CBTM A Trust Model with Uncertainty Quantification推荐.pdfVIP

CBTM: A Trust Model with Uncertainty Quantification and Reasoning for Pervasive Computing Rui He, Jianwei Niu, and Guangwei Zhang Beijing University of Aeronautics and Astronautics, Xueyuan Road 37, Beijing, China, 100083 {harry.he, niujianwei}@, ezhang@263.net Abstract. This paper presents a novel trust model in which we model trust based on an exotic uncertainty theory, namely cloud model. We regard trust between en- tities as a cloud that is called as trust cloud. Based on such a quantification model of trust, we further propose the algorithms to compute propagated trust relation- ships and aggregated trust relationships, which are needed for trust reasoning in pervasive computing environments. Finally, we compare the proposed trust model with other three typical models in simulation experiments, and the results shows the cloud-based trust model performs better in a total sense. 1 Introduction Trust has been researched for more than teen years since Marsh’s work [1]. Because trust mechanism is more flexible and extensible than traditional security approaches such as PKI[20], trust has been introduced into many other cyber fields, i.e. pervasive computing, peer-to-peer networks, etc. In such contexts, trust is always regarded as subjective, therefore, how to measure trust become very important. Till now many approaches have been proposed to quantify trust [1,2,3], which either use discrete numbers such as -1, 0, 1, etc. to indicate different trust levels, or use a real number interval, for instance [0, 1]. However, since trust is subjective, it is not enough to describe trust with deterministic values. As we know, in human society, when we say we trust a person very much, actually