Mutual Valuations between Agents and Their Coalitions

In Multi-Agent Systems (MAS), agents are acting individually and/or in cooperation with other agents in order to fulfill goals set by the initiators of them and thereby maximize some expected utilities. We argue that all form of cooperation is based on an expected raise of some present or future utility and that it should be modelled from a value-perspective. That is, there is no reason for an agent to cooperate if there is no (present or future) gain in doing so. As well as a coalition has a certain value for an agent, the agent is worth something for the coalition as well. Also there is typically a cost for being member of a coalition (e.g. acting according to the norms of the group). Comparing these two measures provide a guideline for future actions/norms aimed at reducing the differences. We conclude that in order to achieve a higher degree of cooperation in coalitions, agents will expose themselves to exploiters but that exploiters themselves will probably not be able to create and support coalitions.