Research On P2P Trust Evaluation Model Based On Fuzzy Theory

In recent years, P2P technology is a research focus. Because P2P network isapplied widely, the open and anonymity nature brings various security risks. P2Pnetwork environment is complicated, factors of trust evaluation are various and thetrust itself is of great subjectivity, uncertainty and ambiguity. Aiming at thephenomenon, the thesis researches trust evaluation model based on subjective theoryroundly, and pay the most attention on the trust evaluation model based on fuzzytheory.Firstly, by improving the traditional P2P trust evaluation model based on fuzzytheory, we propose a new P2P trust evaluation model. In the model, we applytwo-stage fuzzy comprehensive evaluation of fuzzy theory to evaluate therecommendation trust vector of P2P peers, which fully reflects the effect of variousfactors on trust evaluation of the objective peer, at the same time, dynamically reflectsthe demand of different peers in the different transaction context, so the trustevaluation result is more truthful and reliable.Then, aiming at P2P trust model based on fuzzy comprehensive evaluation, thethesis establishes membership function of trusted set according to the trusted level inthe P2P trust model. The results of simulation and analysis show the membershipfunction of trusted set can validate whether the trust evaluation sufficiently close totrusted level.Finally, the thesis applies the thought of maximal tree to propose a peerclassification method. It’s based on fuzzy clustering. The method classifies effectivelythe recommendation peers according to their trusted level. If choosing therecommendation peers in the higher trusted level, their recommendation results willbe not only more reliable, but also can reduce the computation workload ofrecommendation trust.