials in the Learning NetLogo section of the NetLogo help e Check in the NetLogo dictionary the meaning of each term indicated in blue in these slides 3 First example A very simplistic predator prey model A very simple model With Lambs running on then grass and eating the grass on the green patches with e grass randomly growing on patches with a probability of 3 a patch becomes green lambs eating the grass from the patches on which they are placed the patch becomes black and the lamb get some energy from grass lambs moving randomly choose a random direction and advance one step each step costs one unit of energy e lambs making a baby if they have enough energy to give birth birth energy and the babies taking this energy from their parents e lambs dying when they do not have any energy left Step 1 Declarations and setup e Declaration of global variables Number of lambs nbLambs Quantity of grass grass e Declaration of the agent breed lambs e Declaration of the breed variables only one here energy Step 2 Setup e Clear the memory and reset the ticks e Setup patches by setting their co
8. t new value my prices 1 5 15 25 1 3 Controlling the flow of the program Repeating instructions e To repeat n times a set of instructions repeat n instructions e To repeat a set of instructions as long as a condition is fulfilled it evaluates to true and to stop as soon as it becomes false while condition instructions while any other turtles here fd 1 The turtle that executes these instructions goes forward fd forward as long as it finds another turtle on the patch where it arrives Setting the flow depending on conditions e Executing certain instructions if and only if some conditions are fulfilled e We use instructions if or ifelse e if condition instructions instructions executed only if condition true e ifelse condition instructions1 instructions2 if condition true in structionsl are executed otherwise instructions2 are executed e let x 10 ifelse x gt 0 show sqrt x print negative number 1 4 Computations with agent lists Doing computations with agents properties e We can use the plural name of a breed firms to address the set of all agents of this type e we can then ask them to execute some instructions and each one of them will execute them in its turn determined in a random order by NetLogo ask firms set price market price fix production execute the command that computes the output compute profit execute the command that computes the profit
9. ves sheep consumers firms etc e links connect two individual agents turtles They can be oriented or not Network models are easy in NetLogo Other elements A NetLogo model may also contain global variables visible by all agents agent sets containing a list of agents of the same breed specialized procedures dedicated to operations used by the model including setup and go comment lines that must all start with a semi colon This is a comment 1 Elements of language Lists 1 1 Lists play a central role in NetLogo We delimit the elements of a list with square brackets and separate these elements with at least one space the comma is not the separator in NL A list of three numbers 0 5 1 2 A list containing two lists 0 5 1 2 2 5 1 33 25 An heterogeneous list 0 5 2 5 1 33 25 P Paul Jacques Aliye An empty list We can also build a list using the instruction list instruction with parentheses list 0 5 1 2 We refer to specific members of a list using the instruction item item 0 0 5 1 2 gt 0 5 Other useful instructions on the lists replace item fput lput sentence map reduce sort sort by Variables and breeds Declaring and using variables Variables can be declared In the GUI when we create controls a text input field a slider a selection combo etc At the start of the program as a global variable declared int he list called globa
Download Pdf Manuals
Related Search