Research On Optimal Dispatching Problem Of Railway Shifting Yard Vehicles

The marshalling yard plays a key role in the railway network.Its main function is to disassemble the incoming trains and organize the vehicles in the station to form the train leaving the station.At present,the railway marshalling is carried out according to the marshalling plan in China's railway marshalling yard.When the trains are unbalanced,arriving densely and the destination stations of the vehicles are far more than the number of station channels,it is easy to cause the vehicles to stay in the stations,thus affecting the transport efficiency of the whole railway system.Therefore,it is of great significance to study the optimization method of multi-track vehicle dispatching in railway shunting yard for improving railway freight transport capacity and improving railway economic benefits.This paper studies the problem of optimal vehicle dispatching in railway shunting yard.Aiming at the problem of vehicle channel scheduling in the station,a mathematical model of vehicle scheduling on multi-track in the dispatching yard is established,and the greedy strategy based on graph coloring and DQN algorithm based on Q learning are designed to solve the model respectively.The specific work of this paper is as follows:(1)The establishment of actual dispatching operation and dispatching model of single-in-single-out multi-track vehicles.There are many types of shunting yard,such as single in and single out type,double in and single out type,double in and double out type,etc.This paper takes shunting yard as the research object,and establishes a mathematical model according to the actual situation of railway multi-track vehicle scheduling and the actual dispatching infrastructure of railway shunting yard.The mathematical model established for the vehicle scheduling problem is more in line with the actual railway multi-track vehicle scheduling process.(2)Based on the mathematical model of single-in-single-out railway vehicle scheduling established above,this chapter proposes a greedy algorithm based on graph coloring.Marshalling yard vehicles can be converted to the points on the graph,the number of marshalling yard tracks to two parts graph to figure the number of maximum color dyeing vehicle scheduling fields to marshalling yard into the figure on the coloring problem,using the greedy strategy to all the edges on the drawing for dyeing,with different colors represent different tracks,The validity of the proposed algorithm is verified by experiments.(3)Based on the railway stock vehicle scheduling mathematical model established in(1),this chapter proposes the DQN algorithm based on Q learning to solve the model.Based on the greedy algorithm using graph coloring in(2)to solve the model,when the amount of data processed is large,the operation time is longer and the scheduling scheme cannot be provided in time.The neural network in deep reinforcement learning has the ability to process a large amount of data quickly,and the vehicle scheduling in the yard is a continuous decision-making process,which conforms to the Markov decision process.Therefore,a more general and convenient vehicle scheduling algorithm in the railway yard is proposed.Finally,the experiment proves that the proposed algorithm can achieve good results of vehicle track scheduling,which provides a certain theoretical reference for the vehicle track scheduling of railway freight transport yards.