Constraining Self-Interested Agents to Guarantee Pareto Optimality in Multiagent Coalition Formation Problem

Abstract

In this article, we tackle the problem of coalition formation in hedonic games. In such context, agents have their own preferences over coalitions which they may join and could be endowed with different behavioral strategies in order to search for appropriate partners and better assess the proposals of the other agents. However when agents are self interested and free to decide which strategies they use, the system cannot expect them to form a coalition structure with specific properties. Our paper addresses such issue and studies the constraints that could be enforced on self-interested agents, in order to form coalitions which guarantee significant solution concepts. In this article, we focus on the Pareto optimality of the formed coalition structure. We develop our coalition formation model and propose a theoretical analysis of this model.