Conflicts resolution in heterogenous multiagent environments inspired by social sciences

Abstract

Conflict, which is an inherent part of multiagent environments, is also a natural element of any social structure. Therefore, it might be useful to research social sciences in order to find out what is the state of the art concerning conflicts and then try to investigate how it can be utilized in multiagent settings. A synergy of the two fields may also lead to significant insights about conflicts for social sciences. In our approach we focus on Thomas and Kilmann’s classifications of conflict resolution strategies. Following them, we design the behaviour semantics of five different styles of dealing with conflict and obtain a new method of conflict resolution in heterogenous multiagent environments where agents differ among themselves either physically (e.g. do not have the same sensors) or concerning their roles (e.g. their goals are different). Then we conduct series of simulations in order to understand the nature of modeled strategies. Investigation lets us answer the questions about existence of a dominant strategy, influence of proportions of agents of various types and influence of number of conflicts in a population on the performance of distinct strategies. Finally, we are able to find the best circumstances for each strategy in which it can be adopted.