一种改进的社会网络动态信任模型  

An Improved Dynamic Trust Model in Social Network

作　　者：刘雪亮[1] 胡晓辉[1] 

机构地区：[1]兰州交通大学电子与信息工程学院,甘肃兰州730070

出　　处：《计算机技术与发展》2016年第6期51-55,共5页Computer Technology and Development

基　　金：甘肃省高等学校特色专业(080901);甘肃省科技支撑计划(2014GS03891)

摘　　要：社会网络中,针对信任计算过程中出现的恶意节点通过提供低质量服务和节点的动态摇摆的攻击从而达到某种利益而导致网络中恶意服务的增多,导致节点之间的信任不可信或者不可靠,给节点的交互带来一定的不安全性等问题进行了研究。通过分析节点的信任评估行为,节点共谋产生不信任的推荐,提出了一种改进的信任形成和计算方法。通过引入节点的主观评价和客观评价,结合GTFT和TFT博弈激励策略对推荐过程中不诚实的信任推荐给予相应的惩罚策略。仿真结果表明,根据节点主观评价与客观评价所形成的直接信任和通过博弈计算的推荐信任在一定程度上能够促进节点理性地参与网络行为,改善网络中信任计算的不可信和不可靠问题,使得网络中攻击节点逐渐减少。In social network,the problem of large number of malicious services and the trust value unlikelihood or unreliable with the malicious nodes and the dynamic nodes which provide low quality service and network attack in trust calculating is researched. By analyzing the trust evaluation behavior of the node and the recommendation from the conspiracy node, an improved method to calculate and form trust value is proposed. Through introducing the subjective and objective evaluation of the node,combined with the GTFF and TFT,a re- lated punishment strategy is given for dishonest trust recommendation in recommending process. The simulation shows that the direct trust and recommendation trust based on the subjective evaluation and objective evaluation can promote the rational participation of the nodes, improving the unlikelihood or unreliable problem in trust calculation, and decrease the attack node in the network.

关 键 词：博弈论 反馈值 激励策略 信任推荐 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]