资源受限环境下基于QoS的Web服务组合研究

发布时间：2018-07-16 08:37

【摘要】：随着Web服务技术的发展和广泛应用，越来越多的服务提供者将服务以Web服务的形式发布出来，网络上出现了大量功能相同或相似的服务。由于单个原子服务无法满足服务请求者复杂的服务需求，需要将现有的单个原子服务进行整合，促使了服务组合技术的研究和发展。同时服务使用者也不仅仅满足于服务的功能属性，对于服务的QoS属性也提出了更高的要求。因此如何选出一组合适的服务，保证单个原子服务质量的同时，，也能确保组合服务质量全局最优，成为Web服务组合领域的研究热点和难点。此外，越来越多的服务部署在资源受限的环境中，环境资源的有限性会影响到服务能否成功执行，服务质量的高低以及服务使用者的满意度。本文在分析传统Web服务的基础上，结合资源环境特点，扩展服务描述，增加服务运行所需资源描述信息、服务所处环境资源信息等，建立资源受限环境下服务的QoS模型，支持服务的动态部署。从QoS和环境资源两个角度出发，计算服务的可用度，用于衡量资源受限环境下服务的质量。提出了资源受限环境下基于QoS的服务组合实现框架，并在此框架下进一步研究了资源受限环境下的服务组合优化问题。算法的主要思想是：该方法首先对候选的实例服务根据服务的可用度进行筛选，降低环境资源对服务质量的影响，提高服务的可靠性，然后根据服务的QoS属性，在每个服务组进行聚类操作，选取服务簇中质量较优的服务，降低候选实例服务的规模，减少搜索空间，最后采用带备选组合服务方案集的改进后的粒子群优化算法实现组合服务的全局优化。实验结果表明，本文提出的算法能够快速地实现组合服务的全局优化，有效地保证组合服务的质量，提高了组合服务的成功率。
[Abstract]:With the development and wide application of Web services technology, more and more service providers publish their services in the form of Web services. Because the single atomic service can not meet the complex service requirements of the service requester, it is necessary to integrate the existing single atomic service, which promotes the research and development of the service composition technology. At the same time, the service consumer is not only satisfied with the functional attributes of the service, but also puts forward higher requirements for the QoS attribute of the service. Therefore, how to select a suitable group of services to guarantee the quality of service of a single atom, and at the same time to ensure the global optimization of the quality of service composition, has become a hot and difficult research topic in the field of Web service composition. In addition, more and more services are deployed in resource-constrained environment. The limitation of environmental resources will affect the success of service execution, the quality of service and the satisfaction of service users. Based on the analysis of traditional Web services, combining the characteristics of resource and environment, this paper extends the description of services, increases the description of resources needed for service operation, and establishes the QoS model of services in resource-constrained environment. Support dynamic deployment of services. From the point of view of QoS and environmental resources, the availability of services is calculated to measure the quality of services in resource-constrained environments. This paper proposes a QoS based service composition implementation framework in resource-constrained environment, and further studies the optimization of service composition in resource-constrained environment. The main idea of the algorithm is as follows: firstly, the candidate instance services are screened according to the availability of the service, which reduces the impact of environmental resources on the quality of service and improves the reliability of the service. Clustering operation is carried out in each service group to select the best quality service in the service cluster, to reduce the scale of candidate instance service and to reduce the search space. Finally, the improved particle swarm optimization algorithm with alternative composite service scheme set is used to realize the global optimization of composite services. Experimental results show that the proposed algorithm can quickly achieve global optimization of composite services, effectively guarantee the quality of composite services, and improve the success rate of composite services.
【学位授予单位】：南京邮电大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.09

【参考文献】