【作者】 谢斌；

【导师】 徐世浙； 张登荣；

【作者基本信息】 浙江大学 ， 地球探测与信息技术， 2010， 博士

【摘要】 网络化地理空间信息应用正深刻地影响着地理信息的服务方式,国家测绘局已经启动了“地理信息公共服务平台”等计划,以期全面提升地理信息服务的能力和水平。当前,由于网络环境的复杂性,基于网络的地理空间数据共享和互操作还普遍存在着服务语义模糊、服务组合不够动态灵活、服务异常难以处置等问题。另外,在分布式协同服务资源的主动发现、灵活管理等方面也尚无成熟可靠的方法与产品,这都限制了地理空间数据资源的共享和互操作适用性。本文结合服务计算、自治计算、策略管理以及OGC空间信息服务规范,研究描述空间信息服务协同的共享本体,建立“空间E机构-空间智能体”两层空间信息协同服务模型。通过研究用于服务协同环境的协同控制准则和应用管理策略集,探讨从基于空间E机构的抽象服务流程描述,到具体空间智能体执行实体之间的宏观协同调控方法。研究建立空间E机构的调控模型和空间智能体的具体执行机制,并构建松耦合的两层次服务协同体系,最终实现空间信息服务协同的按需自动建立和自主演化,满足分布式地理空间数据共享和互操作服务系统动态、灵活、快捷、有效、安全的应用需求。本文的具体研究内容如下：(1)研究空间信息服务分布式协同体系框架,提出“空间E机构-空间智能体”两层空间信息协同服务模型。建立空间信息服务协同的共享本体、协同推理规则以提供语义基础。基于共享本体,研究服务协同管理策略的表示方法。(2)研究空间E机构协同调控模型,建立以空间E机构方式描述的空间信息服务协同业务规范模板。研究基于软件Agent的空间智能体的组成结构及其内部组件协作方式,以及外部空间信息服务的配置方法和策略驱动的自主管理机制。(3)基于“空间E机构-空间智能体”协同模型,研究空间E机构宏观业务目标与空间智能体个体行为之间的连接机制,通过策略驱动的自主管理,去宏观调控空间智能体的协同参与行为。制定服务协同环境的协同控制准则和应用管理策略,建立空间E机构的调控模型和空间智能体的具体执行机制,形成松耦合的两层次服务协同体系。(4)设计基于空间E机构、空间智能体的空间信息分布式协同服务架构原型,并给出一个空间信息服务分布式协同应用实例分析,验证策略驱动的空间信息服务分布式协同基础设施的有效性。研究表明,本文提出的“空间E机构-空间智能体”两层空间信息协同服务模型,结合了语义本体、Agent、策略管理以及OGC空间信息服务技术,提供了具备一定智能化、自动化的空间信息服务分布式协同处理能力。基于领域共享本体,空间E机构能根据地学对象的描述建立地理空间模型,并表达抽象的空间信息服务协同流程；基于策略驱动的自主管理方式,空间智能体能遵循空间E机构的流程编排,按需组建空间信息服务链,并进行服务协同执行过程的异常监控和处理。本文构建的基于空间E机构调控和多空间智能体协同的松耦合两层次服务协同体系,既提升了空间信息服务协同的智能化表达和灵活管理能力,同时又能减少人工干预。对于促进地理信息公共服务平台在地理信息服务资源的松耦集成、动态装配及管理调度等方面,提供了一种新的思路和实现途径。更多还原

【Abstract】 Service mode of geographic information are effected by developed applications of geospatial information network significantly. Therefore, recently the national surveying and mapping bureau of China has granted research projects such as "Geographic information public service platform" etc. to enhance the capacity and level of geographic information service. Nowadays, due to the complexity of the web environment, techniques of Web-based geospatial data sharing and interoperation services still facing several common problems e.g. service semantic ambiguity, lack of dynamic and flexible service composition, service exceptions are difficult to dispose etc. Furthermore, regarding to the initiative discovery and flexible management of distributed collaborative service resources, no reliable and well developed methods or products are available, which impedes the applicability of geospatial data/resources sharing and interoperation services.A two-level geospatial information service collaborative model namely Geospatial E-Institution& Geospatial Agent Model (GEI&GAM) is proposed in this thesis by integrating achievements from four related techniques i.e. services computing, autonomic computing, policy management, and OGC geospatial information service specification. After studying the cooperative control guidelines and application management policy set for the service collaborative environment, methods of macroscopic regulation and control are investigated. These methods cover from abstract service process description of geospatial e-institution to the service chain implementation of specific geospatial agent. Then, control model of geospatial e-institution and specific implementation mechanisms of geospatial agent is established to form the loosely coupled two-level service coordination system. Finally, on-demand automatic creating and independent evolving for geospatial information services coordination are achieved to satisfy the dynamic, flexible, fast, effective and safe applications of geospatial data sharing and interoperation. The main research works are as follows:1. A two-level geospatial information service collaborative model GEI&GAM is proposed after studied distributed collaborative framework. Shared ontology of collaborative geospatial information services and cooperative reasoning rules to provide semantic foundation are established. Based on shared ontology, representation method of services collaborative management policy is also introduced.2. Research on cooperative control model of geospatial e-institution is conducted; and business specification template of collaborative geospatial information services described in the way of the geospatial e-institution is established afterwards. Besides, policy-driven self-management mechanism of geospatial agent is studied based on techniques from software agent, composite structure and its internal components collaborative method and external configuration of geospatial information service.3. According to the collaborative service model of GEI&GAM, link mechanism between macroscopical business objectives of geospatial e-institution and individual behavior of geospatial agent is studied. By means of policy-driven self-management, collaborative acts of geospatial agent is macroscopical regulated. Followed the development of cooperative control guidelines and application management policy set of the service collaborative environment, control model of geospatial e-institution and specific mechanisms for implementation of geospatial agent are established, and loosely coupled system of two-level service coordination is formed.4. Distributed geospatial information services collaborative architecture prototype based on GEI&GAM is designed, and an application of the prototype is given to verify the effectiveness of the infrastructure.The results show this proposed GEI&GAM collaborative service model, which combines techniques from semantic ontology, Agent, policy management and OGC geospatial information services is able to provide intelligent, automated distributed collaborative processing capability for geospatial information. Using shared ontology, geospatial e-institution is able to establish geospatial model and abstract geospatial information collaborative service process; besides using policy-driven self-management, geospatial agent is capable of following the process of geospatial e-institution to form geospatial information service chain on demand, and monitoring/handling exception of service coordination. This loosely coupled two-level system promotes intelligent expression and flexible management of geospatial information service, and lessens human intervention. It also provides a new way for geographic information public service platform in the loose coupling integration, dynamic assembly and management of service resources.更多还原