Research On Web Service Selection

With the fleetly development of Web Service, the amount of the Web Services which have the same function is getting more and more on the Internet, the distinctness among these Web Services is the different QoS(Quality of Service). When the users want to use the Web Service, they confront with the problem is that how to select a Service from the Web Services with the same function, QoS of which can content to his request best of all. With the fiercely competition in Web Service, the Web Service provider confront with the problem is that how to select the appropriate request to ensure the accomplishment of the request under the limited system resource, and also realize the most adequately use of the provide's resource, realize the most income of the System. This paper enters on the provider distributing resource, the Multidimensional Knapsack Algorithm is researched to slove the problem of the most adequately use of the provide's resource with insuring the adequate resource of the requested Service, and the QoS Measure Function is used to calculate the income of requester, and select the Service which make the most income of the use.In this paper, firstly, the correlative knowledge of Web Service is introduced, specify the WSBS-Model (Web Service Bidirectional Selection-Model) designed in this paper. The Model includes seven function Engines, they are Service Request Sort Engine, Queue Manage Engine, QoS Measure Engine, Provider Select Request Engine, Requester Select Service Engine, Web Service Engine, Web Service Appraise Engine. The function and the design elements of each Engine are analyzed in detail.Next WSBSW(Web Service Bidirectional Selection Worksystem) which is realized according to WSBS-Model in the paper is introduced, and the design of the develop environment,database of WSBSW is also introduced. The Discrete Binary Particle Swarm Optimization Algorithm is researched to solve the Multidimensional Knapsack Problem in the paper, the analyse and the emulational experiment have been done. According to compare the performance between the Discrete Binary Particle Swarm Optimization Algorithm and other Algorithms in resource distribute optimization,reponse time,the number of the iterative step, the conclusion is that the performance of the Discrete Binary Particle Swarm Optimization Algorithm is better than other Algorithms, so the Discrete Binary Particle Swarm Optimization Algorithm is used to solve the Multidimensional Knapsack Problem. Finally the experiment result prove the feasibility of WSBS-Model.