基于犹豫模糊认知图的非常规突发事件链双向耦合推理方法研究

As a multiple attribute set, the unconventional emergency is coupled and irreversible in the course of occurrence, development and evolution. Its secondary and derivative events are easy to form an unconventional emergency chain. There is a causal link between events and intrinsic connection between attributes of events. Furthermore, there are differences in group decision-making. All of these together with the dynamic evolution of events lead to that we cannot easily identify the event precursor and grasp the development trend of the event in time. Therefore, a bidirectional (causes and results) coupled reasoning method of unconventional emergency chain based on hesitant fuzzy cognitive map is proposed. Approaches to identify the system structure and state feature of unconventional emergency chain based on probability dependency are studied, and on this basis to construct the interaction network model for reasoning causes and results evolution of unconventional emergency. Considering the time varying characteristics, we also develop some information inference and aggregation models to simulate the evolution of unconventional emergency. Through systematic analysis of the causal chain of unconventional emergency, we expect to explore the occurrence mechanism and evolution path of unconventional emergency, and the accuracy of early warning and the ability of accident management and control can be improved. The research takes the power public emergency as the application background, and it can provide theoretical reference and decision-making basis for emergency management of electric power enterprises.

作为一个多属性集合，非常规突发事件的发生、发展、演化具有耦合性与不可逆性，其次生衍生事件形成非常规突发事件链。针对非常规突发事件链中事件的复杂因果与属性的内在关联关系、群体决策的差异及事件的动态演化等导致的难以及时识别事件前兆和准确把握事件发展态势问题，提出基于犹豫模糊认知图的非常规突发事件链双向（诱因与结果）耦合推理方法。研究基于概率依赖的非常规突发事件链的系统结构与状态特征提取方法，在此基础上，研究非常规突发事件的诱因与结果演化的联动推理网络模型，以及时变特征下的非常规突发事件演化信息推理集结方法。通过系统分析非常规突发事件因果链，探索新型不确定环境下的突发事件发生机理与演化路径，以切实提高预警的准确性与事故管控能力。项目以电力突发事件的相关问题为应用背景，以期为电力企业应急管理提供理论参考与决策依据。

针对非常规突发事件发生、发展、演化过程中具有的模糊性、随机性与动态性特点，群体决策差异及事件的动态演化等导致的难以及时识别事件前兆和准确把握事件发展态势等问题，提出基于犹豫模糊认知图的非常规突发事件成因预测、信息融合方法。研究基于概率分布的的犹豫模糊信息特征提取方法；探讨影响应急管理成效的关键风险因素。在对突发事件成因分析的基础上，研究其诱因与结果演化联动推理网络模型，以及动态环境下的信息融合方法。通过系统分析突发事件因果链，探索新型不确定环境下的应急决策方法，以切实提高突发事件推理技术的灵活性与模糊决策的可靠性。