Research Of A User Preference-based Trusted Model Of Service Composition

With the growing of web service type and number, service composition technology is becoming more mature. In an open and dynamic internet environment, research on the trust of service composition is attracting more and more attention. And the hot spots now in this field certainly include how to measure the trust of component services and how to select trusted composition plans.With focus on those two hotspot issues, we propose a user preference-based trusted model of service composition in this paper. Within this model, the trust of a single web service (namely, the component services) is measured from user preference perspective, and base on that, we use user preference as a driver to realize trustworthy and optimal selection of composition plan.To achieve the research objective mentioned above, the main work of this paper includes the following aspects:(1) This paper systematically and comprehensively analyzes some current relevant research and technique in the field of trusted service composition. Through these analyses, this paper elaborates the current status and challenges and future directions of this field.(2) For addressing the problem of measuring the trust of a single web service, this paper builds a trust model based on user preference with Bayesian Network, which calculates the services'trust in the form of statistical probability. In this trust model, not only different trust sources are taken into account, but the users'impact is also introduced into the metric system as weight factors.(3) In order to solve the problem about trusted and optimal selection of composition plan, a method for calculating the QoS of composite services is proposed in this paper. We use this method to associate the QoS of composite service with the QoS(s) of the component services. Then, with the trust of component services and driven by user preference, we select the optimal composite plan and its trust is verified.(4) Based on the above work, a middleware prototype system is designed and implemented to prove the feasibility and effectiveness of the model presented in this paper.