Research On Unconventional Emergency Decision-making Based On Scenario Deduction

In recent years, unconventional emergencies have obtained widely global attentions. The characteristics of the unconventional emergencies are infrequency, highly unstructured, disaster consequences unpredictability and so on, and they possess "ripple effect" at the same time, which lead to secondary derivative disasters, brings serious direct or potential harm to the society. Since the great disparities between unconventional emergencies and normal one on frequency, evolutionary status, potential impact, predictive and disposal ways, that unconventional emergencies emergency decision-making of emergency management become an important problem to solve.The characteristics of urgency, unconventionality and complexity of the unconventional emergency managerial decision, require decision maker able to respond immediately and make decisions timely. The theory and optimization technology of conventional decision are difficult to treat unconventional emergencies. Characteristics of unconventional emergency break through the abilities of conventional disposal methods, nonetheless, disposal method of "Scene-response" is effective in settle the unpredictability and non-structural of unconventional emergencies. Artificial intelligence techniques such as artificial neural network algorithm, ant colony algorithm have strong self-learning and adaptive ability, can effectively deal with complexity, nonlinear problem, which is suitable for unconventional emergency decision-making. According to the unconventional emergency system structure characteristics and different characteristics and requirements of inference stage, we choose the optimization theory, evolution theory and the multiple attribute decision making theory to construct the basis of research, and focus the studies of unconventional emergency decision support model and the method based on scenario deduction.In this paper, we extract and indicate the situational factors and its mechanism of unconventional emergency decision, use ontology to create the domain knowledge model, unify the information standard and implement information sharing and integration rapidly. We proposed methods of scene construction and inference of unconventional emergency ontology, established unconventional emergency model, scene representation inference model, Total force coupling model, mutual force coupling model and the drive coupling model of the unconventional emergency, quantitative described the coupling characteristics and coupling rule of unconventional emergency during the evolution process, identified coupling between key elements, and given the application examples.The article studies the unconventional emergency decision-making problem model system, constructs the comprehensive integration framework of emergency decision making problems. Based on department coordination, and dynamic decision-making, we put forward new concepts such as the coordination coefficient, collaborative network and synergy matrix of different department. In this article, we construct a multi-stage multi-objective multi-sector emergency collaborative decision-making model which based on the idea of multi-objective optimization. This model combines different departments and dynamic adjust decisions with the development of the situation. Considering the assessment of unconventional emergency is not only the basic foundation of disaster reduction decisions, but also one of the important indicators of evaluation of emergency decision disposal, this article presents improved fuzzy neural network model based on chaos ant colony algorithm(CACO-FNN) to solve local search inability and premature convergence of RBF neural network, and give the corresponding design and algorithm implementation. Examples proved that CACO-FNN can well reflect the nonlinear function relationship between the input and output sample data and the evaluation model based on CACO-FNN of unconventional emergencies has good application prospect.In the article, we propose an improved AHP portfolio decision based on group decision theory, and set up an IAHP decision model. According to a case of coal mine gas explosion, we construct an Emergency decision making model based on complexity-measure of the improved entropy method. We take “7.23” Yong wen line especially big traffic accident as an instance to apply model and comb the primary content, and analyze the accident scene evolution and evolutionary path. We utilize CACO-FNN model to assess the accident and make empirical analysis on the decision of “7.23” Yong wen line especially big traffic accident using IAHP decision model, which provides decision basis for the decision-making.