高峰时段郊区线考虑列车容量约束的停站方案优化方法研究

Peak Hour Stopping Scheme Optimization of Suburban Line Considering Train Capacity Constraints

大城市郊区线路的客流往往具有明显的长距离出行和时空分布不均衡特征,尤其在高峰时段存在因列车容量约束导致乘客滞留站台的现象,而开行不同站停方式的列车是解决上述问题的有效手段。在计算上下车、进站/换乘及留乘等客流数量基础上,本文建立以包括候车时间和旅行时间在内的乘客总出行成本最小为目标的优化模型;以是否停站的0-1变量对所有列车进行统一编码,通过设置最大触发灾变次数,设计基于灾变思想的改进遗传算法;定义研究时段相邻列车在各车站发车间隔的偏差均值和偏差离差为评价运行图发车间隔均衡性的指标,实现考虑列车容量约束的停站方案优化。以北京地铁15号线为案例,结果表明:改进遗传算法通过灾变操作可有效跳出局部最优;算法可得到设计的3种场景条件下的最优停站方案,相较于场景1,场景2和场景3目标值分别提升12.20%和4.28%;采用跳停策略,可以有效减少列车运行时间,相较于场景1,场景2和场景3的下行方向在到达终点站时分别节省17.49 s和17.97 s。

The passenger flow of suburban rail transit lines in metropolises often have long travel distance and uneven temporal and spatial distributions.Especially during peak hours,there are stranded passengers on the platform due to limited train capacities.Train operations with differentiated stopping scheme is an effective way to improve the peak hour operations.This paper proposes an optimization model that minimizes the total passenger travel cost including waiting time and travel time based on the calculation of boarding/alighting,transferring/arriving,and remaining onboard passenger numbers.With the uniform encoding of 0-1 variables that indicate whether trains stop at a station,an improved genetic algorithm with catastrophe theory is developed by setting the maximum number of catastrophes triggered.The mean deviation and deviation range of the departure intervals between subsequent trains at stations are used as the indicators to evaluate the equilibrium of the timetable,and the train stopping patterns are then optimized in consideration of the train capacities.The case study in Beijing Subway Line 15 shows that the improved genetic algorithm can effectively escape from local optima through catastrophe operations.The algorithm can obtain the optimal stopping scheme under three designed scenarios,particularly,the objective value of scenarios 2 and 3 was respectively improved by 12.20%and 4.28%compared to scenario 1.The skip-stop strategy can effectively reduce the train's running time,for example,the train running time in scenarios 2 and 3 was respectively saved by 17.49 s and 17.97 s in the downstream direction compared to scenario 1.