【作者】 侯青；

【导师】 张广泉；

【作者基本信息】 重庆师范大学 ， 计算机软件与理论， 2011， 硕士

【摘要】 随着Web服务技术的快速发展,软件的应用已逐步从静态的、封闭的模式向动态的、开放的模式转变。软件可以通过复用与组合的方法开发出新的满足需要的应用,从而提高了开发效率,节省了资源。如何在因特网复杂的环境下,提供有效、按需分配的Web服务成为目前服务计算领域的一个研究热点。本文主要研究了支持QoS约束的Web服务发现、组合和选择策略。首先,根据当前一些Web服务发现模型存在着功能不完善、缺乏反馈机制、缺乏扩展性、不能满足偏好型用户需求等不足,本文提出了一种支持QoS约束的Web服务发现和组合框架(WebDAC)。该框架模型首先对WSDL和UDDI进行了改进,使其支持Web非功能性描述并具有可扩展性;其次引入了Web服务评估系统,支持Web服务监控和反馈,保证服务质量评价的真实性和有效性;该模型还对服务选择功能进行了完善,可以根据用户需求的服务类型(单个Web服务或者组合Web服务),相应进入局部选择服务或全局优化选择服务阶段,同时偏好型用户可以根据自己需求来设定不同的权值,满足个性化需求。其次,本文分析了当前服务质量模型的优缺点,并在此基础上建立了两种客观全面的QoS模型,即单服务QoS模型和组合服务QoS模型。单服务QoS模型通过对吞吐量、执行时间、延迟时间、可靠性、费用、安全性、声誉7个属性参数进行定义和量化,并引入概念权重和服务请求权重,形成一个综合的QoS评价模型;组合服务QoS模型是在单服务QoS模型基础上,考虑了顺序、选择、并行、循环四种结构化流程,建立起综合的组合服务QoS评价模型。最后,本文在服务选择方面虑了两种选择策略,即局部选择和全局优化选择策略。局部选择策略在满足用户对Web服务功能需求的基础上,它以服务注册参数、用户反馈信息和实时监测数据为依据对Web服务的非功能性属性进行量化,通过四阶段阶段选择算法,实现服务质量动态评估,保证web服务QoS信息的公平性、可靠性和实时性;在组合服务优化选择方面,以往人们的研究主要以线性规划和遗传算法为主,本文则将禁忌搜索算法应用在组合服务优化方面,提出了一种基于禁忌搜索算法的全局优化选择策略;禁忌搜索算法是局部邻域搜索的扩展,是一种全局逐步寻优算法,实验结果表明,该算法是实用、有效的。更多还原

【Abstract】 With the rapid development of Web services technology, computer software applications gradually shift from static closed model to dynamic opening model. Software develop to meet new applications by reusing and combining and it can be developed to improve development efficiency and save resources.In the complex environment of the Internet, how to provide effective and needed Web services become a research hotspot in the current service calculation domain.This paper mainly studies Web service discovery, composition and selection strategies which based on QoS-constrained.Firstly, some Web service discovery models have many shortcomings at present , for instance functions is imperfect, lack of feedback mechanisms, lack of scalability,can not meet the needs of preferene users and other shortcomings,and in view of this, this paper propose a QoS-constrained Web service framework which suposes Web service discovery and composition (it named WebDAC). First of all,WSDL and UDDI was modified to support non-functional descriptions and Web scalability in the framework model; Then, in order to ensure the assessment of quality be authenticity and validity, we introduce a Web service evaluation system which support Web service monitoring and feedback ; The model also improve function of service selection which enter to the local selection stage or global optimization selection stage, according to the type of service which user’needs (a single Web service or a combination of Web services), meanwhile individual customers can set different weight based on their own needs to satisfy preferene needs.Secondly, this paper analyzes the advantages and disadvantages of the current model of service quality, and based on this, it have established two models of service quality which is objective and comprehensive. The two models are named single-service QoS model and the composite service QoS model. In the single QoS model of service ,we define and quantify the seven property parameters including execution time, delay time, reliability, cost, safety and reputation meamwhile introduce the concept weights and service requests weight, finally,it has formed a comprehesive QoS evaluation model.;QoS model of composite service is based on the single QoS model of service and consider the four structured process of sequence, selection, parallel and cycle which establish a comprehesive QoS evaluation model of composite service.Finaly,in the service selection, we consider two selection strategies which are local selection and global optimization strategy. Local selection strategy satisfies uers’functional requirements of Web service and estimates Web service’s non-functional criterions dynamically based on the criterions registered by Web service,information fed back by users and service’s realtime data, in order to evaluate Quality of Service dynamically and real-time updates, ensures that QoS information is impartiality,trustworthiness,and real-time by the four-stage selection algorithm;In the past, the people using linear programming and genetic algorithm to study the optimization of composite service.In this paper Tabu search algorithm will be applied to Web services composition optimizatio.We propose a new selection strategy of composite service which is based on Tabu search algorithm;Tabu search algorithm is an extension of local neighborhood search, and it is a global optimization algorithm step by step.The experimental results show that the algorithm is practical and effective.更多还原