Developing a composite trust model for multi-agent systems

The Semantic Web, conceived as a collection of agents, brings new opportunities and challenges to trust research. Enabling trust to ensure more effective and efficient interaction is at the heart of the Semantic Web vision. We propose a composite trust model based on statistical decision theory and Bayesian sequential analysis to balance the costs and benefits during the process of trust evaluating and combine a variety of sources of information to assist users with making the correct decisions in selecting the appropriate service providers according to their preferences. The model proposed by this paper gives trust a strict mathematical interpretation in terms of probability theory.