Enhance trust management in composite services with indirect ratings

Service-oriented computing has been developing rapidly with various architectures and applications for improving consumer satisfaction. However, the increasing complexity of service models has brought many challenges to the trust evaluation accuracy, which is a decisive factor in making interactions between consumers and providers in dynamic environments. In composite services, one of the culprits for inaccurate trust evaluations is the problem of missing evidence caused by the rating convention. To address this, this paper proposes a strong evidence-based approach to manage ratings for members of composite services under uncertainty. Moreover, the approach complements the existing one-to-one towards one-to-many trust evaluation. The experimental results show that our method can enhance the consistency and accuracy of reputation systems, robust against biased ratings and also suitable for distributed environments.