Applying AI Planning to Semantic Web Services for Workflow Generation

Semantic Web Services (SWS) intend to enhance interoperability and encapsulate domain knowledge via semantic descriptions. This paper claims that service composition in a service-oriented computing paradigm is an AI planning problem. We present the Event Calculus (EC) - a well-known action theory, and further discuss the mechanisms of using the formalism for representing and reasoning service capabilities. We describe an architecture for automatic workflow generation in which an EC-based planner builds workflows by reasoning against distributed pools of SWS.