A Structure-based Approach for Dynamic Services Composition

Abstract

Web service composition needs to increase the dynamic feature of adapting complex and unstable business environment. Thus the service selection and evaluation are very important for service composition. Although previous studies have stated some approaches to support dynamic composition, they are unable to balance the flexibility and verification very well in dynamic environment. Moreover, most service selection methods only depend on the similarity of a pair of single services. These methods are useful and concise to find substitutes for unavailable services, but sometimes they might be too strict to find a solution. To overcome these problems, we propose a novel approach named SPACE architecture with six basic structures, and define the service composition based on situation calculus language. SPACE estimates the similarity by the basic structures and constraints rather than the features of single service. It can help us to find non-optimal but acceptable substitutes and guarantee the verifications of composition. Finally, as a case study we consider a health care scenario to demonstrate our approach.