AI Spring 2009

  1. Course webpage
  2. Agent based unit -
    1. Schelling neighborhood Assignment March 9 (OPEN WITH OPENOFFICE for best results - ABIWORD distorts the format, may not print)
      1. tk_demo.py, Mr. Torbert's original demo, moving rectangle/oval
      2. tk_schelling.py, extended demo for this problem. This sample constructs a grid of patches, and each patch can hold 0 or more agents. The overall window size is 800 width, 600 height. Each patch is 25 by 25. This grid has dimensions 200 by 200, which results in an 8 by 8 grid of these patches.
        • Example run:


        • Example trace using initial density = .5:



          with this trace for accuracy of colors on particular patch/grid locations (load with OpenOffice for correct formatting)

      3. Initial assignment, count the number of neighbors each agent has, then the number of dissimilar agent neighbors
        • Determine if an agent is "content" in his/her neighborhood. For now, use these rules, which are considered "moderate" (each agent wants something more than one-third of his/her neighbors to be of the same type). If a person has: ***BLACKBOARD ASSIGNMENT "SCHELLING PRELIMINARY" - see Schelling preliminary assignment (OPEN WITH OPENOFFICE for best results - ABIWORD distorts the format, may not print)
          1. one neighbor, it must be the same color
          2. two neighbors, one must be the same color
          3. three, four, or five neighbors, two must be his color
          4. six, seven, or eight neighbors, at least three must be his color
      4. Animation shells:
      5. FOR A FINAL VERSION: Run experiments testing how quickly (whether or not) the neighborhoods segregate according to variation of parameters for the program. Use as a general form example - Schelling experiment writeup form ***BLACKBOARD ASSIGNMENT "SCHELLING EXPERIMENT"
      6. Extension assignment if you finish early - Axelrod's model for dissemination of culture.
        • Run this sample applet model demonstrating the concept of how cultures interact with each other and become more similar, perhaps.
        • Axelrod's paper The Dissemination of Cultures. The model is described on pages 208, 209.
        • Construct a simple model for this experiment. Each "patch" can be a culture having a list of traits. The cultures don't move (as in Schelling neighborhood), but they blend traits of neighboring cultures.
        • If you're interested, here are more links about this idea
      7. Netlogo example agent based programs. Also we have Netlogo (maybe not the most current version). Type "netlogo"