Research Of Intelligent And Dynamic Web Service Composition Based On QoS

In recent years, with the development of Web service technology and service composition,the number of Web service with the same or the similar function increases gradually, so Webservice selection based QoS has become key technology of Web service composition filed.However, how to enable effective choice Web service from the same or the similar Web servicesand improve the quality of Web service composition at the same time is the important researchsubject.Currently, Web service composition based on QoS emerge some problems. Firstly, eachattribute of QoS use fixed weights in the existing QoS comprehensive calculation model, whichcompletely ignore the relationship between each attribute weights and the actual value and can’treasonable to accurate calculation Web service quality. So the selected candidate Web servicesaccording to the comprehensive value of QoS may not be appropriate service, which reduces thetotal quality of Web service composition. At the same time the service selection algorithm inservices composition mostly adopt classical mathematical optimization algorithm, such as thestandard particle swarm optimization algorithm. But the standard swarm optimization algorithmmay appear premature convergence or slow convergence defects speed and so on, which mayserious impact on the efficiency of Web service composition.To resolve above problems, this paper has the following research based on the analysis andresearch of the existing related QoS calculation model and related services selection algorithm.(1) This paper improves QoS comprehensive calculation method. The improved methodestablishes state variable weight vector and dynamic adjusts each attribute weight by adoptingvariable weight synthesis, which could reasonable to accurate calculation QoS comprehensivevalue.(2) According to the defects and shortcoming of the standard swarm optimization algorithm,this paper proposes variable step and dynamic learning factor. Then according to test function,the parameters of improved algorithm by continuously simulation experiment so that it will beable to better adapt to the needs of the function and improve the execution efficiency of servicescomposition by improving update swarms position speed. (3) According to the requirements and restrictions of users for Web service composition,this paper design particle fitness function including target function and constraints. Comparedwith the execution time of the standard particle swarm optimization algorithm and the improvedalgorithm. Finally, Experimental results show that improved QoS comprehensive calculationmethod gets better result by comparing with the whole quality of the existing QoScomprehensive calculation method and improved QoS comprehensive calculation method.