QoS感知的Web服务群体评价的研究与实现

发布时间：2018-05-13 02:40

本文选题：Web服务 + Quality　；参考：《昆明理工大学》2014年硕士论文

【摘要】：Web服务是由URI标识的一种应用软件,可以在网络上宣传、定位、并通过一系列消息机制建立连接。Web服务常通过彼此协作构建更复杂的应用程序,他们需要用功能属性和服务质量(QoS)属性加以描述。随着Web服务的不断发展,有越来越多的服务具有相似或相同的使用功能。为了方便和加快Web服务的发现,将功能相似的Web服务集中在一个社区,这样不仅便于管理,而且用户可以在社区中搜索所需服务。有许多研究旨在依据Web服务的QoS(服务质量)选择优质服务。比较常见的方法是依赖效用函数来选择最优服务,用户必须将自己的偏好表示成能够参与计算的权重。但是,对于有特定偏好的用户。他们难以准确评估服务QoS属性的权重。此外,对于没有明确偏好的用户,获得偏好信息以及QoS属性权重更加困难。因此,需要一种Web服务选择方法,为那些没有明确偏好的用户提供优质Web服务。传统QoS感知的Web服务选择方法需要用户给出精确权重以表达其对QoS属性的偏好。为解决用户无明确偏好情况下的服务选择问题,本文提出一种QoS感知的Web服务群体评价方法。可对社区内Web服务进行综合评价,并给出优劣次序,便于社区管理者掌握服务的整体情况以及作为用户选择服务的参考。首先,需要获取每个用户选择服务的历史评价信息,以QoS偏好明确的用户对服务的历史选择结果为基础。然后利用群决策理论中的社会选择函集结用户群的评价结果,计算社区中相似服务的群体评价。本文选用了Condorcet和Borda函数进行群体评价和对比。并对两个函数用于Web服务群体评价的适用性和有效性进行了理论分析和实验验证。理论分析对本文提出的方法是否具备一种有效的群体评价方法应当具有的明确性、公平性、匿名性、单调性和一致性进行了证明。结果表明群决策理论中的Condorcet和Borda函数适用于Web服务的群体评价。最后,根据本文提出的方法,设计实现了Web服务群体评价系统。该系统实现了对社区内相似服务的评价和优劣排序功能,并能对所得评价数据进行验证。验证模块包括了服务排序的公平性、一致性、单调性等性质和系统性能测试,实验表明了Web服务群体评价方法的有效性和系统的高效性。
[Abstract]:Web services are applications identified by URI that can be advertised, located, and connected through a series of messaging mechanisms. Web services often collaborate with each other to build more complex applications. They need to be described by functional attributes and QoS attributes. With the development of Web services, more and more services have similar or same usage functions. In order to facilitate and accelerate the discovery of Web services, Web services with similar functions are concentrated in one community, which is not only easy to manage, but also allows users to search for the required services in the community. There are many studies aimed at selecting quality services based on the quality of service (QoS) of Web services. The more common method is to rely on utility function to select the optimal service. Users must express their preferences as weights that can participate in the calculation. However, for users with specific preferences. It is difficult for them to accurately evaluate the weight of the service QoS attribute. In addition, it is more difficult to obtain preference information and QoS attribute weight for users without explicit preference. Therefore, a Web service selection method is needed to provide high quality Web services to users who do not have a clear preference. Traditional QoS aware Web service selection methods require users to give precise weights to express their preference for QoS attributes. In order to solve the problem of service selection without explicit user preference, this paper proposes a QoS aware Web service group evaluation method. The Web services in the community can be comprehensively evaluated, and the order of advantages and disadvantages is given, so that community managers can master the overall situation of the services and serve as a reference for the users to select the services. First of all, we need to obtain the historical evaluation information of each user to select the service, which is based on the historical selection result of the user who has a clear preference for the service by QoS. Then the social selection function in the group decision theory is used to aggregate the evaluation results of the user group and to calculate the group evaluation of similar services in the community. In this paper, Condorcet and Borda functions are selected for group evaluation and comparison. The applicability and validity of the two functions for Web service population evaluation are analyzed theoretically and experimentally. The theoretical analysis proves whether the proposed method has the clarity, fairness, anonymity, monotonicity and consistency of an effective group evaluation method. The results show that the Condorcet and Borda functions in group decision theory are suitable for group evaluation of Web services. Finally, according to the method proposed in this paper, a Web service group evaluation system is designed and implemented. The system realizes the function of evaluating and ranking the similar services in the community, and can verify the evaluation data. The verification module includes the fairness, consistency, monotonicity and system performance testing of Web services. The experiments show the effectiveness of the Web services community evaluation method and the high efficiency of the system.
【学位授予单位】：昆明理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.09

【参考文献】