【作者】 刘晓慧；

【导师】 吴信才；

【作者基本信息】 中国地质大学 ， 地图制图学与地理信息工程， 2014， 博士

【摘要】 我国地貌形态多样、地质构造复杂,地壳断裂活动普遍,加之人类活动影响,是世界上地质灾害最严重、受威胁人口最多的国家之一,崩塌、滑坡、泥石流、地面塌陷等受地质作用影响的地质灾害隐患多、分布广。受极端气象、地震、工程建设加剧等因素的影响,近年来我国地质灾害频发群发,人民群众生命财产遭受了严重的损失,对社会经济可持续发展产生了重大影响。随着社会经济的快速发展,大规模基础设施建设与水资源和矿产资源的不断开发对地质环境的影响仍在持续,而且本世纪前期气候变化和地震均趋于活跃期,因此从引发地质灾害的内外因来看,我国地质灾害的发生还将呈现出灾害种类多、分布地域广、发生频率高的趋势。在地质灾害应急管理与决策过程中通过采取有效的信息化手段,提高对突发地质灾害事件的快速反应能力,从而尽量降低灾害造成的损失,具有重要的理论意义和实用价值。地质灾害应急行动过程通常涉及跨区域、跨领域的多个部门,需要部门之间的相互协调与合作,而信息资源的共享与协同服务是应急决策支持系统的基础支撑。地质灾害应急管理的信息化水平在不断提升,空间信息技术也已经深入地质灾害应急管理与决策支持的各方面。但是,作为应急决策的重要依据和知识载体,由各级部门制定的地质灾害应急预案目前仍然主要以自然语言描述的静态文本形式存在,在地质灾害应急响应与决策过程中存在以下问题。(1)应急响应与决策过程需要多个部门的信息共享与协同服务,而目前很多决策所需信息是在灾害事件发生后通过人工方式向有关部门询问获得,不仅浪费了大量的宝贵时间,而且难以实现快速的信息融合与决策支持分析。(2)静态文本预案是非结构化文档,为应急决策过程提供的信息有限,而且无法实现预案知识的共享和重用,难以有效支持应急决策方案的制定。(3)预案是根据处置突发事件的经验预先制定的处置方法,而灾害的发生通常是不可预测的,在应急处置过程中出现应急资源不充足、应急力量变动、信息不确定等情况时,基于文本预案的决策形式具有很大的局限性。本文的研究旨在为突发地质灾害应急决策提供高效、智能、直观的知识支撑,辅助决策者迅速做出决策,目标是通过有效的技术手段在正确的时机以适当的形式为应急管理或决策人员提供合适的知识。首先,在出现地质灾害灾情或险情时,面对复杂的应急态势,能够快速获取决策所需信息并进行智能化的分析,有针对性的为决策者提供可用的决策支持信息。其次,利用GIS的信息分析与展示优势,为决策者提供友好的人机交互界面与直观可视化的信息,确保决策过程高效、有序进行,但是计算机决不能代替人做出决策。本文在查阅与调研国内外相关文献与资料的基础上,围绕地质灾害应急管理与决策过程中信息融合与协同服务不足、基于文本预案决策形式不足等问题,通过研究应急管理体系与机制、应急决策知识以及决策体系,设计了面向服务的地质灾害智能化应急决策支持体系。重点研究了基于本体建模理论的地质灾害应急决策知识表达方法,对基于预案的地质灾害应急决策所需知识作了形式化表达。并提出了事件态势驱动的地质灾害智能化应急决策支持模型,以实时应急态势为驱动,将知识本体推理技术与空间辅助决策支持技术相结合,对地质灾害应急决策所需基础空间数据、专题数据以及应急决策知识进行无缝融合,为基于应急预案的智能化决策支持提供一种可行的方案。主要开展了以下几个方面的研究：(1)分析与研究了地质灾害应急管理与决策体系。首先明确了地质灾害应急管理的原则与目标以及应急管理的对象范围,对地质灾害应急管理组织体系进行了研究,并从应急管理生命周期与工作阶段两个角度对应急管理机制做了详细的分析,将应急管理生命周期中的预防、准备、响应和恢复的应急管理内容与日常管理、汛期预警、临灾响应及灾后恢复四个阶段的工作进行映射。然后进一步对与地质灾害应急决策紧密关联的地质灾害预防与预警、灾险情速报与处理、应急响应与处置等三个应急业务过程作了详细的分析。在分析应急管理体系与机制的基础上,对地质灾害应急决策的特征及体系作了详细分析。地质灾害应急决策需要大量知识的支持,因此对地质灾害应急决策所需知识的组成、结构与特征进行详细分析,指出应急预案中由人力知识与技能知识构成的隐性知识是决策支持的重要依据,提出利用本体建模理论对以预案为基础的应急决策知识进行形式化表达的思路。最后提出了面向服务的地质灾害智能化应急决策支持体系,明确了本文的研究重点是：在该决策支持体系下,将应急决策知识本体与空间信息技术相结合,为应急响应与处置过程提供智能决策支持。(2)利用本体建模理论构建了基于预案的地质灾害应急决策知识本体模型PEOMG。首先对本体建模的相关理论、技术和方法基础进行研究和了解,包括本体的定义和分类、本体描述语言、语义网规则语言以及本体构建方法论,并且对本体在信息共享、信息检索、知识服务及决策支持等应用场景的功能做了分析。通过分析国家级、省市级、县级等多个不同级别的地质灾害应急预案文本,提取以预案为基础的应急决策知识建模要素,包括应急管理实体、应急行动过程以及应急业务规则。以本体建模的相关理论、技术和方法为基础,以ABC事件本体模型为顶层本体,通过重用其中的概念和属性并结合地质灾害应急管理的特点对概念及属性进行扩充,基于网络本体语言OWL对应急管理实体和应急行动过程进行知识建模。在应急管理实体和应急行动过程本体定义的概念和属性的基础上,利用语义网规则语言SWRL对应急业务规则作形式化表达,为基于知识本体推理的地质灾害智能化应急决策支持奠定基础。最后以某县级地质灾害应急预案为对象进行知识本体建模,验证了PEOMG的一致性和有效性。(3)提出了事件态势驱动的地质灾害智能化应急决策支持模型,重点对该模型中的应急态势感知模型、决策知识服务模型以及决策推理与分析进行了研究。态势感知是应急决策过程中的重要环节,是决策者与应急环境之间的桥梁,为决策方案的制定提供实时应急态势信息。首先提出了地质灾害应急态势感知三层模型,包括感知应急情境中的元素、理解当前的应急态势以及预测将来的应急状态,以实现应急决策事件态势的实时感知。然后从事件态势信息流的角度研究了应急决策知识服务的过程。地质灾害预警信息流与速报信息流是应急态势感知的主要途径,对于应急决策知识的动态获取具有重要的驱动作用。通过建立决策知识本体服务目录提供对应急决策知识的服务描述与获取方式。通过解析预警信息流或速报信息流中的特征元素,获得事件信息、态势信息以及时空信息,与知识本体概念进行动态映射,提取所需知识项。在此基础上,以事件实时态势信息为驱动,研究了基于应急决策知识本体与空间信息技术的地质灾害应急决策推理与分析方法。一方面,通过知识本体推理使决策者在宏观上明确应急响应级别、应急响应与处置的任务,根据预案中的人力知识和技能知识指导决策者组织相应的机构开展应急行动；另一方面,在时间紧迫、资源紧缺的情况下,考虑到预案知识不能为具体的指挥调度过程提供直观的依据,将空间辅助决策分析融入预案智能执行的过程,在微观上以直观可视化的形式为指挥调度与态势标绘提供技术支持,形成优势互补、宏观和微观结合的智能化决策支持技术体系。(4)设计并实现了突发地质灾害应急预案智能化原型系统,对突发地质灾害智能应急决策支持模型的理论方法进行验证。通过梳理应急预案执行流程,对区域地质灾害预警响应和单体地质灾害应急响应的流程进行分析与设计。基于面向服务的架构思想对系统作了体系结构设计,并以预案执行的流程为依据对系统功能进行合理的组织,实现了基于预案的应急管理与决策支持过程的数字化、智能化与可视化,验证了基于知识本体推理与空间辅助决策技术进行智能化应急决策支持的理论方法和技术体系的可行性。通过论文研究,主要的创新和特色在于：(1)在综合各级地质灾害应急预案内容的基础上,通过分析基于预案的地质灾害应急决策所需知识和技术框架,基于网络本体语言OWL和语义网规则语言SWRL建立了基于预案的地质灾害应急决策知识本体模型PEOMG。通过PEOMG可以在人与人之间以及人与计算机之间实现应急决策知识共享、重用与互操作,为地质灾害应急决策过程提供了一种语义协调机制。以PEOMG为模型基础,利用本体推理技术为预案流程的智能化执行提供了一种方法。(2)提出了事件态势驱动的地质灾害智能化应急决策支持模型。以事件态势信息流为驱动,通过Jena与Pellet组合进行决策知识推理,使决策者在宏观上可以明确应急响应级别、应急响应与处置的任务,同时将空间辅助决策支持技术融入知识本体推理过程,在微观上以直观可视化的形式为指挥调度与实时标绘提供支持,从而形成优势互补、宏观和微观结合的智能化决策支持技术体系。更多还原

