【作者】 秦学；

【导师】 唐胜群；

【作者基本信息】 武汉大学 ， 计算机软件与理论， 2013， 博士

【摘要】 智慧城市建设的主要目标是“公共服务更加普惠、社会管理更加高效、产业体系更加优化”,智慧政务是智慧城市建设的基础和核心。智慧政务的建设内容主要包括政务云基础设施统一管理调度、信息资源整合与共享、业务集成与协同、智慧应用及服务、基于大数据处理的智慧决策以及安全保障体系等。智慧政务应在电子政务的基础之上,解决大范围的整合多部门信息资源、集成多种业务系统功能,提供按需动态组合的智慧服务。其关键技术是异构信息资源的深度整合、异构系统之间的动态协同等,而现有的基于中间件的集成方法已无法完全解决这些问题。对面向服务架构(SOA)和Web服务技术进行扩充,使其具有数据和功能的形式化描述、语义自识别、语义符合性组合等特点,对于智慧政务建设具有广泛的实用价值和理论价值。本文以武汉市智慧政务云平台规划设计为背景,研究解决我国智慧政务建设中普遍存在的跨领域、跨时空的业务集成与协同的关键问题。现有的SOA和Web服务技术在智慧政务系统集成和业务协同时面临着挑战,这是由于智慧政务覆盖范围巨大,但又要求服务按需定制与组合。如何发现定位需要的服务功能、如何快速地组合多个服务满足新的应用需求是需要解决的两个主要难题。引入语义技术解决智慧政务系统中的四个基础性问题：对业务服务进行精确语义描述,提供服务的准确识别；利用推理技术准确地发现和定位服务；利用语义信息辅助实现服务组合；利用语义技术实现服务的替换和调用等是本文的主要研究工作。结合电子政务的特点和语义技术发展,本文提出了政务Web服务的语义模型,并给出了本体定义。为了有效的组织和管理数量庞大的服务,本文研究建立了电子政务的业务分类模型和本体。同时,根据描述服务功能语义的需要,依据相关的国家标准开发了电子政务信息资源分类本体和公共数据元本体,为标注服务的IO语义标注提供依据。对于服务功能中的PE语义,则使用通用的OWLDL机制进行描述。为了保证服务具有准确和足够的语义信息,本文专门设计了服务的语义注册接口。通过这些工作,研究解决了服务的语义建模问题,为服务的发现和组合两个关键问题的解决创造了条件。基于语义推理的服务发现能准确定位服务,但是囿于推理技术固有的复杂性,在面对智慧政务巨大的服务空间时,推理的效率是无法接受的。本文研究了分阶段多策略的服务发现方法。该方法引入了关系查询技术,采用关系数据库来存储管理服务注册信息,先进行语义辅助的基于关键字的检索,以快速缩减服务数量,然后在这个相对较小的服务空间中进行语义推理来准确定位候选服务。得益于关系数据库高效的检索效率,该方法能很好的平衡服务发现的效率和定位准确性。对于智慧政务业务协同中的另一个关键问题——服务组合,本文研究一种基于可组合性的服务组合方法,该方法是对基于与或图组合方法算法的改进。由于与或图搜索算法具有很高的时间复杂度,因此必须降低问题搜索空间的规模,以提高方法的实用性。本文研究的方法在构造服务组合问题状态图时,不是直接使用注册服务来生成,而是依据服务的业务类别提取抽象服务来构造,从而有效降低搜索状态图的规模并能保持一定稳定性。对于与或图只能描述服务的IO语义而不能刻画PE语义,引入了服务可组合性模型来解决,它全面考察服务之间的语法、语义的可组合性,通过计算候选组合方案的可组合度,可以保证组合方案中组件服务之间的语义一致性。可组合性模型另一个用途则是作为替换服务选择的重要参考指标,它可以考查替换之后的组合服务执行成功率。本文研究了智慧政务中系统集成和业务协同过程中的关键问题,针对SOA和Web服务技术在解决这些问题面临的困难,提出了有效的解决方法。并且,将这些研究应用到了武汉市智慧政务云计算平台的规划和设计。更多还原

