RepTrust : P2P环境下基于声誉的信任模型

声誉是一种新型的信任建立机制，在P2P网络中，利用基于声誉的信任机制可以孤立劣质节点，选择更为可信的节点进行交互，以规避和减少风险。针对目前研究中存在的与网络结构藕合过紧、防止联合欺作和声誉时效性处理不佳等问题，提出了一种新的基于声誉的信任模型RepTrust。模型采用与获取路径甚至评价者rater无关的声誉信息计算信任度，利用基于大多数原则的算法过滤不实声誉，利用衰减因子处理声誉的动态性。仿真实验表明模型在信任度计算、对付欺作和动态声誉处理等方面效果明显。

RepTrust:Reputation Based Trust Model in P2P Environment

Reputation is a new method to build trust. In Pecr-to-Peer networks, trust mechanism based on reputation can isolate inferior peers and recommend excellent peers in order to evade potential risk and provide better service. Some shortcomings were found in related research such as overly dependency on network structure, neglect of collusion cheat and disregard of dynamic reputation change. Aiming at these shortcomings, RepTrust, a novel trust model based on reputation was presented. In this model, reputation unrelated with raters and routes through that it been gotten was used to compute trust. Mendacious reputation was filtrated and an attenuation factor was introduced into to amend the dynamic reputation. According to the simulation experiment, this model is effective in trust computing, dealing with collusion cheat and dynamic reputation.