【Abstract】 Crustal fracture activities is universal because of China’s diversity morphology and complex geology, and human activities are very frequent, thus China is one of the countries in which there are the most serious geological disasters and the most threatened population in the world. Many geological disasters related to geological process, such as collapse, landslide, mudslide and ground sink, is widely distributed in China. Because of the impact of extreme weather events, earthquake and construction, geohazards in recent years have occurred frequently which affected sustainable development of social and economic. With the rapid development of social and economic, infrastructure construction and development of water resources and mineral resources continues to influence geological environment. Meanwhile climate changes and earthquakes tend to be more active in the early years of this century. From the inside and outside cause of geohazards, the occurrence of geological disasters in China will present a trend of variety, wide distribution and high frequency. It is of practical significance to improve the geohazard emergency response ability so as to minimize the loss caused by disasters through making full use of information technology in the process of emergency management and decision-making.Emergency response process of geohazards usually involves several cross-regional and cross-domain departments, thus coordination and cooperation between departments is necessary, and information resource sharing and collaborative service is the basis of the emergency decision support system. With the improving of the geohazard emergency management informatization level, spatial information technology has also been applied to many aspects of the geohazard emergency management and decision support. But as the main basis of decision-making, emergency plans of geohazards made by all levels of departments are still mainly existed as static text in the form of natural language at present. So much shortage is still existed in the process of emergency decision-making.(1)Process of geohazard emergency decision-making requires information sharing and services collaboration among several departments. Now much information needed during decision-making is got through asking to the relevant departments by artificial way after geological disaster happened, which wastes a lot of precious time, and it is difficult to quickly realize information fusion and decision support analysis.(2)Knowledge of static text plans is unstructured, which cannot provide sufficient information for emergency decision-making process or implement sharing and reuse of emergency plan knowledge. So it is difficult to support decision-making effectively.(3)Emergency plans are established based on the experience of emergency disposal, but the disaster is usually unpredictable. Thus it has great limitations to cope with unexpected situations based on text plans. For example, emergency resources are not enough, emergency power is changed, and information is uncertain.The study of this thesis aims to help decision-makers make decisions quickly by providing efficient, intelligent and intuitive knowledge. The goal of this study is to provide effective knowledge for people of emergency management or decision-maker through effective technology in the appropriate form and at the right time. Firstly, when a geological disaster happened, facing the complicated emergency situation, it is capable of providing decision support information available for decision-makers through accessing to information rapidly and intelligentized analysis. Secondly, taking advantages of information analysis and showing of GIS, it is possible to ensure that the decision can be made efficiently in order by providing friendly man-machine interface and visual information, but the computer can never make decision replace decision-makers.In order to solve the above problems, this paper designed service-oriented intelligent decision support system for emergency of geohazards on the basis of reading relevant literature at home and abroad. Plan-based emergency ontology model of geohazard was constructed by using of the ontology modeling theory, on the basis of which knowledge of emergency plans of geohazards was represented formally. And then, making emergency situation information as driver, and combining with the ontology reasoning technology and space decision support technology, this research achieved seamless integration of spatial data, thematic data