Multiagent Traffic Management, traffic lights

• Sample: run the heatbugs.htmlHeatBugs demo in Netlogo, MASON, and Repast (also sugarscape)
  1. APIs:
     • MASON classdocs, main site
     • NetLogo primitives, programming guide, and main docs page. Models library
     • RePast API, main site
     • Sante Fe Institute - working papers on agent based modeling research

  2. Make an Applet of your current model
     • NetLogo
       1. Copy your .nlogo source code into web-docs
       2. Load this source code into NetLogo, choose Save As Applet in the File menu
       3. Check if this new .html file that has been created loads correctly in a browser
          If not, edit the archive= and value= lines in the applet tag and add a complete url path:
          "http://www.tjhsst.edu/~yourname/..."
       4. Samples: Aaron Salls, Mr. Latimer's
     • MASON
       1. Copy mason.jar into a directory in web-docs
          (or copy mason.jar if it exists in your mason directory)
       2. In this same web-docs directory, create directories /sim/app/schelling/ for example.
       3. Copy all your code from your actual mason/sim/app/schelling/ directory
          (this example is for the SchellingWith UI code) into this web-docs version
       4. Sample html file for this Schelling applet
       5. If you want more info, see How to Make an Applet from the How To page
     • RePast RePast creating an executable

  3. Starter assignments to try:
     • drawtree.nlogo - sample starter NetLogo program to draw a fractal tree.
     • Variations on Conway's Life (scroll down)
     • Modeling civil violence
       • Modeling civil violence: An agent-based computational approach
       • Modeling civil violence: An agent-based computational approach by Axtell/Epstein - the paper for the article above
     • More papers from Brookings Institute on modeling and social policy

  4. See Hanneman's Schelling's model of residential segregation
     • To run: download .jar file (scroll down), run with java -jar schelling_random.jar
     • To write code:
       • download java files - use "unzip"
       • easy version (without using java packages): comment out the "package" line, and compile using javac, all in the same folder.
     • First assignment: implement one of Hanneman's suggestions at the bottom of the page.
     • In Simulation.java, I changed the calculation of greenHappy and redHappy to (for greenHappy):

       	if (redCount < (max*(greenTol/100.0)))	greenHappy++;
       
           

  5. RePast modeling package. To run: repastj.sh (Java version) or repastpy.sh (Python/XML version)
     Repast Python version: java -jar /usr/local/repast/RepastPy/repastpy.jar (schelling version included)
     See the Sugarscape model as a good example (RepastJ)
     Repast study guide
     RePast examples, and Tutorials
     To edit and recompile for RePast:
     • Make a directory for your projects, for example: mkdir Repast
     • Make a directory for each different Repast project:
       • cd Repast
       • mkdir MyLife (start with the Conway's Life program, for example)
     • copy the Repast Life files (or whichever Repast project you're working with) into your new directory:
       • cd MyLife
       • If you want to see the Repast java source files available:
         ls /usr/local/repast/RepastJ/demos/src/uchicago/src/repastdemos/
       • cp /usr/local/repast/RepastJ/demos/src/uchicago/src/repastdemos/life/*.* .
       • In each source file comment out the package line:
         //package uchicago.src.repastdemos.life;
       • compile:
         javac -classpath /usr/local/repast/RepastJ/repast.jar:. LifeModel.java
       • This should compile all your Java source files.
     • Change something in the source file. For example change the width and height to 10, and infinite universe to false:

            public void setup() {
               super.setup();
               width = 10;//50;
               height = 10;//50;
               percentFull = .75f;
               infinite = false;//true;
               initialPattern = new RPentSetup();
             . . .
             

       • recompile this file with the command from the previous step.
       • Load this new version. In Repast, select open model, and Add. Find your new LifeModel.class.
       • Select open, and run.

  6. To run RePast in Python: java - jar /usr/local/repast/RepastPy/repastpy.jar
     There is a Schelling model in Python/XML

  7. Implement predator/prey prototype model, also see all Hanneman's chapters

  8. Here's MASON's Schelling model. Schelling Segregation Model and Sugarscape
     To "install" this Schelling model:
     • install MASON, download the mason.tar file
     • cd mason/sim/app/
     • mv the schelling.tar file to mason/sim/app
     • tar xvf schelling.tar
       this creates the schelling directory: mason/sim/app/schelling/
     • cd schelling
     • compile this java code for schelling: javac SchellingWithUI.java
     • go back to the mason directory: cd ../../../
     • run the schelling code: javac sim.app.schelling.SchellingWithUI
     • OR use: java sim.display.Console
       this starts the MASON generic interface