A Trust Evaluation Mechanism Based on Certified Trust and Fuzzy-Derived Reputation

Introducing the trust evaluation into the access control for the service requester in a distributed network environment could be helpful for enhancing the precision and security of authorization decisions, special to the strange requesters. The existing researches of the trust evaluation could be divided into the objective means and the subjective means. The subjective evaluations could be used to dynamically measure the user's trust degree according to the reputation based on its past actions. But it cannot work when the interaction experience with the user is absence. Compared with other subjective evaluation models using the probability theory, the fuzzy based model could exactly depict the unsure and fuzzy phenomena in process. But the existing fuzzy method based on the fuzzy synthetic evaluation needs to do complex calculations between trust vectors. And the chosen operators are not enough accurate. In this paper we promote a comprehensive trust evaluation framework was a fusion of objective and subjective means. An initial trust view for user is provided based on the certified trust and a reputation is derived by the fuzzy synthetic evaluation method based on user's past actions. The proposed fuzzy method can avoid complex calculations between trust vectors, and adopts a more accurate operator to avert the information wastage.