A Bayes Net Approach to Argumentation Based Negotiation

Abstract

Negotiation is one of the most fundamental and effective mechanism for resolving conflicts between self-interested agents and producing mutually acceptable compromises. Most existing research in negotiation presumes a fixed negotiation context which cannot be changed during the process of negotiation and that the agents have complete and correct knowledge about all aspects of the issues being negotiated. In practice, the issues being negotiated may change and the agents may have incorrect beliefs of relevant issues updated during the negotiation process. Argumentation-based negotiation approaches have therefore been proposed to capture such realistic negotiation contexts. Here we present a novel Bayesian network based argumentation and decision making framework that allows agents to utilize models of the other agents. The agents will generate effective arguments to influence the other agent’s belief and produce more profit.