Research On Emergency Decision Methods Based On Case-based Reasoning

Emergency decision-making is a kind of decision-making scenario in which various emergencies are accurately determined and unconventional emergency measures are taken under uncertain conditions.Emergency decision-making is the core of emergency management,different emergency decision-making behaviors often have different treatment effects on emergencies.The existing emergency decision-making theories and methods can not solve the emergency decision problems with complex features.We can use the historical emergency cases to assist the decision-making,that is,the case-based reasoning technology.This article has studied in-depth the emergency decision-making methods based on case-based reasoning,mainly completed the following five aspects of research work:(1)Research framework for emergency decision-making based on case-based reasoning.Through summarizing and refining the advantages and disadvantages of existing research results,researching and combing case-based reasoning techniques,RBF neural network algorithms,genetic optimization algorithms,particle swarm optimization algorithms,and ant colony optimization algorithms,emergency decision-making merhod based on case-based reasoning is proposed.The technical route and research framework of the method are divided into three parts:preprocessing,emergency decision analysis,and program generation,which lay a theoretical foundation for the research work of this paper;(2)Description and representation of emergency cases,storage of emergency cases,extraction of emergency case attributes,and calculation methods of feature weights.For the description,representation,storage,and attribute extraction of emergency cases,the historical case set construction methods for coal mine datasets and typhoon datasets used in this paper were presented,formal descriptions of historical cases and target cases were presented;Aiming at the problem that the attribute feature weights are too dependent on the subjective decision of the decision maker,an improved particle swarm algorithm(SGAPAO algorithm)was proposed,which provides a theoretical basis for subsequent research on similar historical emergency case retrieval;(3)Retrieval method for similar historical emergency cases.Aiming at the retrieval problem of target cases and similar historical cases,a method of emergency case retrieval based on RBF neural network was proposed,which solved the problem that the traditional retrieval algorithm in the retrieval module based on case-based reasoning has a long time,low accuracy,and not a wide range of fitness.A case retrieval system based on case-based reasoning and RBF neural network was constructed.According to the setting of the optimal learning rate of RBF neural network under this system,an improved particle swarm optimization algorithm(OEPSO-ACO algorithm)was proposed,making case retrieval more effective and outstanding,laying the foundation for the method of generating emergency plans;(4)Dynamic generation and optimization methods of alternative emergency decision-making plans based on case-based reasoning.Aiming at the problem of how to dynamically adjust emergency plans with the development of emergency situations,a new dynamic generation,adjustment,and optimization method for emergency plan based on case-based reasoning was proposed.Through the analysis of the occurrence,development and evolution process of emergencies,the emergency decision-making process of emergencies is considered as a multi-stage and multi-agent emergency decision-making process.According to the different emergency information fed back from each stage,different emergency plans are generated,and emergency plan at different stages are evaluated.So as to provide the optimal and corrective methods for the emergency plan,and provide emergency decision makers with emergency plan;(5)The construction and application research of emergency intelligent decision support system model based on case-based reasoning and RBF neural network.Through the combing of case-based reasoning technology,RBF neural network,Bayesian belief network and intelligent decision support system,an emergency intelligence decision support system model based on case-based reasoning and RBF neural network is proposed,and targeted to solve the shortcomings of the original IDSS.The proposed emergency IDSS based on CBR and RBF neural network was applied to the generation and optimization of coal mine gas explosion emergency plan and typhoon emergency plan,verifying its feasibility and validity.Providing a useful reference for the application of emergency decision-making method based on the case-based reasoning proposed in this paper.In summary,the emergency decision-making method based on case-based reasoning proposed in this paper can provide emergency decision-makers with guidance and help,and lay a solid foundation for related research on theoretical and application levels.