| Web services are software systems designed to support machine-to-machine inter-operation over the Web. Many researches have been carried out for the web service standards, and these efforts have significantly improved flexible and dynamic functionality of Service Oriented Architecture (SOA) in the current web services. However, there still remain a number of research challenges one of which is the web service composition (WSC) problem---when a single web service does not satisfy a given requirement, one wants to automatically combine web services to satisfy the requirement entirely. In this dissertation, we tackle this WSC problem in three levels, i.e., a signature level, a behavior description level and a QoS description level based on the web service descriptions.;For a signature-level approach where each web service is described with its signature in WSDL, we first analyze the topological landscape of the web service networks formed by real-world web services based on graph theory. We then propose a SAT-based algorithm based on the analysis.;Since web service composition based on signatures has a limitation due to a lack of information provided by a WSDL signature, we need additional information, e.g., a semantic annotation of data and/or a behavioral description of a web service function. Focusing on the latter one, we first define a realistic model for the WSC problem on behavioral descriptions, and investigate the computational complexities for the composition of web services on restricted (i.e., with full observation) and general cases (i.e., with partial observation). We then prove that the WSC problem with full observation is EXP-hard and the WSC problem with partial observation is 2-EXP-hard. To solve these high complexities, we propose approximation-based algorithms using abstraction and refinement.;The previous two approaches consider only functional requirements specified WSDL or BPEL. However, non-functional ones, such as Quality of Services (QoS) constraints help clients to select a service provider with good quality. In this case, the main aim of the WSC problem is to find a composite web service which satisfies a given complicated task with the optimal accumulated QoS value, which is called QoS-aware WSC problem. We first propose applying anytime algorithm based on beam stack search to the QoS-aware WSC problem. Moreover, to improve the basic anytime algorithm, we propose dynamic beam width with more heuristics, i.e., short backtracking and upper bound propagation. |
