不确定全面风险分析框架下的供应链风险建模与优化研究

Risk analysis is a decision-making process to reduce the negative impact of risk. It has been widely applied to solving real supply chain problems. Compared with the traditional methodologies, the advanced risk management has turned to comprehensive risk management. Moreover, most of the supply chain decision-making problems are under the complex and uncertain environments. Therefore, how to build an uncertain comprehensive risk analysis framework to study the supply chain risk is worthy of the further related research. Supply chain risk modeling and optimization under the framework of uncertain comprehensive risk analysis is designed for decision-making process with the overall risk under a complex uncertain environment in this project. We will aggregate supply risk, demand risk, sustainable manufacturing risk, financial risk, strategic risk and sudden crisis risk to construct a comprehensive risk model. The designed project includes: 1) establish a overall uncertain comprehensive risk analysis framework, construct the models to describe risk events based on empirical study of risk events, and establish the comprehensive measurement of the risk; 2) based on the sustainability of the supply chain performance, design a sustainable risk optimization model, derive uncertain mathematical programming models with different decision criterions, and design the algorithms; 3) by using the method of financial tools and operation management tools, study the comprehensive risk management strategy and management mechanism design problems.

风险分析是识别风险的消极影响并进行主动管理的有效手段，该决策方法已经被广泛运用到现实供应链问题中。而国际先进的风险分析已经开始向全面风险分析方式转变，此外，现实的供应链系统和其外部环境都呈现高度复杂的不确定性。因此，如何在复杂的不确定环境中应用全面风险分析来整体研究供应链风险是一个值得探讨和研究的课题。本项目将研究不确定全面风险分析框架下的供应链风险建模与优化，它综合供给风险、需求风险、制造风险、融资风险、可持续经营风险和突发危机风险进行全面整体的风险建模以及优化决策研究。具体内容包括：1)建立不确定全面风险分析框架，对风险事件进行实证研究，建立描述这些不确定性因素的模型，并建立全面风险的度量；2)以供应链价值表现的可持续性为目标，构建可持续性风险优化模型，并基于不同决策准则，建立不确定数学规划模型，并设计算法求解；3)结合金融工具和运作管理方法，研究全面风险管理策略和管理机制设计问题。

现实的供应链系统外部环境呈现高度复杂的不确定性，本项目对外部风险进行了全面的识别、分类和建模分析。创新性的利用大数据方法和机器学习等技术探讨了风险识别的原理与算法。并利用风险评估的技术对单一风险、多源风险的频率、强度和相互作用机制建模研究。针对直接突发外部冲击风险，则利用计量经济学方法研究风险在供应链中的影响结果，利用公开数据分行业和区域研究风险发生后不同时点的影响程度。对于巨灾风险和信息安全风险，则通过网络传染模型和系统动力学建模，利用参数校准和随机模拟比较不同决策目标下的最优性结果。使用随机时滞和随机切换微分方程描述会产生长期影响的经济环境变量状态，建立最优控制模型，分析资源配置策略降低风险，还探讨利用金融衍生工具研究全面风险管理策略和管理机制设计问题。预测生产要素和产业结构的长期变化趋势，进行可持续性风险分析，探讨转型升级可行方向。