A Fuzzy Trust Model with Punishment Mechanism

One way to minimize threats in P2P system is to exploit reputation to help evaluate the trustworthiness and predict the future behaviors of peers. In this paper, a set of parameters are identified for describing the trust of a peer. And a trust model based on fuzzy techniques is presented. The trust model uses Amplifier of Malicious Behavior (AMB) to intensify malicious behaviors and helps peers identify the sly malicious peers. The recommendation credibility is evaluated by Recommendation Weight (RW) and Recommendation Accuracy (RA). In simulation, this trust model has been shown to improve the efficiency of P2P system by significantly decreasing the number of inauthentic files on the network.