Diversification by Self-reinforcement of Preferences and Change of Interaction Type

Abstract

There are situations where agents can benefit from coordinating their actions, i.e., agents gain payoffs by taking the same action as the others. Such situations can be modeled as coordination games. Other situations, i.e., agents gain payoffs by taking different actions, are modeled as complementary, asymmetric coordination, or minority games. In this study, we treat coordination and complementary games as types of interactions. Each agent can change the interaction type and faces binary choices. We examine the effects of decision errors on collective behavior and we show that a small decision error stabilizes collective behavior in a pure population of all agents who play either the coordination or complementary game. We also examine populations in which each agent can change the interaction type. We show that in a mixed population of agents who play either game, the behaviors and games become heterogeneous and the ratio of cooperative agents becomes 80% and the average utility is stable high.