and knowledge of emergency plans which is required in emergency decision of geological disaster. Through these studies, this paper provided a viable solution for intelligent decision support of geohazards based on emergency plans. The key contributions of this study are as follows:(1)Emergency management and decision system of geohazards is analyzed and researched. First of all, the principle, goal and target of geohazards emergency management were understood, emergency management organization system was studied, and emergency management mechanism was analyzed from aspects of lifecycle and work phase in detail. Lifecycle of emergency management contains four steps of prevention, preparedness, response and recovery, and the main contents of the four work phases were mapped to the4-step lifecycle. Then three emergency businesses which were closely related to emergency decision-making of geohazards were analyzed further in detail, including pre-warning, disaster reporting and emergency response. Based on the analysis of emergency management system and mechanism, the characteristics and system of emergency decision-making were analyzed in detail. Much knowledge is needed by emergency decision of geohazards, so the composition, structure and characteristics of emergency decision knowledge of geohazards were analyzed. And it was pointed out that the tacit knowledge composed of know-who and know-how was important for emergency decision, and ontology modeling theory could be used of formalizing emergency decision-making knowledge based on the emergency plan. Finally, service-oriented intelligent decision support system for emergency of geohazards was proposed. And the focus in this system is to provide intelligent decision support for geohazards emergency response by combining knowledge of emergency plans and spatial information technology.(2)Plan-based emergency ontology model of geohazard(PEOMG) was constructed through ontology modeling theory. Firstly, theory, technology and method of ontology modeling were researched, including definition and classification of ontology, ontology description language, semantic web rule language and ontology construction methodology. And function of ontology in scenario of information sharing, information retrieval, knowledge service and decision support was analyzed. Then According to the analysis of geohazards emergency plans text in a number of different levels, three elements of emergency decision-making knowledge modeling based on plan were extracted, including emergency management entities, emergency actions and emergency business rules. Based on ontology modeling theory and OWL, making ABC event ontology model as the top ontology, the ontology model of emergency management entities and emergency actions were constructed by reusing the concepts and properties of ABC and extending new concepts and properties of geohazard emergency management. On the basis of the ontology model of emergency management entities and emergency actions, emergency business rules of geohazards were formalized by using SWRL. And these rules are the foundation of intelligent decision support for emergency of geohazard through the knowledge-based ontology reasoning. Finally, an application study was carried out for modelling a county-level emergency plan by using PEOMG, the consistency and effectiveness of which was validated.(3) Event-driven intelligent decision support model for emergency of geohazard was proposed, and the focus of the research was emergency situation awareness, decision-making knowledge service and reasoning and analysis of decision. Emergency situation awareness which provides real-time situation, takes an important part in decision-making process as a bridge between decision-makers and emergency environment. Firstly, the3-level model of geohazard emergency situation awareness was proposed in order to get real-time event situation for emergency decision-making. The three levels are the perception of the elements in the emergency environment, the comprehension of their meaning and