THE EQUIVALENCE OF EVOLUTIONARY GAMES AND DISTRIBUTED MONTE CARLO LEARNING

This paper presents a tight relationship between evolutionary game theory and distributed intelligence models. After reviewing some existing theories of replicator dynamics and distributed Monte Carlo learning, we make formulations and proofs of the equivalence between these two models. The relationship will be revealed not only from a theoretical viewpoint, but also by experimental simulations of the models by taking a simple symmetric zero-sum game as an example. As a consequence, it will be verified that seemingly chaotic macro dynamics generated by distributed micro-decisions can be explained with theoretical models.