Abstract: It is difficult to deal with the collusion in malicious nodes and spy actions in large scale P2P networks. Aiming at these problems, a model for trust management based on fuzzy theory in P2P networks is proposed, in which the reliability of recommendation trust can be computed by aggregating similarity and Bayes estimate. With the help of chord protocol, the overheads of communication can be reduced well. Simulation results show that this model can distinguish malicious nodes effectively and improve the rate of successful transaction.

Key words: trust, fuzzy theory, similarity, Bayes estimate

摘要： 针对大规模P2P系统不能有效处理恶意同伙合谋和间谍行为的问题，提出一种基于模糊理论的P2P信任模型，采用相似度和贝叶斯估计加权的方式计算节点推荐可信度，结合Chord协议来放置推荐值，减少了系统通信开销。仿真实验结果表明，该模型能有效识别恶意节点，提高成功交易率。

关键词: 信任, 模糊理论, 相似度, 贝叶斯估计

CLC Number: 

• TP309.2