Research On Emergency Decision Of Urban Rail Transit Based On Train Graph Adjustment

With its advantages of public,green,fast,punctual and comfortable,urban rail transit has gradually become the backbone of China’s urban transportation system.The mileage of operation has been increasing year by year,and the degree of network has also been gradually improved.Under the background of the increasing dependence of citizens on urban rail transit,the improvement of its service level has gradually been placed in a more important position.In the daily operation of urban rail transit system,emergencies are difficult to predict and ubiquitous.The occurrence of sudden delay will affect the normal driving order of a line,and may spread the delay in a large area,resulting in the accumulation of passenger flow in multiple lines and stations,and difficulties in station organization.In order to make specific traffic scheduling decisions scientifically,quickly and automatically in the case of sudden delay,this paper establishes a set of decision-making method system based on the existing operation diagram adjustment scheme.The specific work is as follows:1.All the existing urban rail transit traffic scheduling schemes are analyzed.The characteristics of each scheme’s operation diagram structure are shown,and their internal principles and specific situations are described.2.Based on the applicability and characteristics of the drawing adjustment scheme,the decision model of the operation diagram adjustment scheme selection is established.In this paper,the selection of operation diagram adjustment scheme under emergency is regarded as a multi-criteria decision-making problem under the determination scheme.The 13 existing operation diagram adjustment schemes are taken as the pending schemes,the adaptability between the characteristics of emergency and the operation diagram adjustment schemes,and the economic practicability of the schemes are taken as the decision criteria,and the decision model based on TOPSIS method is established,and the TOPSIS method is improved in three aspects according to the characteristics of the decision problem.Taking Kanzimen Station of Nanjing Metro Line 3 as an example to verify the rationality of the result of the decision model,the proposed scheme is to cut down the train and slow down the speed.Through analysis,the mapping scheme obtained from the decision model has high efficiency,economy and applicability in this case.3.A time point migration model is established based on Speed up / Extend pit stops/slow down.This model is based on the assumption that the recommended scheme based on the improved TOPSIS method in an emergency is to Speed up / Extend pit stops/slow down,then the next step is to solve the specific operation diagram schedule,so as to realize the full automatic solution of the scheme.The model takes the deviation amount of the arrival and departure time of all trains that do not meet the normal operation constraints at each station after the occurrence of an emergency as the solving object,and minimizes the sum of the total deviation amount as the optimization objective to establish a single-layer nonlinear programming model.In order to deal with the problem that the dimension of solving variables is too high,the solution process of the optimization problem is divided into the iteration of solving the train schedule one by one,and the improved genetic algorithm of mutation probability and crossover probability successive optimization is used to solve the problem.The feasibility of the model is verified by taking two sudden delays of train diagram insertion in Nanjing Metro Line 1 during evening rush hour(17:00-20:00)as an example.The results show that the model can calculate the schedule offset required to resume normal operation within one minute,which meets the requirements of rationality and rapidity.4.The decision model of backup train operation based on BP neural network is established.Assuming that the recommended scheme based on the decision model of scheme selection based on the improved TOPSIS method is to call the standby train in an emergency,the next step is to work out the specific standby train operation scheme according to the actual situation of the line.BP neural network is used to construct the decision model.After 50000 times of training with 20 training cases,the structure with minimum training error was selected,and the structure of the neural network was 9×12×3.Taking three simulated emergencies as examples,using the neural network to make decisions,the recommended schemes obtained are reasonable after analysis.