An Intelligent Marshalling Model For Enterprise Station Freight Railway

Railway marshalling and transportation is an important component of the production supply chain for large and medium-sized enterprises in China. Traditional inefficient manual-made marshalling plans usually are not optimal in time or energy consuming. An efficient method needs to be developed to find the optimal marshalling plans automatically. This paper mainly studies the railway train automatic marshalling in large and medium-sized enterprises in China. Based on the investigation at the train station of a certain enterprise, according to the railway track information, carriage information, and production task information, this paper designs the abstracted railway state definitions of the station. Then based on the state definitions, the scheduling rules, and the objective function of time cost and economic cost, this paper converts abstract scheduling instructions into a general railway automatic marshalling model which can be executed by computers. By introducing the greedy strategies into different situations to optimize the algorithm of tracks occupation, carriages selection and train path selection in the model, the planning efficiency can be improved while ensuring the economic benefits of the enterprises and the quality of the formation plan. The experimental results show that the proposed model can generate fewer marshalling plans and find the optimal one faster in most cases, which proves the feasibility and availability of the model.