Agent-Based Modeling and the Evolution of Cooperation: Methods and Assumptions
University of California, Davis
Last modified: September 14, 2007
Agent-Based modeling is an important tool for implementing increasingly realistic game-theoretical models of the evolution of cooperation by, for example, explicitly modeling space, movement and environmental structure. Seemingly harmless assumptions can greatly affect the solutions to cooperative games. For example, if agents move about in a 2-dimensional discrete space, should more than one agent be allowed to occupy the same cell? For some strategies in prisonerís dilemma games, whether or not agents can occupy the same cells dramatically alters solutions in such games. In this paper, I will discuss several examples including spatial, reproduction, and payoff accumulation assumptions. I conclude that more work is needed to determine what aspects of game-theorectical models of cooperation are relatively sensitive to assumptions and which are not.