由于P2P网络的开放、匿名等特点,节点问的信任关系往往很难通过客观的信任机制建立.本文引入模糊理论的方法对信任进行度量,通过改进Einstein算子来解决信任向量的传递和合并问题,并把算子用于求全局信任关系模糊矩阵的传递闭包,结合分布式哈希表的机制来存储全局信誉值,较好的解决了信任的传播问题.仿真实验表明该模型能有效的提高P2P网络交互成功率,同时系统开销相对较小。
Abstract:The feature of the peer-to-peer system such,as user anonymity, open nature makes that it's hard to establish the trust among participants by the objective confirmation mechanism. This paper presents a new P2P reputation system based on the fuzzy set theory. It employs fuzzy operators to deal with the issues of trust vector's transfer and coalition, and then apply them to calculate the transitive closure of global trust relation matrix. Plus, with the help of distributed hash tables, the trust value can propagate easily. Experiments show that the model can improve the rate of successful barter greatly, and the computing cost of the whole peer-to-peer network is comparatively less.