| Web services applications prevail in mobile devices such as the iPhone and Blackberry as well as among numerous networked computers. Web services are a state-of-the-art, leading technology based on Service Oriented Architecture in the World Wide Web environment. Each Web service is described in a standard language, such as WSDL (Web Service Definition Language), and is published in a global registry called UDDI (Universal Description, Discovery, and Integration). Individual users or software agents invoke Web services based on the Web service descriptions registered with UDDI. A simple information or service request can be fulfilled by a single Web service, but complicated requests cannot be satisfied by a single Web service. Consequently, a composition of multiple Web services in an appropriate sequence is required.;As the number of Web services increases in dynamic business environments, the automation of Web service composition becomes an essential feature for commercial Web services. Automatic Web service composition software should be able to consider not only functional requirements, but also quality of service (QoS) aspects. Functional requirements force the composition software to generate feasible solutions, while QoS aspects make the composition software satisfy user objectives, such as cost, time, or reliability.;Web service composition problems can be classified in two ways: (1) syntactically and (2) semantically. Semantic issues have recently presented challenges to Web service composition. Due to insufficient understanding of semantics, Web service composition solutions can be inferior, or even impossible to generate.;In this research, a mathematical solution framework that guarantees the optimal solution to semantic Web service composition is introduced. The Integer Programming (IP) based mathematical framework considers not only functional requirements but also QoS aspects of Web service composition. Furthermore, the framework can incorporate semantics-processing mechanisms into its IP formulation. The proposed approach guarantees the optimality of the solutions from both syntactic and semantic perspectives.;Finally, a k-best solution method for Web service composition is presented. Using k-best solutions provides a holistic view of the Web service composition solution space rather than a myopic view that is focused only on the optimal solution. Knowing k-best solutions and the summary statistics among them (such as the range of objective values) provides a broader view when composing Web services.;Key words: Web service composition, Semantic Web, Integer Programming, quality of service, k-best solution approach. |
