Research On Methods Of Semantic Web Service Discovery And Composition

Distributed computing has obtained a significant development with the rapid development of Internet technology and the growing number of web services as well as the great popularity of the Internet. how to discover and combine services flexibly, fast and accurately, which meets users’ needs, from the numerous web services, is the premise to realize service sharing and multiplying and also one of the key issues facing the present stage. In order to solve the problem of low efficiency of Web Service in Discovery and Composition, a click-through rate index and adaptive genetic algorithm-based web service discovery and combination method is proposed in this thesis based on the analysis of web services discovery and combination, and the evolution process of genetic algorithm.The main work of this paper is as follows:Firstly, the establishment of the web service index module. By analyzing the characteristics of the present semantic web service discovery framework and the user search service, the web service discovery method was proposed based on the click rate index. Indexing the web services with higher click-through rate in the Universal Description, Discovery and Integration repository based on the click rate of service in a certain period of time, we can firstly match the services with a higher access rate in the query process. Thus improve the efficiency of Web service discovery.Secondly, the establishment of the semantic web service description model. On the basis of analysis and research of existing semantic web service description language, a new web service description model is proposed, adding the description of the click-through rate of the service, to record the clicking rate of service and provide basis for the establishment of the index.Thirdly, the design of self-adaptive Genetic algorithm. Through the analysis of the evolutionary process of genetic algorithm and the characteristics of web service composition, a web service composition method based on adaptive genetic algorithm was proposed. In the evolutionary process of genetic algorithm, individual crossover and mutation operators is adjusted according to the individual’s Fitness Value and evolutionary stage to improve the convergence speed of genetic algorithm. Fourthly, the introduction of tabu list of genetic algorithm. To prevent the genetic algorithm from converging to false optimal solution, tabu list was introduced into genetic algorithm based on the analysis of the taboo search algorithm and web service composition process, to record the optimal solution in the evolutionary process and avoid the duplication local search in the optimal solution, thus improve the quality of composite services.Finally,The paper verifies the performance of the web service discovery method based on the click rate index through the simulation of Web service discovery process firstly, and then a experiment was carried on to simulate web service composition process with adaptive genetic operators to verify the validity of algorithm, it’s ability to meet user preferences through setting different weights to quality.The experimental result shows that, the web service discovery method based on the click rate index not only ensures the precision ratio,but also improves the efficiency of Web service discovery; Web service combination genetic algorithm that Genetic operators changed with the evolutionary stage and individual fitness can quickly improve the fitness of individuals and get better optimal solution under certain evolution number conditions,it demonstrates that the proposed method is effective;The changes of optimal solution’s quality attributes under different weights of quality shows that the Web service composition method in this paper can meet the user’s preferences well.