Planning for Semantic Web Services

Using Semantic Web ontologies to describe Web Services has proven to be useful for various different tasks including service discovery and composition. AI planning techniques have been employed to automate the composition of Web Services described this way. Planners use the description of the preconditions and effects of a service to do various sorts of reasoning about how to combine services into a plan. OWL-S 1.1 will support the description of the preconditions and effects of services using OWL statements similar to atoms in Semantic Web Rule Language (SWRL). Thus, planners are required to understand the semantics of OWL in order to evaluate such preconditions. However, planners typically support only fairly limited reasoning capabilities which cannot handle the expressivity of Semantic Web ontologies. In particular, planners typically make the closed world assumption, whereas OWL has open world semantics. In this paper, we demonstrate how an OWL reasoner can be integrated with an AI planner to overcome these problems. We identify the challenges of writing the service descriptions and reasoning about them when OWL is used to describe preconditions and effects. We also investigate the efficiency of such an integrated system and show how OWL reasoning can be optimized for this system. Finally, we present the performance results of our prototype implementation.