Integrating experience management and theoretical decision-making for optimizing train service network: A case study of China railway operation

Abstract

Decision-makers in railway companies often face the challenge of balancing the differences between theoretical optimization decision-making and practical experience management when designing train service network. In response, this paper proposes a practical car-to-train strategy for train service network, which originates from the operational experience of yard (marshalling station) operators. It is noteworthy that existing studies usually focus on theoretical research, while neglecting practical operation strategies at yards. These real-world operation strategies always rely on experience management by highly-experienced railway operators. In this paper, we first design train service network optimization models based on the proposed practical car-to-train strategy and theoretical optimization strategy, respectively. Subsequently, a semi-experience car-to-train strategy is developed. The established mathematical models are bi-level programming. To solve the models, a simulated annealing algorithm is developed. Taking Beijing-Shanghai Railway as an example, three numerical experiments based on three car-to-train strategies are carried out.

In the results of the experiments, the objective function of the semi-experience strategy increases by 4.82 % compared to the theoretical optimization strategy, and the proposed practical car-to-train strategy only increases by 2.02 %. For the operations of yards, the proposed practical car-to-train strategy is more in line with railway operation regulations.