Automatic Web Service Composition Based on Behavior Network
As Web services become more widely disseminated, tools are needed to help users find and integrate these services. As these services become more available, users and service providers will want to obtain new and value-added services by combining and reconfiguring existing services. We propose novel Web service composition based on a behavior network that selects services automatically through internal and external links with environmental information from sensor and goals. Optimal services are selected at each step, resulting in a globally optimal service sequence for achieving a preset goal. We developed our prototype using Amazon Web services to verify its feasibility. Experimental results comparing rule-based system and user tests demonstrated the potential of the behavior network for Web services applications.