【Abstract】 Smart city construction’s goals mainly include three aspects:public services will be more popular and favorable; social management will be more efficient; and industrial systems will be more optimized. The realization of smart government is the foundation of smart city construction. The construction of smart government mainly includes the following aspects:the unified management of the e-government cloud infrastructure; integration and sharing of information and resources; business collaboration and cooperation; application and service of wisdom; and intelligent decisions based on large data processing and safety guarantee system, etc. Based on the e-government, smart government is to integrate multi-sectoral information resources in a wide range, integrate a variety of business functional systems, and to provide intelligence services according to the combination of dynamic needs. Its key technology is the deep integration of the heterogeneous information resources, and dynamic collaboration between heterogeneous systems, etc., while the existing integration method based on the middleware has been unable to solve these problems completely. Expanding of the service oriented architecture (SOA) and Web service technology will make them charactered by formal description of function and data, semantic self-recognition, and combination of semantic conformity, etc., which will bring extensive practical and theoretical value to the smart government construction.Taking the smart government cloud platform design of Wuhan as the background, this paper was to solve the key problems of business collaboration and cooperation in the construction of smart government in our country, which were widespread across domains and across time and space. The existing SOA and Web services techno logies were faced with challenges in business collaboration and system integration, the reason of which was the great coverage of the smart government, and to ask for a service on customization and combination of needs. How to find the positioning service function, and how to quickly combine multiple services to meet the demand of new applications were the two main problems to be solved. The main research work in this paper was to apply semantic technology to deal with four basic problems in the smart government system:a precise semantic description of business services to provide accurate identification service; Using reasoning technology to accurately find and locate services; Using semantic information to assist service composition; Using semantic technology service to realize replacement and invocation.According to the characteristics of the e-government and semantic technology development, this paper presented a semantic model of e-government Web services, and presented the ontology definition. In order to effectively organize and manage a large number of services, this paper studied the establishment of a smart government business classification model and ontology. At the same time, based on the needs of describing services function semantics, according to the relevant national standards to develop smart government information resources classification ontology and public data elements onotology, the paper was to provide the basis for annotating service10semantics and use the general mechanism of OWL DL to describe the PE semantics of the service function. In order to guarantee the service with accurate and enough semantic information, this paper designed the semantic register service interface. Through what mentioned above, this paper solved the problem of the service semantic modeling, and created foundation for the solution of service discovery and combination.Service discovery based on semantic reasoning could accurately locate service, but due to the inherent complexity of reasoning technology, in the face of the great smart government service space, the efficiency of reasoning was unacceptable. This paper introduced a relational query technology, and used relational database to store management service registration information, which first conducted semantic retrieval based on keywords for fast shrinking of service quantity, and then conducted semantic reasoning to accurately locate the candidate services in this relatively small space. Because of the efficient retrieval of relational database, this method could be a very good balance between positioning accuracy and efficiency of service discovery.For another key issue in the business collaboraion of smart government-service composition, this paper took the search algorithm based on "AND/OR graph", namely the combination problem was converted to AND/OR graph solution. Because there existed high time complexity in AND/OR graph search algorithms, it had to reduce the size of problem searching space to improve the practicability of the method. In constructing the service composition problem state diagram, it did not directly use the registration service to generate, but to extract the abstract service to construct according to the business categories, which could effectively reduce the searching size of state diagram and maintain stability. With the problem that AND/OR graph could be only used to describe the IO semantics of the service but could not be used to depict PE semantics, service composability model was introduced, which comprehensively studied service grammatical and semantic composability among services, and by calculating the candidate combination degree, which could guarantee the semantic consistency between the component services. Another purpose of composability model was that as an important reference indicator replacing the service selection, it could examine the success rate of the combination of service execution after replacement.This paper studied the key problems in the process of smart government system integration and business collaboration. In view of the difficulties of using the SOA and Web services technologies to solve these problems, the author of this paper put forward an effective solution and applied these studies in the planning and design of cloud computing platform of Wuhan smart government.更多还原