Research On A Dynamic Trust Evaluation Model Of Internetware

With the rapid development of the Internet and computer technology, Internetware emerged as a new software paradigm, which is built upon the coordination of the heterogeneous software services entities in the Internet environment. There is no centralized management authority can rely on the Internet environment, trust is an important thing for security and reliability of software services entities and systems. Therefore, how to conduct dynamic trust evaluation on Internetware, which has the characteristics of subjectivation and dynamic evolution, is a problem which we have to face and resolved.For the existing trust evaluation models cannot reflect the uncertainty, fuzzy of the trust relationships, thereby affecting the trust evaluation accuracy and dynamic adaptability. This paper presents a dynamic trust evaluation model for Internetware, which can support the automatic formation and update of the trust relationships among entities.In the research of Internetware dynamic trust evaluation model, we have done some researches as follows:(1) The existing trust evaluation model and Internetware have been studied, analyzed their advantages and disadvantages, then a dynamic trust evaluation model for Internetware is proposed, which is consistent with human cognitive habits.(2) To reflect subjective and fuzzy nature of trust, a fuzzy comprehensive evaluation method based on fuzzy theory for Internetware is proposed, in which determining the various factors weights we introduce information entropy theory. The method can overcome the shortcomings in determining the weight by the subjective judgments method, and improve dynamic adaptability of trust evaluation model, which is not strong enough as a result of subjective weight distribution by traditional methods.(3) In the feedback trust aggregation computation, the feedback trust information is searched by using Direct Trust Tree. At the same time, two new parameters, feedback level threshold and trust threshold, are introduced to adjust the scale of aggregating computation automatically.(4)In the direct trust and feedback trust comprehensive calculation, two new parameters, confidence coefficient and feedback coefficient, are introduced to establish and adjust the weight of direct trust and feedback trust automatically.(5) According to the dynamic character of behavior trust, a new model of trust renewal is provided, which is based on the degree of approximation in fuzzy theory, and which do better in preventing the malicious feedback. Finally, we use a multi-agent modeling platform for the simulation, and simulation’s results show that, compared to the existing trust quantization metrics, the model in this paper is more robust on trust dynamic adaptability...