Study on rolling stock rescheduling under passenger demand fluctuation
To meet the dynamic demand of passengers,the need for train timetable rescheduling is becoming increasingly frequent,and the subsequent rolling stock plan also requires rescheduling. To efficiently and effectively carry out rolling stock rescheduling when changing train timetable,we focused on the rolling stock rescheduling under passenger demand fluctuation,and constructed a multi-day time-space connection network of rolling stock plan to describe the utilization,the maintenance,the connection and some basic constraints of rolling stock so as to transform the generation of feasible paths into a path search problem on the time space connection network. On this basis,the objectives were set to minimize the number of canceled additional trips,the deviation from the original plan,the cost of using rolling stock,and the deviation from the desired inventory. By considering connection and maintenance constraints,we established a path-based model of rolling stock rescheduling. According to the partition structure characteristic of the model,a Lagrangian relaxation algorithm decomposed the original problem into several independent shortest path sub-problems of rolling stock utilization. Based on the set of paths generated in the iteration process,a solver was used to solve the original model to obtain an upper bound for the problem. We tested our method on the real data of basic and semaphore timetables from several high-speed railway networks. The results show that our method can solve all cases within 65 s,with Gaps below 0.64％,and the obtained plans are better than the manual plans in terms of both solution quality and efficiency. It can be concluded that the proposed method can solve the problem quickly,improve the utilization efficiency of rolling stock and reduce the operation cost.
high-speed railwayrolling stock reschedulingshort term planningtime space connection networkLagrangian relaxation
万方数据
维普
