Evolutionary dynamics of partition games

Abstract

We introduce a model of evolutionary game dynamics for partition games - general class of resource allocation games. Our focus is on a computer modelling of the repeated partition games where the pure payoff matrix strategies are the actors of the population. At each instant, actor from one concurrent population is randomly matched with actor from other population, and they play a strategic form game Lotto. The payoffs of the game are population's fitness level. Populations with the strategies that have higher payoffs expand and those that have lower payoffs shrink. We discuss several experimental results with unexpected populations behavior.