Response Fusion in Multi-Agent Environment

Abstract

Multi agent environment are often complex, while each member of agent community has vision about only a part of the environment. In such environment, information fusion of agent community is vital to attain a better accuracy or quality of information. Agents have different ability and can access to information sources with different degree of credibility. As a result, information which is obtained by agents, should be fused with considering their response credibility. In this research, at first we have introduced a method to obtain agent credibility based upon the opinion of agent community. In the proposed method, we measure the possibility of credibility for each agent. Next, we assume that in response to a query, every agent produces a fuzzy answer set in response to a query. On this assumption, a new fuzzy operator with some desired properties for agent responses fusing with their credibility consideration is proposed.