铁路技术站阶段计划滚动优化方法

Rolling Optimization Method for Stage Plans of Railway Technical Stations

针对目前铁路技术站阶段计划主流研究多聚焦于对某一固定周期进行优化、较少考虑相邻计划间动态衔接的不足,提出一种阶段计划的滚动优化方法。从计划编制、车流调配和资源占用3个方面分析技术站阶段计划的滚动编制特性;以技术站正点出发的列车最多、车辆周转率最高和车站整体作业效率最高为字典序多目标,构建通用的技术站阶段计划滚动优化模型;设计带初始解的混合迭代算法求解模型;依托某技术站实际数据构造算例进行验证。结果表明:求得最优解的总耗时不超过90 s,共配流成功出发列车13列;考虑相邻计划衔接后,考虑前序阶段计划即“向前看”使满轴出发列车优先级总和增加了5.83%,考虑后序阶段计划即“向后看”使所有作业用时减少了6.91%;相比基本深度优先回溯算法和随机重启算法,混合迭代算法在求解效率上分别提高了73.8%和70.8%,在求解质量上分别提高了692.3%和77.6%。滚动优化方法得到的结果稳定性更好、更符合现实中阶段计划的连续滚动特性,且算法的求解效率更高。

Aiming at the deficiency that the current mainstream research on stage plan of railway technical station mainly focused on the optimization of a fixed period and gave less consideration on the dynamic connection between adjacent plans,a rolling optimization method of stage plan was proposed.The rolling planning characteristics of the technical station stage plan were analyzed from three aspects:planning,wagonflow allocation and resource occupation.A general rolling optimization model of the technical station stage plan was constructed,which took the maximum number of trains departing on time,the highest railcar turnover rate,and the highest overall operation efficiency of the technical station as lexicographic multi-objective.A hybrid iterative algorithm with initial solutions was designed to solve the model.A calculation example based on the actual data of a certain technical station for verification was constructed.The results show that the total time required to obtain the optimal solution does not exceed 90 s,and a total of 13 trains are successfully assigned to set off.After considering the connection between adjacent plans,the total priority of full-axis departure trains is increased by 5.83%when considering the pre-stage plan“looking forward”,and the time spent on all operations is reduced by 6.91%when considering the post-stage plan“looking backward”.Compared with the basic depth-first backtracking algorithm and the random restart algorithm,the proposed hybrid iterative algorithm improves the solving efficiency by 73.8%and 70.8%respectively,and improves the solving quality by 692.3%and 77.6%respectively.The results obtained by the rolling optimization method have better stability and are more in line with the continuous rolling characteristics of stage plans in reality,and the algorithm has higher solving efficiency.