In order to relieve the issue of persistent capacity strain in the main passenger flow direction and significant underutilization in the non-main direction caused by tidal passenger flow, a flexible express skip-stop strategy was proposed, based on asymmetric train operation patterns for a dual-depot double-line metro system to optimizing the train schedule during peak hours. The objective function aims to minimize both total passenger waiting time and train dwell time. A mixed-integer nonlinear programming model was formulated, integrating constraints related to timetables, flexible skip-stop operations, rolling stock circulation, and passenger flow balance. A hybrid algorithm combining genetic algorithm and simulated annealing was developed to solve the proposed model. The coordinated optimization of train schedules, stop plans, and rolling stock circulation plans was achieved. Finally, the validity of the model and algorithm was verified through a case study of Jinan Metro Line 3. The optimization outcomes of three operational schemes: symmetric operation, flexible skip-stop, and a combined approach integrating asymmetric operation with flexible skip-stop were compared. The results show that the combined asymmetric transportation strategy and flexible skip-stop approach reduces passenger waiting time by 3.43%, shortens train dwell time-related duration by 10.93%, and lowers the objective function value by 7.56%.It is concluded that the proposed asymmetric transportation strategy combined with a flexible skip-stop operation scheme effectively addresses tidal passenger flow phenomena, enhances operational efficiency and service quality during peak hours in urban rail transit, and provides new insights and references for optimizing operational management.
地铁实施不成对运输组织模式的核心问题在于车底运用计划的制定。由于线路双向开行的列车数量存在差异,需设计合理的车底周转方案,才能在有限的车底资源条件下实现运力的高效配置。目前,关于列车运行图与车底运用协同优化的研究已取得丰富成果。Zhou等[4]提出一种混合整数线性规划(mixed-integer linear programming,MILP)模型,用于联合优化潮汐客流下的列车时刻表与车底周转计划,并设计变邻域搜索算法(variable neighborhood search,VNS)以快速生成高质量解。周文梁等[5]基于给定的列车牵引策略集,综合考虑列车牵引能耗与车底运用时间,采用非支配排序遗传算法Ⅱ与熵权法相结合的方法,对列车到发时刻及车底运用计划进行求解。李思杰等[6]结合潮汐客流背景下的客流时变特征,构建了不成对列车运行时刻表与车底运用优化模型。Gong等[7]建立了包含节能策略的列车时刻表与车辆循环计划综合优化模型,并采用自适应大邻域搜索(adaptive large neighborhood search,ALNS)算法求解列车运行计划。苏子贞等[8]以最小化乘客与企业总成本及出入段作业接续占比的加权和为目标,构建不成对运行图优化模型,通过贪心算法求解车底周转计划。
列车跳停应覆盖到每一对起讫点车站(origin to destination,OD)以满足乘客的出行需求,本文进一步提出一种改进灵活跳停策略,具体而言,设列车i跳停车站集合为β1,列车i+1跳停车站集合为β2,则列车i+2在满足相邻两车不能连续跳停约的束基础上须在车站集合β1∪β2实施停站,该改进灵活跳停策略能够有效覆盖各类乘客的出行需求,改进灵活跳停策略列车停站方案示意图如图3所示。
2 问题描述
2.1 问题背景
选取一条具有潮汐客流背景的双向地铁线路为研究对象,线路两端各配备一个车场。线路共有2N个车站,为便于区分,将车站按照上下行方向分别编号,上行方向车站编号为1,2,…,N,下行方向车站编号为N+1、N+2,…,2N。上行方向为非主客流方向,下行方向为主客流方向。该线路下结合快慢车运营的不成对运行组织模式如图4所示。本文中研究时段为07:00—10:00,考虑到不同时段乘客出行需求的差异性,将研究时段分为早高峰(07:00—8:00)、次高峰(08:00—9:00)、平峰(09:00—10:00)3个分时段,根据自动售票系统(automatic fare collection,AFC)刷卡记录整理各时段客流OD数据,分时段分向灵活调整列车发车间隔,满足乘客出行需求。
$\begin{aligned}q_{i, n}^{\text {arr }} & =\lambda_{k, n}\left(d_{i, n}-d_{i-1, n}\right), \quad \forall i, i-1 \in I ; \\& n \in S ; k \in K\end{aligned}$
候车乘客数计算公式为
$q_{i, n}^{\text {wait }}=q_{i-1, n}^{\mathrm{z}}+q_{i, n}^{\text {arr }}, \quad \forall i \in I ; n \in S$
$\begin{aligned}q_{i, n}^{\text {board }}= & \min \left(y_{i, n} q_{i, n}^{\text {wait }}, C_{i, n}^{\text {remain }}\right), \quad \forall i \in I ; \\& n, n-1 \in S\end{aligned}$
列车i在区间(n,n+1)车内人数约束由上一区间车内人数、n站上车人数、下车人数共同决定,表达式为
$\begin{aligned}p_{i, n}= & q_{i, n}^{\text {board }}+p_{i, n-1}\left(1-y_{i, n} \mu_{k, n}\right), \\& \forall i \in I ; n, n-1 \in S ; k \in K\end{aligned}$
