Research On Dynamic Trust Model Based On Multiple Behavior Trust And Decision Factors In Trusted Network

In the trusted networks, behavior of nodes and their result can be expected and controlled.Dynamic trust model which evaluates nodes’ behavior in the network is the core content of trustednetwork research. Because of the complexity and diversity of malicious behavior, the research ondynamic trust model is to solve how to evaluate nodes’ behavior, protect networks from maliciousbehavior, and achieve itself in reality.This paper proposes a distributed dynamic trust quantitative model based on multiple behaviortrust and decision factors (referred to as MTFMT). This model evaluates different behavior as tradebehavior trust and feedback behavior trust. It also takes multiple decision factors into account,including direct trust, feedback trust and incentive function. The weight of each decision factordepends on its information and effect. MTFMT evaluates behavior trust more accurately and canreduce the impact of malicious behavior effectively in networks.This paper proposes the algorithm to calculate direct trust based on trust vector with penaltyfactor and fading time. The algorithm to calculate feedback trust based on feedback trust tree ispresented, too. It also proposes an encrypted mechanism to transport feedback information andstorage structure of local data in distributed environment to improve the implementability oftraditional dynamic trust model.At last, simulation results verify the performance of this proposed model. The advantages anddisadvantages of MTFMT are also compared and analyzed.