基于随机网络的非常规突发事件情景推演模型及其应用研究

发布时间：2018-05-13 13:00

本文选题：非常突发事件 + 情景推演　；参考：《南京航空航天大学》2011年博士论文

【摘要】：非常规突发事件往往涉及大量人员生命安全，经济与社会代价巨大，其研究也无法进行物理实验。此外，非常规突发事件的发生和发展还具有不可预测性、演变复杂性及影响广泛性等内在特点。因此对非常规突发事件的研究急需从传统的“预测-应对”型转变到“情景—应对”型，更需要将非常规突发事件的定性认知转变为定量计算的方法路径。论文以非常规突发事件应急管理重大研究计划培育项目“基于GERTS仿真网络的非常规突发事件情景推演与测算问题研究”为依托，借鉴已有的非常规突发事件情景应对研究成果，，采用定量研究与实例分析相结合的方法，讨论了非常规突发事件情景推演网络的构建方法，研究了非常规突发事件情景推演中事件间的耦合作用模式、应急对冲资源的优化配置、情景推演仿真系统的搭建等问题，以期为非常规突发事件的应急辅助决策提供理论参考。论文的主要研究成果有：（1）搭建了基于随机网络的非常规突发事件情景推演网络结构。在总结分析非常突发事件情景推演特征的基础上，提出了基于随机网络的情景推演网络的描述原理，设计了非常规突发事件情景推演广义随机网络的基本结构，研究了非常规突发事件广义随机网络向随机网络的转化方法，并给出了非常规突发事件情景推演随机网络的求解方法。进一步地，研究了非常规突发事件情景推演随机仿真网络模型，并扩展了仿真网络仿真节点的结构和功能。该研究为非常规突发事件的情景推演提供了实验平台。 (2)提出并构建了基于随机网络的非常规突发事件情景推演耦合模型。通过对非常规突发事件间作用形式的分析，创新地提出了非常规突发事件情景推演共力耦合、互力耦合与驱力耦合概念；结合随机网络的特征，设计了三类耦合节点的逻辑结构，构建了三类耦合结构的情境演化模型，并给出了求解三类耦合模型参数的算法。该研究提供了分析非常规突发事件耦合的定量方法。 (3)构建了非常规突发事件应急资源配置情景推演模型。针对应急抢险过程中资源需求与供应不匹配这一问题，通过定义非常规突发事件情景推演随机网络的基本单元，建立一种综合考虑灾害自身演化过程及外界作用的相互关系的非常规突发事件情景演化随机网络，设计了求解不同资源配置情况下突发事件状态转移概率的极大熵模型；在此基础上，研究了应急资源的动态配置的情景推演模型。该研究提供了应急资源配置定性与定量结合的分析框架与工具。 (4)建立了基于不确定参数的非常规突发事件情景推演随机网络模型及其求解算法。针对非常规突发事件的情景推演具有的不确定性，提出了带有不确定参数信息的非常规突发事件情景推演随机网络模型，研究了不确定变量随机网络的矩母函数的构造，并对其具有的性质做了详细的探讨；结合信号流图原理和不确定性理论研究成果，研究了不确定信息非常规突发事件情景推演的随机网络仿真算法。 (5)搭建了基于随机网络的地震生命工程情境推演平台系统。以地震灾害中人员伤亡情景推演为研究背景，在考虑该事件情景推演中存在的不确定性，将构造的情景推演仿真随机网络方法、情景推演过程中的耦合模型、应急资源的配置模型等应用于地震生命工程情景推演；在Microsoft Visual Studio2005开发环境下，建立上述相关模型的计算机可执行语言，开发出地震生命工程情景推演平台系统。该平台系统为近一步的研究和应用提供了可实际操作的工具。
[Abstract]:Non - conventional emergencies often involve a large number of people ' s lives , safety , economic and social costs , and their research is impossible to carry out physical experiments . In addition , the occurrence and development of non - conventional emergencies has the inherent characteristics of unpredictable , evolutionary complexity and wide impact . Therefore , it is urgent to change the traditional " prediction - response " type from traditional " prediction - response " type to " situation - response " type , and to change the qualitative cognition of non - conventional emergencies into quantitative calculation method path .

Based on the research results of non - conventional emergency scenarios based on GERTS simulation network , this paper discusses the construction method of non - conventional emergency scene deduction network by using the method of combining quantitative research and case analysis , and discusses the problems of coupling action mode , optimal allocation of emergency hedging resources , setting up of scene deduction simulation system and so on , with a view to providing theoretical reference for emergency auxiliary decision - making of non - conventional emergencies .

( 1 ) The network structure of non - conventional emergency scenarios based on random network is established . Based on the summarization and analysis of the characteristics of critical incident scenarios , the paper puts forward the basic structure of the generalized random network based on random network , studies the transformation method of generalized random network to random network , and gives out the structure and function of simulation network simulation nodes . The research provides an experimental platform for the scenario deduction of unconventional emergencies .

( 2 ) The coupling model of non - conventional emergency scenarios based on random network is proposed and constructed . Through the analysis of the interaction between unconventional emergencies , the concept of co - force coupling , mutual force coupling and displacement coupling is proposed .
Based on the characteristics of random network , the logic structure of three kinds of coupling nodes is designed , the situation evolution model of three kinds of coupling structures is constructed , and the algorithm for solving three kinds of coupling model parameters is given .

( 3 ) A scenario - based model for emergency resource allocation of non - conventional emergencies is constructed . In the light of the problem of mismatch between resource demand and supply during emergency rescue , a stochastic network of non - conventional emergency scenarios is established by defining non - conventional emergency scenarios , and a great entropy model is designed to solve the emergency state transition probability under different resource configuration conditions .
On this basis , the dynamic configuration of emergency resources is studied , and the framework and tools for qualitative and quantitative analysis of emergency resources are provided .

( 4 ) The random network model and its solution algorithm based on uncertain parameters are established . The random network model of non - conventional emergency scenarios with uncertain parameter information is presented , and the structure of the moment mother function of uncertain variable random network is studied , and the properties of the random network model are discussed in detail .
Based on the principle of signal flow graph and the research results of uncertainty theory , a random network simulation algorithm for uncertain information unconventional emergency scenarios is studied .

( 5 ) The platform system of earthquake life engineering based on random network is built . In this paper , based on the background of the research background of the scene deduction of the casualties in the earthquake disaster , the simulation random network method , the coupling model in the scene deduction process , the configuration model of the emergency resources and so on are applied to the earthquake life engineering scene deduction in consideration of the uncertainty existing in the event scenario deduction ;
In the development environment of Microsoft Visual Studio 2005 , the computer - executable language of the above - mentioned relevant model is established , and the platform system of seismic life engineering is developed . The platform system provides practical tools for the research and application of the next step .

【学位授予单位】：南京航空航天大学
【学位级别】：博士
【学位授予年份】：2011
【分类号】：D035

【参考文献】