| Web service matching is a process that service requesters search the service their required in the different web services. Due to the traditional web service matching based on key words and OWL-S exists defects such as low recall rate and low precision rate, and not global optimization with matching information and so on. And at present, in the majority of study, researchers only concern about the function matching in web service matching, and don't consider of the non-functional property which includes quality property. In the paper, for the above problem, web service QoS ontology model,web service matching model supporting QoS and web service matching method based on global optimization information are researched for improving the process of web service matching. The study work in the paper are follow:Firstly, web service matching model supporting QoS is builded. In the matching model, a three hierarchy matching method is proposed. At first, a web service matching method based on function information is proposed. So that a primary web service set meeting user is got.Then a semantic web service matching based on ontology volabulary is presented, in which user's feedback is to as attributes' weight, ontology volabulary is devided into numeric type and interval type, and maching algorithm of web services is proposed. Thus a second-rate web service candidacy set is got.Last according to servicer's QoS request, whether a semantic web service matching method based on improving particle swarm algorithm is implemented or not. And a final web service set meeting user is got.Secondly, web service QoS ontology model is constructed. Due to web service QoS information are indefinite, QoS is expanded based on previous QoS model information and QoS ontology model is established. In the QoS ontology model, ervery QoS element is definite.Thirdly, according to servicer requst based on the accureate type QoS, a semantic web service matching method based on improving particle swarm algorithm is proposed. To implement web service global optimization matching idea, a web service matching method based on improving particle swarm algorithm is proposed. Although particle swarm algorithm is a maturing multi-objective optimization algorithm, a premature convergence problem and a not well distributed Pareto-optimal solution problem exist. So in order to keep some particles with smaller constraint violations, the comparison strategy of particles using threshold is revised. Then, to find a set of diverse and well distributed Pareto-optimal solutions, a new crowding distance function is designed and a new mutation operator of total force is presented. Last, improved particle swarm optimization algorithm is proposed, and it is brought to web service matching method.Fifthly, according to servicer requst based on the fuzzy QoS, a semantic web service matching method based on fuzzy particle swarm algorithm is proposed. A fuzzy particle swarm algorithm is revised by fuzzy theory applied to the particle swarm algorithm. To improve algorithm precision, the moving direction of a particle is influenced by multi-particle instead of best particles. Then to improve premature convergence problem, increasing inertia factor is presented.At last, experiments based on recall radio and precision radio are made, and the the alogorithm convergence and stability of PSO-WSM(web service matching alogorithm supporting QoS based on particle swarm optimization),IHPSO-WSM (web service matching alogorithm based on improving hybrid particle swarm optimization)and FPSO-WSM (web service matching alogorithm based on fuzzy particle swarm optimization)are verified. Experiment results have shown the effectiveness of the proposed web service matching model. |
PDF Full Text Request