移动自组网中基于多约束和协同过滤动态信任机制  

Dynamic trust mechanism based on multi-constraints and collaborative filtering in mobile Ad hoc networks

作　　者：彭三城[1] 陈建平[1] 胡忠望[1] 

机构地区：[1]肇庆学院计算机学院,广东肇庆526061

出　　处：《计算机应用研究》2011年第3期1065-1069,共5页Application Research of Computers

基　　金：国家自然科学基金资助项目(60773013);肇庆学院科研启动基金资助项目(201032)

摘　　要：由于具有分布式特性,导致移动自组网容易遭受攻击,为了增强移动自组网的安全性,提出了一种移动自组网中基于多约束和协同过滤的动态信任机制。在根据节点自身经历的基础上,采用一个带多约束的信任更新算法来评估直接信任,其中:时间衰减因子保证信任度随时间进行衰减;奖励因子保证良好的节点应受到奖励;惩罚因子保证恶意节点应受到惩罚。采用协同过滤技术来评估推荐信任,以此来阻止不诚实的推荐。通过定量评估分析和模拟仿真,结果表明所提出的方法比Bayesian模型能更精确地计算节点之间的信任度。该方法能有效地提高移动自组网的安全性。Due to their distributed nature,MANETs are vulnerable to various attacks.To enhance security of MANETs,this paper proposed a DTM based on multi-constraints and collaborative filtering in MANETs.The main idea of this method introduced a trust updating algorithm with multi-constraints（time aging factor ensured that the trust fades with time;rewards factor ensured that good peers deserve to be rewarded;and penalty factor ensured that the selfish or the malicious deserve to be punished） to assess direct trust by self-experience of nodes,and adopted collaborative filtering technique to evaluate recommendation trust in order to prevent false recommendations.The analytic results show that DTM not only can compute trust between nodes more precisely than Bayesian model,but also can improve security of MANETs through quantitative evaluation analysis and extensive simulations.

关 键 词：移动自组网 信任评估 多约束 协同过滤 动态信任机制 直接信任 推荐信任 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]