【作者】 王辉；

【导师】 冯志勇；

【作者基本信息】 天津大学 ， 计算机应用技术， 2010， 博士

【摘要】 随着互联网上Web服务(Web Service)数量的增长和Web服务应用的普及,以服务为中心的互联网正在形成。然而,无论是互联网上服务的发现和彼此感知,还是服务的自动组合与交互,都存在着不同程度的困难。基于语义和社会化关系的Web服务基础设施——服务网络的提出为解决此问题做了有益的尝试并取得了相应的成果。然而,最初的服务网络模型是粗糙的、基于弱语义和不易用的。围绕上述问题,本文进一步在语义层面提出了更为完善的服务网络模型,给出了服务网络结构和其中各元素的明确定义;提出了Web服务自动标注算法,完成了服务网络的完全语义化;解决了Web服务彼此感知的问题,在语义层面上构建了服务各层次之间的关系;给出了基于服务网络的需求分析和服务发现流程,为Web服务自动化应用做了有益的尝试。具体说来,本文的研究成果如下:1.研究了Web服务组织管理的现状,提出了更为完善和可操作的服务网络模型。对比之前的服务网络版本,本文提出了更为完善的三层网络结构,给出了各层结点的详细定义,明确定义了服务网络中的横向和纵向关系。2.研究了Web服务和领域本体的相关特征,提出了一种Web服务自动语义标注方法,为基于语义的服务网络的构建和服务关系挖掘奠定了基础。在详细对比分析了WSDL文档和领域本体的结构特征的基础上,提出了一种Web服务自动标注方法,实践表明该方法可以明显的提高标注准确性。3.研究了Web服务之间的关联特性,提出了一种Web服务关系挖掘方法,实现了基于语义关系的服务网络的自我演化。基于标注后的带有语义的Web服务,借助本体推理等手段,挖掘Web服务之间隐含的协作或竞争关系,既是服务网络中关系构建和演化的基础,又是自底向上的服务组合的有效手段。4.提出了一种基于服务网络的Web服务发现方法,设计了服务网络的基本运行机制。从用户需求分析出发,充分利用服务网络的三层结构和服务各层次级别的关系,给出了基于服务网络发现的基本流程和方法,定性的说明了服务网络在服务发现和服务推荐两方面的优势。总之,本文将Web服务组织成更为合理的网状层次结构,并提出了相应的算法来构建这样一个网络,实现了服务网络的自我演化,同时给出了基于服务网络应用的基本流程。这些使得服务网络真正成为面向互联网的Web服务基础设施。更多还原

【Abstract】 As Web services proliferate and being widely used, a service-oriented Internet is forming. However, there are still challenges in service discovery and sensoring each other, and also in service composition and interoperability. Service Network (SN), which is a Web service infrastructure based on semantic and social relations, was proposed to solve this problem. Unfortunately, the early version of SN was coarse, weak semantic-based and uneasy to use.Based on these problems, a polished and easy to use Service Network (SN), which has three-layer architecture and clearly defined relations, was proposed in this thesis. Furthermore, the thesis proposed a Web service annotation approach and used it in SN constructing; proposed a Web service relation mining approach and used it in SN evolving; proposed a requirements analysis and service discovery process based on SN, and did useful effort on Web service automatic processing. In details, the contributions of this thesis are listed as following:1. Proposition of a polished and easy to use SN model based on the research of Web service organization. Compared to the early version SN, vertical and horizontal relations are more clearly defined on semantic level, thus this three-layer architecture is more useful and better facilitated to service discovery.2. Proposition of a Web service annotation approach based on the contrastive analysis between domain ontology and WSDL document, which builds a foundation of constructing complete semantic SN and mining service relations. The practice proved that the approach proposed in this thesis could solve annotation problem with high efficiency. It has been used in SN constructing process.3. Proposition of a Web service relation mining approach in semantic level, which achieves self-evolution of SN. Based on the methods such as ontology reasoning, Competitive and Collaborative relations have been mined among annotated Web services. Service relation mining is both the foundation of SN evolution and the way of bottom-up Web service composition.4. Proposition of a process of requirements analysis and service discovery based on SN and also design of SN running mechanism. Fully use of the three-layer architecture and service relations, the process from user requirements to suitable services is designed, which qualitatively demonstrates that SN has the advantage in service discovery and recommendation.In summary, this thesis organized Web service in a more rational network, proposed relevant methods to construct and evolve such a network. Furthermore, the basic process of Service Network application was proposed. The researches in this thesis make Service Network a real Internet-oriented Web service infrastructure.更多还原