Norm Diversity and Emergence in Tag-Based Cooperation

In multi-agent systems norms are an important influence that can engender cooperation by constraining actions and binding groups together. A key question is how to establish suitable norms in a decentralised population of self-interested agents, especially where individual agents might not adhere to the rules of the system. It is desirable, in certain situations, to establish multiple co-existing norms within a population to ensure a diversity of norms, for example to give agents alternatives should one norm collapse. In this paper we investigate the problem of norm emergence, and the related issue of group recognition, using tag-based cooperation as the interaction model. We explore characteristics that affect the longevity and adoption of norms in tag-based cooperation, and provide an empirical evaluation of existing techniques for supporting cooperation in the presence of cheaters.