基于高阶矩风险的非常规突发事件应急管理优化模型研究

Supply chain risk management under general event mainly focuses on the modelling of mean and variance of random variables. However, the occurrence of unconventional emergencies will lead to the sharp fluctuations of supply chain system variables. Therefore, it is inadequate to model unconventional emergency management relying only on mean and variance, and the high order moments risk should be considered. This project will investigate the unconventional emergency management by modelling the high order moments risk. Firstly, we give a comprehensive analysis about the causes, factors, characteristics, and evolution mechanism of unconventional emergencies, and then give a determination about high order moments. Finally, the discrete distribution is obtained by scenario tree which is generated by applying moment matching method. By taking the discrete values on the scenario tree as inputting variables, a multi-period dynamic stochastic programming risk management model is proposed in our project. Based on our model, a study about unconventional emergency management is showed, and the risk control strategy of unconventional emergency is obtained by combining the scenario tree analysis method. And further the robustness of the model and the problem about computing speed are also discussed in our project. Finally, by means of empirical analysis for the unconventional emergency management and risk control strategy based on the high order moments risk to verify and improve the unconventional emergency management theories and methods.

常规事件下的供应链风险管理主要集中在对随机变量均值和方差进行建模。然而非常规突发事件的发生会导致供应链系统变量出现剧烈波动，仅仅依赖均值和方差对非常规突发事件应急管理进行建模显得非常不足，我们必须把变量的高阶矩风险纳入进来。本项目将通过对高阶矩风险的建模来对非常规突发事件应急管理给出研究。首先对非常规突发事件产生的原因、影响因素、特点及演化机理给出分析。以此为基础对随机变量高阶矩给出判断，通过矩匹配方法生成情景树给出变量的离散分布。以情景树上的离散取值为输入变量构建多阶段的动态随机规划风险量化模型，对非常规突发事件风险管理给出研究，并结合情景分析方法给出非常规突发事件下的风险控制策略，并进一步对模型的鲁棒性和快速求解问题给出讨论。最后，通过实证分析的手段对基于高阶矩风险的非常规突发事件风险度量和控制策略进行验证反馈，完善非常规突发事件应急管理理论和方法。

常规事件下的供应链风险管理主要集中在对随机变量均值和方差进行建模。然而非常规突发事件的发生会导致供应链系统变量出现剧烈波动，仅仅依赖均值和方差对非常规突发事件应急管理进行建模显得非常不足，我们必须把变量的高阶矩风险纳入进来。本项目将通过对高阶矩风险的建模来对非常规突发事件应急管理给出研究。首先对非常规突发事件产生的原因、影响因素、特点及演化机理给出分析。以此为基础对随机变量高阶矩给出判断，通过矩匹配方法生成情景树给出变量的离散分布。以情景树上的离散取值为输入变量构建多阶段的动态随机规划风险量化模型，对非常规突发事件风险管理给出研究，并结合情景分析方法给出非常规突发事件下的风险控制策略，并进一步对模型的鲁棒性和快速求解问题给出讨论。最后，通过实证分析的手段对基于高阶矩风险的非常规突发事件风险度量和控制策略进行验证反馈，完善非常规突发事件应急管理理论和方法。