A Multiagent Approach for Configuring and Explaining Workflow of Semantic Web Services

In this article, we describe a multiagent application that configures Semantic Web services following a design problem solving method. For that, a propose-and-revise strategy was modeled and implemented in OWL-S. The other contribution of this method is the extension of OWL-S for recording the service execution flow followed by the agent. The trace of the agent’s reasoning is stored in node sets represented in a proof markup language, which make up a framework for explanations on the Web. The proposed approach is shown to be useful in the context of e-business where the workflow definition of services is automatically performed by the agents in order to configure a computer from a customer order.