the projection of their status in the near future. Then knowledge service of the emergency decision-making based on event situational information flow was studied. The main way for emergency situation awareness is the information flow of pre-warning and reporting, which has an important role for dynamic access to knowledge of emergency decision. The service descriptions for knowledge of emergency plans were provided through establishing services directory of decision-making knowledge. And the real-time information of geohazards, such as event information, situational information and spatio-temporal information, was got by parsing the information flow of pre-warning or reporting and extracting the characteristic elements. The necessary knowledge items for decision-making can be extracted through mapping between the real-time information of gehazards and concepts of emergency plan ontology model. Finally, Reasoning and analysis based on ontology and spatial information technology for geohazard emergency decision-making were studied, which was driven by the real-time situational information. On the one hand, decision-makers can get information about the grade and task of emergency response at the macro level by ontology reasoning, and to guide the emergency action according to the knowledge of human and skill. On the other hand, in order to provide intuitive basis for the processes of command and scheduling under constraints of time and resource, this study integrated spatial decision support analysis into the process of knowledge reasoning of emergency plan, which can provide intuitive information support for the real-time dispatching and plotting at the micro level in a visual form. So an intelligent decision support system was provided for emergency of geohazard by combining with advantages of these two aspects.(4)An prototype system of intelligent emergency plan for geohazard was designed and implemented in order to validate the theory of intelligent decision support model for emergency of geohazard. At first, emergency response process of regional geohazard pre-warning and monomer geological disaster were analyzed and designed by teasing out the implementation procedure of emergency plan. Then, an service-oriented system architecture was designed, and the system functions was organized based on the process of emergency plan, which achieved digital, intelligent and visual emergency management and decision support based on emergency plans. The theory and technology of intelligent decision support which combined knowledge ontology of emergency plan with spatial information technique, was validated being feasible.The main in novations and features of this study are as follows:(1)In the basis of considering geohazard emergency plans at all levels, the knowledge and technical framework required for geohazard emergency decision based on plan was analyzed, and then plan-based emergency ontology model of geohazard(PEOMG) was constructed by using OWL and SWRL. Sharing, reuse and interoperability of knowledge of emergency plan between people or between people and computers can be achieved through PEOMG. which provides a semantic coordination mechanism for emergency decision of geohazards. And ontology reasoning based on PEOMG provides a way for intelligent execution of emergency plans.(2)Event-driven intelligent decision support model for emergency of geohazard was proposed. Decision-makers can get information about the grade and task of emergency response at the macro level by knowledge reasoning based on ontology by using Jena and Pellet, which is driven by information flow of event situation. At the same time, the research integrated spatial decision support analysis into the process of knowledge reasoning of emergency plan, which can provide intuitive information for the real-time dispatching and plotting at the micro level in a visual form. So an intelligent decision support system was provided for emergency of geohazard by combining with advantages of knowledge reasoning and spatial information technology.更多还原