Robust stage plan for a railway marshalling station

In order to reduce the average dwell time of the railcars at railway marshalling stations, previous studies on stage plan are almost exclusively based on the assumptions of accurate and unchanged input data. However, delays in rail freight transport are common. The occurrences of these unpredictable disturbances are unavoidable to lead to the stage plan should be rescheduled. In this paper, we firstly present a robust model for the stage plan based on minimal perturbation in dynamic cumulative scheduling theory with lexicographic multi-objective, such that robustness is considered as the primary objective and the benefits of railway marshalling stations is the secondary one. Then, a hybrid multi-point constructive search algorithm is designed to solve the model iteratively. Experimental results show that the hybrid algorithm and minimal perturbation in dynamic cumulative scheduling theory approaches could be a promising way to improve the stage plans' robustness.