基于记忆原理的信任值更新模型  

Trust value update model based on the memory theory

作　　者：左赛哲[1] 郭玉翠[1] 宫尚宝[1] 吴旭[1] 张虎[1] 

机构地区：[1]北京邮电大学理学院,北京100876

出　　处：《东南大学学报（自然科学版）》2010年第A02期307-312,共6页Journal of Southeast University：Natural Science Edition

基　　金：国家自然科学基金资助项目(60973146)

摘　　要：针对开放网络中节点信任值随时间和节点交易行为动态变化的特性,通过对记忆原理与社会网络记忆特性的分析,认为节点服务的重要性与事件的影响力等属性都是一种具有记忆特性的信息,网络中具备这些属性的节点的信任值的变化过程与记忆原理相似.因此以记忆原理为基础,建立了与时间及节点交易相关的变速率的信任值更新模型.该模型可以通过调整衰减因子的大小对节点交易行为起到一定激励作用.实验结果表明该模型能够准确地反应节点信任值随时间和节点交易行为变化的规律,同时还可以通过对交易频率的监控,实现对恶意节点的检测.Nodes＇ trust values change dynamically with time and transactions in an open network.By studying the memory principle and memorial characteristics of social networks,this paper proposes that the importance of nodes＇ service and events＇ influence are kinds of information with memory features,and the change of nodes＇ trust values with these attributes is similar to the memory principle.Therefore,a trust value update model,which has a changing rate and takes time and transaction behaviors into account,is established based on the memory theory.The proposed model plays a positive role in encouraging nodes to provide high-quality services by adjusting the fading factor.The experimental results show that it can describe the evolution process of the trust value more accurately.Meanwhile,the malicious nodes can be detected by monitoring the frequency of transactions.

关 键 词：记忆原理 信任信息 信任衰减 交易特征 

分 类 号：TP309.2[自动化与计算机技术—计算机系统结构]