WSC: An Automated Web Service Composition Framework Based On Semantic Similarity

As Web technology continues to evolve, more and more Webservices have been deployed in the network. Meanwhile as users’ needkeep changing and become gradually complex, a single service cannotmeet the actual needs of the user. How to manage so many Web serviceseffectively, how to find suitable services rapidity, and how to combinemultiple services automatically, are attracting more and more researchers’attention.To solve these problems, this paper introduces the concept of servicetypes, which grouped services with similar functions into the sameservice type. The introduction of the service type is to help manage alarge number of Web services effectively by adding a function identifierto each service. Then, based on Artificial Intelligence Planning, composeexisting service automatically to meet user’s needs.In order to realize the goal of service’s automatic classificationmanagement and automatic combination, this paper our research works asbellow:1. This paper analyzed the existing Web Service description methodsand the mainstream of semantic similarity calculation methodssystematically and comprehensively, introduced related concepts ofservice type, and then we introduced the AI planning method SHOP2indetail.2. Based on the Web service ontology language OWL-S, this paperincrease the service type information description, and puts forward thesemantic similarity algorithm for service, and realized the automaticclassification management of service.3. In order to select a service with high quality from so manyservices those have different qualities, service’s quality information wasneeded, so we introduced the semantic labels of QoS, and proposed the QoS calculation method based on user’s satisfaction, the calculation ofQoS is more reasonable.4. Based on the service semantic model, service classification andAI planning, we put forward a dynamic Web service compositionframework WSC which have the function of classification managementand dynamic Web service combination. We abstract out the core functionmodule of service classification, plan combination, matching from WSC,and studied the key problems in WSC.5. Finallly, we design and realize the prototype system of WSC, andverified WSC with a simulated service request-buying books, andconfirmed the feasibility of the system of dynamic service compositionframework WSC.