Research On Railway Emergency Decision Methods Based On Bayesian Networks And CBR

As a modern means of transportation,railway transportation is the backbone of the national comprehensive three-dimensional transportation network,which is of great significance to the social economy development and the improvement of the national living quality.With the innovation and development of railway technology and the rapid expansion of railway scale,the correlation between multi-standards and the coupling of disaster-causing factors have brought more severe tests to the railway safety transportation,and also put forward higher requirements to the "one case and three systems" emergency management system in our country.Based on the current research of railway emergency management system,emergency situation evolution trend and emergency decision-making method at home and abroad,in this paper,the emergency decision-making of railway emergency was taken as the research object,and the construction of railway emergency situation prediction,emergency decision-making method and emergency auxiliary system were discussed.First,this paper clarify the definition,classification and characteristics of railway emergency,the implement and decision-making method of the railway emergency management system in our country were carried on the thorough research,on the basis of the previous research,aiming at the deficiency of railway emergency decision model,the " scenarioresponse " model of emergency decision-making method were proposed.Secondly,on the basis of the traditional situation assessment method,this paper applied Bayesian network to construct the situation prediction network model to study the situation evolution of railway emergencies.The node,structure and probability distribution method of the Bayesian situation prediction network model are introduced in detail.Taking a train derailment event as an example,the situation deduction results of the accident are completed by using the model to support the generation of targeted emergency decision plan.Thirdly,under the situation prediction of railway emergencies,combined with the application of CBR technology,starting from the logical process of emergency decisionmaking,the case attributes are firstly extracted and weighted with the help of cloud model,and then the scene retrieval algorithm of "structural similarity-attribute similarity-global similarity" is designed.By modifying the retrieved similar cases,the emergency plan of the target case was obtained,and the validity and feasibility of the method in the emergency decision reasoning of railway emergencies were verified by the application of a practical case.Finally,based on the above key technology research,the railway emergency auxiliary decision-making system is constructed,the decision-making process,module function and basic interface of the system are designed,and the basic function and application effect of the prototype system are demonstrated,which verifies the practical application value of the method proposed in this paper.