【作者】 王一军；

【导师】 罗大庸；

【作者基本信息】 中南大学 ， 控制科学与工程， 2009， 博士

【摘要】 随着地理信息系统、计算机技术、智能控制技术的发展以及环境信息管理系统的普及,环境决策支持系统的研究和应用成为目前的研究热点之一;环境决策支持系统能为环境决策者提供更好的计算机辅助决策支持手段,对环境规划、管理和保护的科学化和现代化,环境事件的应急快速处理以及提高专家在环境保护决策中的作用都具有十分重要的意义。环境决策支持系统是包括环境规划子系统、环境评价子系统、突发事故应急处理子系统、环境预警子系统、生态环境子系统等子系统在内的一个涉及多个学科的复杂大系统,学位论文主要针对环境决策支持系统的基础理论和实现方法以及各子系统的有关关键技术展开研究,主要包括环境决策支持系统的实现框架、重点污染企业选址算法、环境评价方法以及应急最优路径选择算法等研究,并通过在环境决策支持仿真系统中的编程实现,验证算法和技术的有效性和可行性,以期能提高环境决策支持系统的决策效果,促进环境决策支持系统的发展,更好地为环境事业服务。学位论文的主要工作和成果可以概括为以下五个方面:1、在分析环境系统的数据流和业务流以及管理决策过程特点的基础上,对环境决策支持系统的技术需求和功能进行了总结归纳,提出了一种实现环境决策支持系统的实现框架,并对环境决策支持系统的实现平台进行了初步设计。2、重点研究了基于证据理论的环境评价方法,针对目前证据理论在合成冲突证据时存在的合成效率不高,计算复杂度较大等缺陷,提出了改进的基于证据可信度分级的证据理论,并具体提出了模糊分级和线性分级两种证据分级方法以及算法的快速计算定理并加以了证明,最后通过典型案例实验证明了提出的方法可以更高效地处理高度冲突的证据;在此基础上,探讨了群决策系统在环境评价的应用,提出了两级证据融合的环境评价应用模型,并通过水质评价的应用实例证明了提出的环境评价方法的有效性和可行性。3、针对目前的环境规划子系统的污染企业选址算法中存在的图像信息利用较少,选址精度不够以及搜索时间过长等问题,利用遥感图像处理技术和改进的粒子群优化算法,提出了一种结合遥感图像处理技术的企业选址算法,可以充分利用GIS的空间和非空间信息,考虑的因素更为全面,极大地避免了人为因素的影响,是对选址决策支持系统的改进。4、在应急子系统中的最优路径选择算法研究中,针对目前应急路径选择算法存在的实用性和实时性差的问题,提出了一种新的最优路径搜索算法,算法首先采用分层建模技术,建立优化的层次化路网,然后在高层路网上,采用结合道路状况的启发式A*优化路径搜索算法进行最优路径的搜索,并通过在环境决策支持仿真系统的应急子系统中的应用验证了提出方法的有效性;同时对应急系统中普遍存在的目标点动态变化情况下的应急最优路径展开研究,提出一种结合直线型椭圆排除法的动态最优路径算法,并进行了仿真试验。5、探讨了环境决策支持仿真系统的实现问题,对环境决策支持系统涉及的上述关键技术进行了编程实现和考核,为环境决策支持系统的具体实施和应用打下了基础。学位论文最后一章在总结全文工作的基础上,对环境决策支持系统的发展方向进行了展望,并提出了后续研究工作的思路,对进一步的研究具有一定的指导意义,最后,就学位论文中开展的研究工作在实际场合中的应用进行了展望。更多还原

【Abstract】 With the development of Geographic Information System, Computer Science and Technology, Intelligent Control Technology and the popular of Management Information System in environmental system, the study of Environmental Decision and Support System (EDSS) is becoming the key and hot research field. EDSS can provide the deciders in environmental system with more useful and instant support information in the process of making decisions, and make it possible to make more effective and scientific decisions and modernization and standardization of the management of environmental protection system. At the same time, it can help develop the role of the exports in making decisions and dealing with the environmental accidents more instant and effective.EDSS is a sophistical large system including environmental planning subsystem, environmental evaluation subsystem, environment pollution emergency subsystem, environment alarm subsystem and ecological environment subsystem, which is involved in many scientific subjects. So the main task of the dissertation is to study the basic theory and application method in EDSS and the key technology in related subsystems including the study of the frame architecture of EDSS, location selection of the polluting factory, expert group decision system based on modified D_S evidence theory applied in environmental evaluation, the optimization route in emergency subsystem and their applications in emulation system whose aims are to testify all developed methods in dissertation. The study of the dissertation can help obtain more effective decision results, contribute to the popular of EDSS and provide better services to environmental management. The main contributions can be summarized in the following five parts.1. After analyzing the data flow, the business operation flow and the characteristics of the environmental system, the technique demands and functions are summed up in this dissertation. Then an application frame is presented and the platform of EDSS is designed. 2. In order to overcome deficiencies such as difficult to deal with conflict evidence and the large time cost in combing the evidences, a modified evidence combination rule is developed through rating the evidences according to their belief functions. Meantime, two different rating ways to rate the evidences are also presented based on the fuzzy sets or linearization the evidences’ distances. Meanwhile, a quick calculation method is developed and testified. The typical and real instance shows this combination rule can combine highly conflicting evidences more quick and satisfactory. Moreover, in designing the environmental evaluation subsystem based on expert group decision system, a new model including two-layers evidence combination processes is developed, and the application of it in a water evaluation system testifies the feasibility and validity of our developed method.3. In establishing the location selection mathematical model in spatial decision support system, the conventional methods have deficiencies such as large time cost in searching process and difficult to locate the exact position because of less considering the images information. To improve the mathematical model, based on the remote image processing and modified particle swarm optimizer, a new method is introduced to the research of spatial decision support system for location selection of factory. As a remarkable improvement on the conventional methods, the mathematical model can be easily established by fully utilizing spatial information and attribute information in GIS.4. In studying the algorithms to select the optimization path in emergency subsystem, a modified optimization path finding algorithm with higher speed and better results than old ones is presented. In which, by the use of the hierarchical spatial modeling method, the optimization hierarchical traffic network is firstly set up. Then the start part and the end part of the optimization path are found through querying the pre-stored databases. At the same time, by applying a modified heuristic A* shortest path search algorithm which considering the traffic states of the road in searching path at the dynamic traffic environment, the main part of the optimization path is also obtained. And the real application in real-time traffic system shows the effective of our developed method. Finally, based on the study of the shortest path algorithms when the destination is changing, a modified shortest path algorithm with high speed and good results is presented. And the simulation using data from the real traffic system shows the developed method can get good balance between the time cost and the exact results.5. The emulation system of EDSS is designed, in which all new developed methods are applied using software program and testified. The results show that the developed methods can be used in the future real EDSS system.In the last chapter, after summarizing the contents of this dissertation, the further research issues are presented in detail, which will guide the future research direction in this field. Finally, the future applications of algorithms are also discussed in this dissertation.更多还原