Discrete Event Simulation Based Train Departure Interval Optimization For Urban Rail Transit

The development of urban mass transit is an important means for major cities to cope with urban traffic congestion.With the expansion of the scale of urban rail transit network,its passenger traffic is growing,bringing new challenges to the operation of urban mass transit.At present,most of the city's train departure interval are determined by the maximum section passenger flow of the line.The main problems are as follows:(1)Establishment of departure interval is based on the macro level indicators,and insufficient attention is paid to the micro-operation status of the station;(2)The consideration of dynamics of the passenger flow to the station(especially the passenger flow in the bottleneck area of station service capacity),randomness,state correlation and effects of congestion propagation are lacking.Therefore,the existing operation plan may lead to overcrowding of the station and increase the security risk.As a meso-level model,the queuing network can better describe the passengers' running process in the station and reveal the passengers' operating rules.And the computational efficiency of the queuing network model is higher than the simulation model at the micro level.Therefore,a queuing network model is introduced to describe the operation process of passengers on the station and on the train online to ensure that passengers do not over-aggregate at the station and on the train line.It is difficult to establish a queuing network analysis model considering dynamic randomness.Therefore,the discrete event simulation method is used to establish a queuing network simulation model for passenger and train operation.Based on the data from the Automatic Fare Collection(AFC),the dynamics,randomness,state correlation and effects of congestion propagation of the passenger flow in the station have been considered,and the corresponding train interval time optimization method has been proposed.The main research contents are as follows:Firstly,in view of the shortcomings that the traditional train departure interval optimiztion does not adequately consider the dynamic randomness of arrival and service time of passenger flow in the rail transit system,the PH distribution with infinite approximation ability and queuing theory model has been used to model the urban mass transit system.The innovation quantification considered the state correlation of passenger service time in the queuing process of service facilities such as passages and stairs,and provided a quantitative description method of dynamic random environment for the optimization of train running schemes of urban mass transit system.Secondly,the networked urban mass transit queuing simulation model has been built by Simulink discrete event simulation platform,which solved the key problems of passenger-train dual-entity hybrid simulation,state-related traffic service facility service process simulation,commutating passenger flow diversion and merge simulation.Based on the data from AFC of Chengdu Metro,the efficiency and accuracy of the simulation model have been verified.Finally,the optimization model and algorithm of train departure interval based on simulated queuing network have been established.By comparing and analyzing the optimization results of considering state-related service time and regardless of staterelated service time in a specific example,a conclusion has been drawn that the optimization method considering the service time state can better configure the appropriate train departure interval for the peak passenger flow congestion.The model in this paper can provide decision-making basis for for rail transit operators,so as to optimize train departure interval on the basis of considering passenger flow aggregation at stations,maximize the advantages of urban rail transit and alleviate passenger congestion.