基于多智能体的多维证据动态信任计算模型  

A multi-dimensional evidence dynamic trust computing model based on multi-agent

作　　者：蒋伟进[1,2] 许宇晖[3] 郭宏 许宇胜 

机构地区：[1]湖南商学院计算机与信息工程学院,湖南长沙410205 [2]武汉理工大学计算机科学与技术学院,湖北武汉430070 [3]湖南商学院设计艺术学院,湖南长沙410205 [4]北方导航控制技术股份有限公司,北京100176

出　　处：《山东大学学报（理学版）》2015年第1期1-11,19,共12页Journal of Shandong University(Natural Science)

基　　金：国家自然科学基金资助项目(61472136;21106036);教育部人文社科研究一般规划基金资助项目(11YJAZH039);湖南省社会科学基金资助项目(07YBB239);湖南省重点学科建设项目

摘　　要：信任度评估是基于行为的信任管理模型关注的核心内容。为使信任度评估所依据的证据源完备,运用信任管理思想方法,提出一种基于多维证据的信任度评估模型,该模型将主体网络操作行为层面的信息引入传统仅考虑交易反馈信息的信任度评估模型,基于交易反馈和网络操作行为两个层面的多维证据源进行信任计算,扩展了证据源,突破了只依据单一种类证据源进行信任评估而引起的缺陷;另外,应用改进的D-S证据理论来合成多维证据,较好地解决了证据不确定性的问题。The trust assessment is the core content of the trust management model based on the behavior. In order to complete the evidence sources of the trust assessment,a trust assessment model based on multi-dimensional evidence was proposed by using the trust management thinking. And this model introduces information on operating behaviors level of the main network into the traditional trust assessment model which only considerates feedback information of transaction,and computes trust based on multi-dimensional evidence sources on feedback information of transaction and network operating behaviors level,and extends the evidence source,and overcomes the defect when trust is assessed only according to a single type of evidence source. In addition,the application of the improved DS evidence to synthesize multi-dimensional evidences can solve the problem of uncertainty of evidences.

关 键 词：信任管理 信任评估 多维证据 动态信任计算 

分 类 号：TP274.2[自动化与计算机技术—检测技术与自动化装置]