The Research Of Trust Model Based On Fuzzy Theory

With the rapid development of Internet, Online transaction is widely used by people. But the users enrolled in the transaction are anonymous, which increases the risk of online transaction activities .The research in the trust model provides a solution to this problem, because it will help decrease the risk and unpredictability. By the analyzing of current trust model research, we found most of them ignored the dynamicity and fuzzification. So this paper is focusing on the fuzzification and dynamicity of trust, and exerts the fuzzy math theory into every part of the trust model research.This thesis firstly proposes two different methods to evaluate trust relationships. One is the direct recommendation trust evaluation based on fuzzy Petri-net, this method combines the direct recommendation about some one who is evaluated and then get a comprehensive trust value. Because the fuzzy Petri-net is a kind of graphical tool,so it will make trust reasoning to be like traditional flow chart which is simply and intuitively. The other method is the indrect recommendation trust evaluation of many roads based on fuzzy reasoning of similarity ,this method gets trust value about some on who is evaluated with the indirect recommendation of many roads. The searching of the route uses the traversal search algorithm in graph theory .In a single route, the transferring of trust value is using the method of fuzzy reasoning of similarity. This method is simple and direct. We can calculate the trust value of one route by using a regulation of fuzzy reasoning and finally combine the trust value of many roads.This thesis also proposes a new trust model of online transaction based on fuzzy theory. It treated the evaluation of trust value as a dynamic evaluation procedure. In the trust establishment stage, it uses the fuzzy Petri-net to establish initial trust relationship between entities. Regarding the weight determination in the trust value dynamic accumulation, it adopts a new weight determination based on fuzzy reasoning. After we get the trust values, it uses fuzzy judgment to judge the trust, which makes the judgment via building the regulation related with transaction amount and trust value. In this model, we use Matlab fuzzy toolbox to simulate the weight determination and fuzzy judgment.