A rough guide to CF 902

Document Sample
A rough guide to CF 902 Powered By Docstoc
					A rough guide to CF 902




    CF 902 Part I: Computational Models of
    Agents Networks and Self-Organisation
                          Dr Dietmar Maringer

                           Autumn Term 2005
A rough guide to CF 902

         Why use agent based models?
              The story so far
              The “representative” market participant
              Reaching the limits
              Overcoming the limits
CF 902
A rough guide to CF 902

         Agents at Work: The Game of Life
              John Convey (Scientific American, Oct 1970:120-123)
              Starting Point:
               » Checkerboard
               » Some (randomly selected) cells are living

              The Rules for cells:
               1. Living cell with 2 or 3 neighbours          survives
               2. Living cell with <2 or >3 neighbours        cell dies
               3. Empty cell with 3 neighbours                cell is born
CF 902
A rough guide to CF 902

         Convey’s Game of Life
              Some Examples
CF 902
A rough guide to CF 902

         Matlab version of Convey’s Game of Life
              X = sparse(m,m);
                     population matrix: 1 = alive, 0 = empty
              survive = [ 2        3 ]; list of “survival” condition
              birth = [ 3 ];              list of “birth” condition

                   Pseudo-Code:
                   Initialize environment and X;
                   WHILE halting condition not met;
                     Determine current number of neighbours;
CF 902




                     Check if cells survive / become alive;
                     Update X;
                     graphical output;
                   END;
A rough guide to CF 902

         What CF 902 (Part I) is concerned with
              Reinforcement learning
              evolutionary search
              heuristics
CF 902

				
DOCUMENT INFO