【作者】 蔡海尼；

【导师】 唐远炎；

【作者基本信息】 重庆大学 ， 计算机科学与技术， 2014， 博士

【摘要】 Web服务集成是构建基于面向服务体系架构（Service Oriented Architecture，SOA）的企业信息框架的关键技术，其基本思想是把服务作为基础构件，支持快速、低成本的分布式甚至异构环境的应用组合开发，解决不同系统/平台间的应用整合问题。服务集成的目标是实现企业现有内部软件资源和外部服务的重用，并在此基础上组合成具有一定功能的软件系统，以提高软件构造的效率，满足不断变化的用户需求。从技术角度分析，服务集成是一个服务发现、选择和组合执行的业务流程。现行的Web服务技术在服务发现效能、动态服务选择与组合方面，缺乏有力的支撑。语义Web服务为上述问题提供了新的解决途径，利用其良好的语义描述能力和逻辑推理能力，能自动、准确、高效地进行服务集成，动态生成和执行业务流程。目前，在语义Web服务的发现、选择和组合领域已取得一定研究成果，但尚未提出一个完整的基于语义的Web服务集成方案。本文从基于语义的服务描述模型、服务匹配与发现、服务选择与组合、数据服务、服务集成框架等方面进行研究，构建了勘察设计企业协同设计平台原型。论文主要研究工作如下:①针对现有Web服务描述语言WSDL(Web Service Description Language)不能提供语义信息的局限性，建立了基于语义的Web服务描述模型。该模型将描述Web服务语义的主流规范OWL-S(Web Ontology Language for Services)和现有的WSDL、UDDI等成熟技术充分结合，用WSDL描述服务名称、服务描述，URI地址信息等具体信息，用OWL-S描述服务功能等抽象信息，同时支持基于语法和语义的匹配与发现；并在功能语义的基础上增加服务的QoS(Quality of Service)本体描述，为服务选择和组合提供支持。②针对基于UDDI(Universal Description, Discovery and Integration)的传统Web服务发现机制只能提供关键字查找，不能进行语义匹配的问题，提出了一种基于语义的Web服务分层匹配模型。模型分为服务类型匹配、服务描述特征聚类和服务功能匹配三个层次，各层采用相应的算法，逐步缩减匹配问题的规模，改进了服务匹配的效率和精准度。在此基础上，对UDDI进行语义扩展，构建支持语义的服务发现框架，并针对实际SOA系统中服务访问规模大、频率高的特点，设计了一种服务发现优化方案。③为适应勘查设计领域复杂多变的运行环境和任务需求，提高服务选择与组合的效率，研究了基于任务分解的语义Web服务选择与组合机制。提出了一种基于迭代学习的任务分解算法，优化服务的选择与组合过程；并针对服务组合过程中运行环境的异构性和资源负载的动态变化，设计了一种面向任务的自适应服务组合机制。④从数据集成和应用集成两方面进一步分析Web服务集成中的关键问题，提出了一种基于语义的Web服务集成框架，并应用于勘察设计企业协同设计平台；针对传统异构数据访问方式存在的问题，采用数据服务的方式，对底层数据源直接进行封装，提供更加灵活的数据处理功能，并向上层提供易于集成的标准Web服务。更多还原

【Abstract】 Web Service integration is the key technique for constructing the enterpriseinformation framework based on Service Oriented Architecture (SOA). Through serviceintegration, the services are treated as basic components to support rapid, low-costapplication development in distributed or even heterogeneous environment. The aim ofService Integration is to reuse both the enterprise’s internal software resources and theexternal services to improve the efficiency of software construction and to meet theever-changing needs of customers.From the technology prospective, Service Integration is a business process ofservice discovery, service selection, service composition, and service execution. Theexisting Web Service technologies cannot provide strong support for efficient servicediscovery, service selection and dynamic service composition. It is well appreciated thatSemantic Web service provides a new solution for the aforementioned problem. Withpremium capability in terms of semantic description and logic reasoning, Semantic WebService can complete the process of service integrating in an automatic, accurate, andefficient way. Semantic Web service can also create business processes and executethem dynamically. Though extensive research has been performed and abundantfindings has been achieved in the field of Semantic Web Service Discovery, Selectionand Composition, a complete semantic-based Web service integration solution has notemerged.Based on the above background, this thesis presents the following key issues:(1) A semantic web service description model has been established to tackle thelimitations that the Web Service Description Language (WSDL) cannot providesemantic information. This model combines the Web Ontology Language for Services(OWL-S) and the existing technologies. It can support both syntax-based andsemantic-based service matching and discovery by using WSDL to describe concreteinformation such as service name, service description and URL address, and OWL-S todescribe abstract information such as service function. The model also adds QoSontology to semantic description of service function to support service selection andservice composition.(2) The traditional web service discovery mechanism based on the UniversalDescription, Discovery and Integration (UDDI) is limited to offering the function of keyword matching instead of semantic matching. To mitigate this problem, asemantics-based Web service hierarchical matching model has been proposed. Themodel is divided into three layers including service category matching, servicedescription feature clustering and service function matching. It improves the efficiencyand accuracy of service matching by performing relative algorithms for each layer toreduce the scale of matching problem. In line with this, a service discovery frameworkis built with semantic extension to UDDI in this study, and a service discoveryoptimization scheme is designed against large-scale and high-frequency serviceaccesses in actual systems based on SOA.(3) In order to meet the needs of the complex tasks and running environments inthe survey and design field, this study focus on a service selection and compositionmechanism to improve the efficiency. A task decomposition algorithm based on iterativelearning is proposed to optimize the service selection and composition process, and atask-oriented adaptive service composition mechanism is designed aiming at theheterogeneity of system environment and the dynamic changes of resource load.(4) The solution of Web Service Integration are discussed within the aspects of dataintegration and application integration. An integrated framework of semantic Webservice is built and applied in the field of exploration and design. To solve the problemsexisting in traditional heterogeneous data access mode, this study utilized the dataservice to directly encapsulate the underlying data. The data service provides moreflexible data processing, and can be integrated easily as a standard Web service.更多还原