Exploring indirect reciprocity in complex networks using coalitions and rewiring

Abstract

It is generally known that cooperation can be achieved in complex real-world interactions that are not limited to direct interactions only. In particular, cooperation can consider prior interactions with other players, i.e., indirect reciprocity. Moreover, coalition based mechanisms have shown to facilitate cooperation among self-interested agents. Also, research on games over dynamic topologies has found empirical evidence showing that partner switching leads to cooperative behavior. In this paper we present a new mechanism to improve cooperation among self-interested agents placed in a complex network. Our mechanism is based on three main pillars: indirect reciprocity, coalitions and rewiring. Thus agents play against each other an indirect reciprocity game where they can create coalitions to share information about agents' reputation or change their personal network (social contacts). Altogether, we explore the conditions to enhance cooperation in complex networks. Finally, in our experiments we determine how, by using our mechanism, cooperation is improved in our reputation-based game, and how topology highly influences cooperation in our scenario.