Semantic Web service composition using Graphplan

Web service composition has become a key area of research in the services oriented architecture (SOA) community. The idea of using AI planner to compose Web services has been suggested in a lot of papers. Since the different application domains and the constraint of precondition that AI planners have, none of the existing planner can solve this problem, perfectly. In this paper, we present an algorithm that translates Web service composition problem to AI planning problems, and using Planning Domain Definition Language (PDDL) to describe the problem which is supported by most of AI planners. The present paper also describes the architecture and several key procedures of the translating algorithm.