PeerStrategy: A Local Strategy for Peers to Evaluate Their Neighbors

Trust management is receiving more and more attention recently, which is also critical to the wide acceptance of P2P computing. Some applications, especially e-commerce, need a mechanism to evaluate the trustworthiness of participating peers and combat the dishonest, selfish, and malicious behavior. Reputation-based trust mechanism has been identified in the literature as a viable solution to the problem. Most reputation-based systems rely on personal feedbacks to generate global trust. Therefore, it´s important for feedbacks to reflect peers´ character, good or malicious. However, the existence of strategic peers who cheat out of several interactions and human judgment error is a great challenge. Current personal feedback calculation methods can´t defend strategic peers as well as neglecting human judgment error. In this paper, we propose a local strategy, named PeerStrategy, to calculate personal feedbacks about neighbors, which can combat strategic peers as well as tolerate the influence of human judgment error. We compare current feedback calculation methods with PeerStrategy in the same experiment settings and find PeerStrategy performs the best. Our simulation shows that PeerStrategy can significantly diminish estimation error in global trust estimation by way of improving the accuracy of feedbacks.