综合功能属性和非功能属性的Web服务组合研究

发布时间：2018-06-06 18:09

本文选题：Web服务组合 + 功能属性　；参考：《西南大学》2014年硕士论文

【摘要】：随着电子商务快速发展和网络上Web服务的逐渐增多,Web服务作为一个Web上数据和信息集成的有效机制已经得到广泛的认可。如何快速、准确地从海量的Web服务中获取并组合成新的、复杂的服务同时并能够满足用户功能属性和非功能属性的需求已成为现阶段Web服务组合研究的热点问题。目前,针对Web服务组合的研究仍然存在负载失衡、效率低下、组合方法不当等问题,本文针对以上问题开展研究,主要工作如下： (1)基于反馈机制的Web服务组合框架的构建。本文在现有的Web服务组合框架的基础上,添加了反馈模块,建立了基于反馈机制的Web服务组合框架。反馈模块不仅能够及时反馈服务组合状态,以便于解决负载均衡、QoS值突变等问题,而且能够反馈服务请求者的评价信息,为其他服务请求者访问该服务提供访问依据。 (2)基于功能属性的候选服务集生成框架的确立。在考虑功能属性的匹配时,即服务选择模块的候选服务集生成阶段,为了提高功能匹配的效率,在已有的Web服务发现框架基础上,添加了基于用户评价和点击率的索引机制,设计了更新索引机制,建立了索引评价函数,提出了带索引机制的候选服务集生成框架。 (3)基于非功能属性的Web服务选择算法的建立。在考虑非功能属性选择时,即服务选择模块的组合服务集生成阶段,为了快速、准确的从海量功能相同或者相似的Web服务中选择出满足用户非功能属性要求的Web服务,提出了基于改进蚁群算法的Web服务选择算法。该算法在蚁群初始化的时候引入了混沌模型,在更新信息素的时候加入了混沌扰动量。一方面使用蚁群算法的信息素正反馈原理加强混沌优化算法中的搜索区域,增加搜索次数,降低了混沌优化算法搜索的盲目性。另一方面使用混沌优化算法的混沌初始化蚁群,提高蚁群的多样性,利用混沌优化算法的混沌扰动防止出现早熟现象。 (4)Web服务组合原型系统的设计与实现。在分析已有的组合系统的基础上,设计并实现了原型系统,该系统包括服务查询、发现和组合三个模块,验证了功能属性的候选服务集生成方法和非功能属性全局优化的Web服务选择算法的可行性和高效性。
[Abstract]:With the rapid development of electronic commerce and the increasing number of Web services, Web services have been widely recognized as an effective mechanism for the integration of data and information on the Web. How to quickly and accurately obtain and compose new and complex services from a large number of Web services has become a hot issue in the research of Web service composition at present. At present, there are still some problems in the research of Web service composition, such as load imbalance, low efficiency and improper composition method. The main work of this paper is as follows: 1) Construction of Web services composition framework based on feedback mechanism. Based on the existing Web services composition framework, this paper adds feedback module and establishes a Web service composition framework based on feedback mechanism. The feedback module can not only feedback the state of service composition in time to solve the problem of load balancing and QoS mutation, but also feedback the evaluation information of service requestor. It provides access basis for other service requesters to access the service. 2) the establishment of candidate service set generation framework based on functional attributes. In order to improve the efficiency of function matching, a new indexing mechanism based on user evaluation and click rate is added to the existing Web service discovery framework when considering the matching of functional attributes, that is, the candidate service set generation stage of service selection module. An index updating mechanism is designed, an index evaluation function is established, and a candidate service set generation framework with index mechanism is proposed. When considering non-functional attribute selection, that is, the generation stage of composition service set of service selection module, in order to quickly and accurately select Web services that meet the requirements of users' non-functional attributes from the massive Web services with the same or similar functions. A Web service selection algorithm based on improved ant colony algorithm is proposed. Chaotic model is introduced in ant colony initialization and chaos perturbation is added to update pheromone. On the one hand, the pheromone positive feedback principle of ant colony algorithm is used to strengthen the search area of chaotic optimization algorithm, increase the number of searches, and reduce the blindness of chaotic optimization algorithm. On the other hand, the chaos of chaos optimization algorithm is used to initialize the ant colony to improve the diversity of the ant colony, and the chaos disturbance of the chaos optimization algorithm is used to prevent the premature phenomenon. Based on the analysis of the existing composite system, a prototype system is designed and implemented. The system includes three modules: service query, discovery and composition. The feasibility and efficiency of the candidate service set generation method for functional attributes and the Web service selection algorithm for global optimization of non-functional attributes are verified.
【学位授予单位】：西南大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.09

【参考文献】