Real-Time Decision Support for Crew Assignment in Double-Ended Districts for U.S. Freight Railways
针对美国货运铁路双端区段，提出优化模型和快速求解算法，在1分钟内找到最优乘务员分配方案，为调度员提供实时决策支持。
Crew costs are a significant portion of rail operating expenses, and so effective deployment of train crews is an important priority for freight railroads. This paper proposes an optimization model and develops an effective solution approach to assign crews to trains and decide when to deadhead crew members in double-ended crew districts in U.S. freight railways. To solve this model quickly, we develop several modeling and methodological enhancements including methods to reduce the problem size, strengthen the model, and generate good heuristic solutions. Computational testing and validation using 140 real-life problem instances demonstrate that our algorithm can find optimal and implementable solutions very quickly (typically within one minute), making it well-suited for practical use to provide real-time decision support to crew planners.