Research On Top-K Web Service Composition Method And Relevant Theory

Service computing, as a novel mode for software deployment, dilivery and maintenance, well meets the interaction requirements between enterprises or between enterprises and users. However, the development and maturation of Web service technology and cloud computing platforms makes the number of Web services available on the Internet has expanded dramatically. Then the quality evaluation of Web Services becomes an important part of Web service technology. Unfortunately, the fairness of services’quality evaluation and the standard specifications for evaluation is not yet conclusive. Besides, when the number of various Web services is very large, it becoms a challenge to automatic composing services based on semantic for geiven user’s requirements in lower time overhead.Based on previous research, this paper has studied the top-k Web service composition method and relevant theories. This paper proposes an personalized requirement-based atomic Web services’evaluation method and a QoS pruning-based top-k automatic service composition method.Personalized requirement-based atomic Web services’evaluation method takes into account the facts, i.e., different functional equivalent atomic Web services can meet different users’personalized requirements and the QoS experience is due to both the service provider and the service consumer. Firsly, the method models the personalized service requests. Then, based on the ideology of collaborative filtering, it performs user-centered QoS prediction. Finally, the performance analysis and efficiency optimization is discussed.Considering the influential facts of atomic services’availability and exception during executing composite service, the automic service composition method produces QoS top-k composite services. Firsly, this method models the service composition problem and starts initialization in an open service set. Then, a QoS pruning-based top-k automatic service compostion algorithm is designed. To improving efficiency, a service filter algorithm is designed for reducing solution space, and a greedy-based pruning algorithm for speeding up top-k composite services’ search. Lastly, the implementation of automic service composition system and its experiment evaluation validate that, the method can get respectable result in both effectiveness and efficiency.