An Efficient Qos-Based Ranking Model for Web Service Selection with Consideration \of User's Requirement

The power of Web services to address the incompatibility issue of standalone systems, has led them to play a major role in business application development. Adopting an efficient and effective method to locate and select desired services among thousands of available candidates is an important task in the service-oriented computing. As part of a Web service discovery system, the ranking process enables users to locate their desired services more effectively. Many of the existing approaches ignore the role of user's requirements which is an important factor in the ranking process. In this thesis we enhance a vector-based ranking method by considering user's requirements. The vector-based model is chosen because of its simplicity and high efficiency. We evaluate all Web services in terms of their similarity degrees to the optimal or the best available values of each quality attribute, and penalize the services that fail to meet the user's requirements. Through our extensive experiments using real datasets, we compare the improved algorithm with other approaches to evaluate it in terms of efficiency (the execution time to return the result) and quality of the results (accuracy). Cherie Ding