基于语义的动态服务组合技术研究

发布时间：2018-07-23 15:51

【摘要】：随着Web服务技术的不断应用与发展,特别是面向服务的体系结构(SOA)的成熟和推广,使得面向服务的计算逐步成为一门新兴的计算学科,并得到了学术界和工业界的广泛支持。与此同时,卫星信息资源服务也趋于多样化和颗粒化,要求各个功能服务单元不仅具备单项服务能力,各功能服务之间还要能够动态组合在一起,以提升整体服务的能力,卫星信息资源的动态服务组合方法成为新的难题。Web服务组合技术提供了一种高度互操作、跨平台和松耦合的框架体系,针对卫星信息资源服务体系的难题提出解决方法。通过小粒度的Web服务之间的通信和协作来实现大颗粒度服务,通过组合多个Web服务,实现服务增值和服务重用,将Web服务的潜能得以最大限度发挥。本文的工作主要包含以下方面： 1.提出了一种在语义层面上描述Web服务语言的模型。本文选取OWL-S为基础来描述Web服务,增加了服务的描述信息,在传统的IOPE的四种属性上,本文增加了语义描述服务目的和非功能属性,前者可以在服务匹配时提高匹配的效率,后者可以在服务匹配和服务组合时及时发现剔除不符合用户需求的服务和服务链。 2.提出一种基于因果关系矩阵和图论的动态服务组合方法。首先通过服务输入输出的参数构建相应的关系矩阵,存储不同服务之间输入输出的关系以及语义之间的联系,每个元素是一个分别包含当前服务、匹配程度和非功能属性值的三元组,以便在构建服务组合链时能够快速准确的找到上下关联服务。其次根据图论的方法,用回溯的方法寻找上一级满足要求的服务,在这个过程中一旦非功能属性值不满足用户要求就会舍弃该组合链,寻找另一条组合链。最后,在满足用户需求的服务链集合中选择最优的服务链。 3.搭建服务组合框架并进行测试以验证可行性。本文准备了大量服务数据,针对上面提到的基于因果关系矩阵和图论的服务组合方法,设计了测试方案,制定了相应的评判标准。通过实验结果证明在本文提出的框架下,能够高效和准确的实现服务的动态组合。
[Abstract]:With the continuous application and development of Web services technology, especially the maturity and popularization of service-oriented architecture (SOA), service-oriented computing has gradually become a new computing discipline, and has been widely supported by academia and industry. At the same time, satellite information resources services are becoming more diversified and granular, requiring each functional service unit not only to have a single service capability, but also to be able to dynamically combine the various functional services together, in order to enhance the overall service capability. The dynamic service composition method of satellite information resources has become a new problem. Web service composition technology provides a highly interoperable, cross-platform and loosely coupled framework system, and proposes a solution to the difficult problem of satellite information resources service system. Large granularity services are realized through communication and cooperation between small grained Web services. By composing multiple Web services, service increment and service reuse can be realized to maximize the potential of Web services. The work of this paper mainly includes the following aspects: 1. This paper presents a semantic model for describing Web services languages. In this paper, we choose OWL-S as the basis to describe the Web service, and add the description information of the service. On the four attributes of traditional IOPE, we add semantic description of service purpose and non-functional attribute, the former can improve the efficiency of service matching. The latter can find and eliminate services and service chains that do not meet the needs of users in time when service matching and service composition. 2. A dynamic service composition method based on causality matrix and graph theory is proposed. First of all, the corresponding relational matrix is constructed by the parameters of the service input and output, and the relationship between the input and output of different services and the relationship between the semantics are stored. Each element contains the current service separately. The triples of the matching degree and the nonfunctional attribute value can be found quickly and accurately when constructing the service composition chain. Secondly, according to the method of graph theory, the method of backtracking is used to find the service that meets the requirements of the upper level. In this process, once the non-functional attribute value does not satisfy the user's requirements, the combination chain will be abandoned and another combination chain will be found. Finally, select the optimal service chain in the set of service chains to meet the needs of users. 3. Build service composition framework and test to verify feasibility. In this paper, a large number of service data are prepared, and a test scheme is designed for the service composition method based on causality matrix and graph theory mentioned above, and the corresponding criteria are established. The experimental results show that the dynamic composition of services can be realized efficiently and accurately under the proposed framework.
【学位授予单位】：河南大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.09;TP391.1

【参考文献】