基于声誉和遗传算法的P2P网络推荐节点选取算法  

A Selection Method of Recommenders Based on Reputation and Genetic Algorithm in P2P Network

作　　者：黄海[1] 许峰[1] 

机构地区：[1]南京航空航天大学计算机科学与技术学院,江苏南京210016

出　　处：《中国电子商情（通信市场）》2011年第4期110-116,共7页

基　　金：航空科学基金资助项目（20101952021）;航空电予系统综合技术国防科技重点实验室和航空科学基金联合资助项目（20085552021）

摘　　要：P2P网络的动态性、开放性、节点匿名性等特性给节点信任评价机制的效率提出了很高的要求，为此本文提出了基于声誉和遗传算法的P2P网络推荐节点选取算法。首先引入推荐信任度量上下文及上下文相似度以量化不同节点的推荐信任传递关系，并结合遗传算法提出的可信路径发现方法以得到可传递的推荐信任关系网络。通过节点间的信任传递与迭代计算，确定具有高声誉值的推荐节点。初步实验结果表明，该方法在保证推荐信息准确性的前提下，可以尽可能缩小推荐节点的发现范围，减少网络资源消耗，从而提高基于协同推荐的声誉评价机制的效率。Dynamic, open, node anonymity and other characteristics set higher requirements to the efficiency of trust evaluation mechanism in P2P network, therefore, this paper proposes a selection method of recommenders based on reputation and genetic algorithm in P2P network. Firstly, recommendations measuring context and context similarity are introduced to quantize recommendation transmission relationships of peers. Then, we can get the recommendation transmission relationship network by combining the credible path found method proposed by using Genetic Algorithm. The recommenders with high reputation are determined through transmission and iterative calculation of peers＇ trust value. It is proved by results of primary experiments that new method guarantees the accuracy of recommendation information, narrows the found scope of recommenders, reduces the consumption of network resources and improves the efficiency of trust evaluation mechanism.

关 键 词：声誉 遗传算法 信任 推荐 上下文 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]