Evolutionary Models of Bargaining: Comparing Agent-based Computational and Analytical Approaches to Understanding Convention Evolution

This paper compares two methodologies that have been used tounderstand the evolution of bargaining conventions. The first is theanalytical approach that employs a standard learning dynamic andcomputes equilibria numerically. The second approach simulatesan environment with a finite population of interacting agents. We compare these two approaches within the context of threevariations on a common model. In one variation agents randomlyexperiment with different demands. A second variation positsassortative interactions, and the third allows forsophistication in agent strategies. The simulation resultssuggest that the agent-based approach performs well in selectingequilibria in most instances, but exact predicted populationdistributions often vary from those calculated numerically. Classification Numbers: C63, C73, C78. © 2002 Kluwer Academic Publishers.