The double-direction marshalling station is a large-scale technical railway station where two independ-ent sorting systems are installed and operations are complex and busy.Its wagon-flow allocation operation re-fers to simultaneously determining the make-up of each departure train and assigning and scheduling tasks for each shunting engine in its two operation systems.In such a station,a departure train in one system can simul-taneously receive ordinary wagon-flows from its own system and transfer wagon-flows from its opposite sys-tem.In this paper,assuming that the transfer wagon-flow was moved from its own system to its opposite sys-tem by utilizing the transfer yard,the large-scale mixed integer linear programming model was established with the obj ective to minimize the total time for wagons staying idle at station.The Lagrangian relaxation algorithm was developed by exploiting the structure of the model.The original problem was decomposed into much simp-ler sub-problems through relaxing the linking constraints.The algorithm was verified by case study.The re-sult shows that the proposed algorithm is obviously better than the commercial optimization software CPLEX and the empirical on-site method used in practice in terms of solution quality and computation time.%双向编组站是铁路大型技术站，有2个相对独立的改编系统，业务繁忙且复杂。每个系统中出发列车的车流既来源于本系统的普通车流，也来源于对向系统的折角车流。配流问题指同时确定2个系统中出发列车的编组内容及车流来源，指派并调度解编调机的任务。本文设定折角车流利用交换场进行转场，以车辆在站总停留时间最小为目标，构建大规模混合整数线性规划模型。利用模型结构，开发拉格朗日松弛算法，通过松弛连接约束，将原问题分解为更为简单的子问题。最后以案例测试算法，结果表明，该算法在求解质量和时间方面明显优于商业优化软件CPLEX和现场使用的经验方法。
展开▼
