Correlated Contribution Analysis for Service Composition in Dynamic Environments

Abstract

Service Oriented Computing Systems can be considered as a type of complex systems consisting of a number of loosely coupled autonomous and adaptive components (i.e., service components). Service quality depends on the performance of the "service group", and many dynamic factors including the expectation of service consumers, the availability of resources, etc. Trust is an important factor for determining the interrelationships among service components. In this paper, we consider the dynamic factors in service composition, and propose a trust management approach, which adopts related methods in information theory, to enable more reliable service composition in dynamic environments. From the experimental results, we claim that the proposed approach can effectively handle dynamic factors in open environments, and obtain better service composition results.