非常规突发事件下供应链的风险扩散动态识别与多阶协同响应决策

In recent years, unconventional emergencies such as earthquakes, conflagrations occur frequently, which have great negative impacts on the proper operation of supply chain, even will cause supply chain disruptions. At present, the risk management of supply chain facing unconventional emergencies focuses on beforehand preventions. There is no a systematic theory and method to conduct how to identify accurately risk diffusions and make rapidly response decisions after an unconventional emergency occurs. Therefore, according to the characteristics of unconventional emergencies including unpredictability, diffusivity and dynamic nature, taking enhancing operation performance of supply chain risk management as research target, case based reasoning method, stochastic network modeling, and rule discovery technology as main research approaches, this project is designed systematically and in detail. Based on the analysis processes of “characteristics analysis-attribute abstraction-semantic modeling”, this project studies the structural description method for risk diffusions and response decision behaviors under unconventional emergencies. Then, this project explores the realization method of multi-case integrated reasoning, and establishes GERT-GSPN dynamic identification model of risk diffusions in supply chain. Also, integrating rule discovery technology, this project studies the multi-phase coordinative response decision method from the two perspectives of time series and network structure. The research will contribute to developing theories and practices of supply chain risk management.

近年来，地震、火灾等非常规突发事件频发，对供应链的正常运作造成了巨大冲击，甚至会引起整个供应链的断裂。目前，供应链非常规突发事件的风险管理主要侧重于事前防范，而在事件发生后，如何针对目标事件具体特性，对风险扩散进行准确识别，并迅速作出响应决策，还缺乏一套系统化的理论和方法支撑。本项目针对供应链非常规突发事件的难以预测性、扩散放大性和动态过程性等特征，以提升供应链风险管理的运行效能为目标，以案例推理方法、随机网络模型、规则挖掘技术为主要研究手段，基于“特征分析—属性提取—语义建模”的解析过程，研究非常规突发事件下供应链风险扩散和响应决策行为的结构化描述方法；探索基于D-S证据链的多源案例融合推理的实现路径，并构建风险扩散GERT-GSPN动态识别模型；集成规则挖掘技术，从时间序列和网络结构两个角度，研究供应链多阶协同响应决策方法。本项目对于拓展供应链风险管理的基础理论和应用领域具有重要意义。

近年来，地震、火灾等非常规突发事件频发，对供应链的正常运作造成了巨大冲击，甚至.会引起整个供应链的断裂。目前，供应链非常规突发事件的风险管理主要侧重于事前防范，而在事件发生后，如何针对目标事件具体特性，对风险传递进行准确识别，并迅速作出响应决策，还缺乏一套系统化的理论和方法支撑。本项目通过现实案例对比、典型特征分析、国内外研究现状梳理，将供应链非常规突发事件管理的探索性研究聚焦在事发之后的风险传递识别和响应决策环节，完成了以下研究内容：基于知识图谱的供应链突发事件静态描述；基于改进型GERT网络模型的风险传递动态描述；基于D-S证据理论的供应链非常规突发事件特征信息合成研究；基于案例的供应链非常规突发事件风险演变规则推理方法；基于可信性理论的三级供应链运作中断恢复模型及其算法。基于上述研究，发表了6篇学术论文，获科研奖励1项，培养硕士毕业生3名。以上研究丰富和发展了供应链风险管理的基础理论和应用领域。