Service-Correlation Aware Service Selection for Composite Service

In service oriented application, value-added service is composed of some component services which are selected from the candidate services. However, in current service selection model, the QoS of a candidate service is independent from other candidate services. Therefore, current service selection approaches can′t deal with the situation in which the QoS of a candidate service has a correlation with other candidate services. This paper presents a systematic approach to construct the service-correlation aware QoS description to characterize the QoS dependence of an individual service on the other related services. Based on this model, it also proposes service correlation aware selection approaches for above situation, including integer programming based selection approach for optimal solution and heuristic based selection approach for sub-optimal solution. A case study is conducted and demonstrates the applicability of the proposed approach. Experiment results indicate that, based on the service correlation aware QoS description model, the QoS of composite service through integer programming based selection approach reach an obviously high level, and a sub-optimal QoS of composite service can be gained efficiently through the heuristic based selection approach.