摘要
信任的动态性刻画和信任信息的聚合能力是信任模型需要考虑的重要问题,然而现有信任模型存在难以描述节点行为的动态变化特性、忽略信任的模糊特征和上下文环境影响等问题。为此,提出一种基于直觉模糊集合理论的动态信任模型DIFTrust,通过兴趣域划分机制刻画节点信任关系存在的上下文环境,基于直觉模糊集合理论刻画信任的模糊特性,以及利用时间窗口机制和自适应权重向量刻画信任的动态衰减特性,解决具有直觉模糊特征的节点动态信任关系的描述问题。实验结果表明,该模型能够有效描述节点的信任特征,在防止恶意节点策略性行为改变等方面优于现有PeerTrust和DyTrust等模型。
The dynamic property description and information aggregation ability is a key function of trust models,but most current models cannot describe the dynamic variation features of node behaviors,and ignore the fuzzy property as well as behavior context relationship of trust.In this paper,an intuitionistic fuzzy theory based dynamic trust model DIFTrust is proposed.In DIFTrust,the behavior context correlated,fuzzy,and dynamic properties of trust are characterized by interest domain partition mechanism,intuitionistic fuzzy theory,as well as time window and adaptive weight vector mechanism.DIFTrust provides a way to describe the node trust relationship with dynamic and intuitionistic fuzzy features.Experimental results show that DIFTrust can describe trust features more effectively,and has good performance in defending strategic behavior changing of malicious peers compared with current models such as PeerTrust and DyTrust.
出处
《计算机工程》
CAS
CSCD
2014年第8期133-137,共5页
Computer Engineering
基金
国家自然科学基金资助项目"高质量自适应P2P流媒体直播服务模型及关键技术研究"(61103225)
关键词
对等网络
信任模型
兴趣域
时间窗口
直觉模糊
推荐相似度
P2P network
trust model
interest domain
time window
intuitionistic fuzzy
recommendation similarity
作者简介
王群（1981-）,男,讲师、硕士,主研方向：对等网络,分布式计算 E-mail：qun．w@263．net
戴秀岳,讲师、硕士
杨莉,工程师、博士.
