Based On The Theory Of Biological Evolution, Web Service Composition Evolution Mechanism Research

Internet is undergoing great change. In the past it is mainly used as a medium of information sharing and exchange, but now with the rapid development of network, it has become a platform with various types of resources sharing and integration. In this open and changing environment, service-oriented computing was born. In the service computing application development model, in order to improve the reusability of Web services, the combination of services is an effective way for resource integration and can also help user-defined combination of objectives achieved, producing value-added services. Because the Internet is an open computing platform, as time passing, the users’ needs and operating environment has also changed, so the Web services management become more important increasingly, good service management system can provide effective, real-time service information for web services users, developers, and providers. But now the majority of Web services management system can not provide global web services management solutions, and only use their own information processing model to deal with some Web services technical, such as QoS, licensing, classification, and so on.Currently, many domestic and foreign research institutions launched the research on how to build Web services evolution mechanism and the corresponding evolution strategy, and have made some valuable results. However, the calculation of the evolution of dynamic Web service composition framework, the theoretical model, the implementation mechanism, the key technology and other aspects are not only mature, but also faces many technical challenges. Mainly in:(1) service-oriented applications architecture. How to manage the quality of service and meet customers’needs efficiently;(2) Static evolution ensurence of services composition. How to ensure the correctness of the data flow of the entire service in the process of evolution;(3) Dynamic evolution ensurence of services compositon. Using the service selection, matching and replacement algorithm. Hw to dynamically modified a combination of services and ensure the quality of services and the needs of users at run time. In response to these problems, this paper made the following work:1. For the architecture of Web services applications, on the basis of the theoretical foundation of architecture and architectural style, our paper proposed a web services evolution cycle theory model which is the theoretical basis for effective controlling services evolution process, and then put forward a workable unified multi-tier web service composition evolution operating framework. The framework gives the abstract information model of scalable Web services and collect all the existing web services in this information model which provides the basis for the evolution of service unified management, while the service evolution strategy is also integrated in it. According to the different needs of users characteristics and technical features, specific service evolution strategy is applicated. When the evolution operation completed, the new service composition sequence will be given, while updating to the corresponding XML database to provide for users to use next time.2.For the correctness of static data flow in the evolution of web services, this paper put forward a Web service evolution strategy based on artificial immune algorithm and the dependency relationship graph. The strategy provides the related operations and using principles. Its characteristics are:The changing needs of Web services evolution in service evolution can be seen as the antigen, and then according to the evolution marked tectonic in WSDL files, construct the changing dependency graph to generate service dependency matrix as the input of the algorithm, and then calculating the antibody affinity and antibody-antigen affinity according to the method based on information entropy and affinity degree of the norm to evaluate the antibody in the population and finally after the updating operation in the group of memory cells, outputing the optimal antibody which can also be called optimal services portfolio evolution results.3.For the dynamic evolution of Web services, this paper proposes a web service combination evolution strategy based on hybrid co-evolutionary algorithm. The first of all, using the improved fuzzy C-means clustering algorithm to classify the web services evolution unit according to the application, then using the weighted value of the particle swarm algorithm to conduct an internal selection of the best for division-well sub-group and finally co-evoluting each subgroup. So as to achieve customer service requests, and can also select the optimal web services evolution combination. The experimental results show that the hybrid coevolutionary algorithm is superior to conventional cevolutionary algorithm in algorithm stability or algorithm computation time, and can provide quality and efficient service for a large number of service requests in the Web environment.