关键词: 铁路运输, 技术站, 协同配流, 多商品网络流, 折角车流

Abstract: The demand for freight transportation has gradually shifted from “quantity” to “quality” with the development of economy and society. As the adaptation site of a large number of transit car flows within the railway transportation network, the technical station is an important network node to reorganize the freight trains and car flows. The operation optimization of technical stations has been mostly studied with unidirectional technical station as the object for a long time. Nevertheless, the technical stations are related to each other actually on the railway network. The operation organization within the technical station affects not only its own work efficiency and benefits, but also other technical stations. If the two adjacent technical stations are regarded as a whole, and the car flow allocation is cooperatively optimized from the perspective of the regional railway network, greater transport benefits to obtain are expected. Under the background of railway freight logistics, studying the coordinated distribution method of freight trains between technical stations based on multi-commodity network flow has both theoretical and practical significance for making full use of transportation capacity and improving freight service quality.
Compared with unidirectional marshalling stations, although the carrying capacity and resorting capacity of bidirectional marshalling stations have been greatly improved, it is inevitable that angular car flows will be generated, resulting in repeated disintegration. The generation of angular car flow is not only related to the station type, car flow structure and operation characteristics of bidirectional marshalling station itself, but also affected by the freight train formation plan of other technical stations on the railway network. Taking the starting and ending points of the station technical operations including train arrival, hump disintegration, car accumulation, freight train formation and train departure as the nodes and the edges between the nodes representing the corresponding technical working process, the marshalling station is abstracted as a network graph. The flow in the network diagram represents the car flow, the capacity on the edge represents the number of cars in a train, and the cost represents the operation time. Based on the analysis of the allocation process of freight trains, this paper proposes a coordinated allocation method of freight trains between technical stations based on multi-commodity network flow. Considering the exchange of car flow at the bidirectional marshalling station in front of train operation, the car flow group numbers of the up-direction and down-direction systems are accumulated separately. This paper establishes a coordinated distribution model of freight trains between technical stations, which is to maximize the flow of trains departing from the station, the total stay time of cars at the two adjacent technical stations, and the number of angular car flows. However, the decision variables and constraints of the model will increase rapidly with the expansion of the number of arrival and departure trains. Consequently, this paper constructs the effective edge set of the car flow to reduce model size. According to the resource allocation characteristics of the transportation problem, an effective coding and fitness function is designed, and the heuristic genetic algorithm is used to optimize the freight trains to obtain the disassembly sequence and allocation plan of freight trains. This model can be solved and verified by Gurobi solver.
The experimental analysis show that the total stay time of cars in the two adjacent technical stations has been saved by 122.5 hours, and the angular car flow of the bidirectional marshalling station has been reduced by 23 cars. The improvement effect is obvious. The coordinated distribution operation between stations is beneficial to improve the overall efficiency and overall efficiency of the global transportation organization benefit.
The conditions for establishing the model in this paper are relatively ideal. Factors such as local car, the arrival and departure tracks and shunting locomotives at the technical station can be considered to find a high-quality feasible solution that is more in line with the actual situation in the future. The method and model in this paper have important guiding significance for realizing the coordinated distribution between railway technical stations, providing decision support for decision makers, and improving railway transport capacity.

Key words: railway transportation, technical station, coordinated car flow distribution, multi-commodity network flow, angular car flow