ANT-BASED METHODS FOR SEMANTIC WEB SERVICE DISCOVERY AND COMPOSITION

The behavior of biological swarms offers us many clues regarding the design of efficient optimization methods applicable in various domains of computer science. Inspired by the behavior of ants, we propose in this paper two methods for semantic Web service discovery and composition. For efficiently discovering the services that will be involved in the composition process we choose to organize the available services in clusters according to the similarity between their semantic descriptions. To semantically compose services we propose a graph model that stores all the composition solutions for a specified user request. We further apply an ant-inspired selection method on the composition graph to identify the optimal composition solution according to user constraints. We have tested and validated the two ant-inspired methods on an extended version of the SAWSDL-TC benchmark collection