Evolving Automata Negotiate with a Variety of Opponents

The rapid growth of a global electronic market place, together with the establishment of standard negotiation protocols, currently leads to the development of multi-agent architectures in which artificial agents can negotiate on behalf of their users. Ideally, these agents should be able to negotiate successfully against a variety of opponents with different tactics and different preferences. Furthermore, they should be able to adapt their strategies to deal for instance with agents with different preferences. We show that such flexible and powerful bargaining agents can be obtained using the combination of finite automata and evolutionary algorithms (EAs). Finite automata allow the bargaining agents to behave differently against different opponents. EAs can be used to adapt the agents' bargaining strategies (consisting of finite automata) in successive steps to generate more and more successful strategies in the course of time. The performance of the evolving automata is assessed in a competition against a broad variety of bargaining strategies. Highly-efficient bargaining strategies, which discriminate successfully between opponents with different bargaining tactics, are generated by the EA. We also investigate the situation in which the opponents are also co-evolving (and have different preferences). Positive results are obtained in this setup as well. The evolving automata perform especially well when the bargaining game is very short and a fast discrimination between different opponents becomes necessary.