为了更全面地研究早高峰时段乘客出行需求,选取07:00—10:00为研究时段,乘客出行OD统计如图8所示,为精细化满足乘客动态出行需求,将研究时段划分为早高峰(07:00—08:00)、次高峰(08:00—09:00)、平峰(09:00—10:00),分别统计3个时段内乘客出行数据。为适应潮汐客流需求,结合客流数据、列车容量及乘客候车体验综合测算,规定上行方向开行24趟列车,下行方向开行31趟列车。各类参数取值如表2所示。α1取值0.4,α2取值0.6。根据上述线路信息、客流数据及列车运行相关参数,在运行环境为Windows11 系统,Intel(R) Core(TM) i5-1035G1 CPU @1.00 GHz 1.19 GHz,采用Python程序编写遗传-模拟退火混合算法(GA-SA)求解该模型。
5.2 优化结果分析
采用遗传-模拟退火混合算法求解该案例,算法迭代收敛曲线图如图9所示,在第53代后区域收敛,最终得到最优解为402 500.4 s。求解得到优化后的列车不成对运行图如图10所示,可以看出,在研究时段内上行方向共上线12趟车底,下行方向共上线18趟车底。优化后的列车发车间隔随时间推移呈现逐渐增大的趋势。在早高峰时段07:00—08:00内,上行平均发车间隔为6 min 40 s、下行平均发车间隔为5 min 4 s;次高峰时段08:00—09:00内,上行平均发车间隔为7 min 26 s,下行平均发车间隔为6 min 1 s;平峰期09:00—10:00内,上行平均发车间隔为10 min 3 s,下行平均发车间隔为7 min 29 s。优化后的列车停站方案如图11所示,其中0代表列车跳站,1代表列车停站。其中末班车为站站停列车。
JiaJianlin, ZhangWanting, WangZhenbao, et al. Passenger flow prediction for newly opened rail transit stations based on artificial neural network[J]. Science Technology and Engineering, 2024, 24(22): 9636-9644.
XiangHongyan, YuanFatao. Passenger flow identification model for rail transit stations based on automatic fare collection data[J]. Science Technology and Engineering, 2021, 21(4): 1568-1573.
[7]
ZhouH S, QiJ G, YangL X, et al. Joint optimization of train timetabling and rolling stock circulation planning: a novel flexible train composition mode[J]. Transportation Research Part B, 2022, 162: 352-385.
ZhouWenliang, HuangYu, DengLianbo. Optimization of urban rail transit timetable considering energy-saving operation and rolling stock utilization[J]. Journal of Railway Science and Engineering, 2023, 20(2): 473-482.
LiSijie, ShangYuxin, LiuZhigang. Collaborative optimization of train timetable and rolling stock circulation for urban rail transit under tidal passenger flow[J]. Journal of Transportation Systems Engineering and Information Technology, 2023, 23(4): 175-183.
[12]
GongC C, LuanX J, YangL X, et al. Integrated optimization of train timetabling and rolling stock circulation problem with flexible short-turning and energy-saving strategies[J]. Transportation Research Part C: Emerging Technologies, 2024, 166: 104756.
SuZizhen, ChenShaokuan, YangAn’an, et al. Optimization of asymmetric train diagram with variable marshaling for metro tidal passenger flow[J]. Journal of Railway Science and Engineering, 2024, 21(6): 2183-2193.
LiZongran, BaiYun, ChenYao, et al. Optimization model of asymmetric train diagram for subway under tidal passenger flow[J]. China Railway Science, 2022, 43(6): 207-218.
[17]
DengL B, CaiL, ZhangG Q, et al. Energy consumption analysis of urban rail fast and slow train modes based on train running curve optimization[J]. Energy Reports, 2024, 11: 412-422.
SunYuanguang, DengChengyuan, PengLei, et al. Energy-saving optimization method of asymmetric subway train diagram driven by tidal passenger flow demand[J]. Journal of Transportation Systems Engineering and Information Technology, 2024, 24(5): 128-139.
YangPenglei, GuoJianyuan, QinYong, et al. Joint optimization of train timetable and rolling stock circulation under spatiotemporally uneven passenger flow[J]. Railway Transport and Economy, 2026, 48(1): 160-172.
MengFanting, YangLixing, LuYahan, et al. Cooperative optimization of urban rail train timetable and station passenger flow control under skip-stop strategy[J]. Journal of Transportation Systems Engineering and Information Technology, 2021, 21(3): 156-162.
[24]
张霁月. 大客流条件下城市轨道交通列车跳停方案优化[D]. 大连: 大连交通大学, 2023.
[25]
ZhangJiyue. Optimization of train skip-stop strategy for urban rail transit under large passenger flow conditions[D]. Dalian: Dalian Jiaotong University, 2023.
[26]
EslamiA, ShafahiY, BafandkarS. Optimizing and synchronizing timetables in an urban subway network considering trains’ speed profiles and skip-stop strategy[J]. Journal of Rail Transport Planning & Management, 2025, 34: 100520.