交通运输系统工程与信息

交通运输系统工程与信息 ›› 2025, Vol. 25 ›› Issue (6): 249-264.DOI: 10.16097/j.cnki.1009-6744.2025.06.023

• 系统工程理论与方法 • 上一篇    下一篇

区间重叠下多线公交车辆调度和泊位设置联合优化

胡宝雨，刘文磊，程国柱*   

  1. 东北林业大学，土木与交通学院，哈尔滨150040
  • 收稿日期:2025-07-21 修回日期:2025-08-26 接受日期:2025-09-01 出版日期:2025-12-25 发布日期:2025-12-24
  • 作者简介:胡宝雨(1987— )，男，黑龙江宾县人，副教授。
  • 基金资助:
    黑龙江省自然科学基金(YQ2022E003)；中央高校基本科研业务费专项资金 (2572023CT21-04)。

Joint Optimization of Multi-line Bus Scheduling and Berth Settings with Overlapping Sections

HU Baoyu, LIU Wenlei, CHENG Guozhu*   

  1. College of Civil Engineering and Transportation, Northeast Forestry University, Harbin 150040, China
  • Received:2025-07-21 Revised:2025-08-26 Accepted:2025-09-01 Online:2025-12-25 Published:2025-12-24
  • Supported by:
    Natural Science Foundation of Heilongjiang Province (YQ2022E003)；Fundamental Research Funds for the Central Universities (2572023CT21-04)。

PDF

摘要： 为解决因多条公交线路重叠运行所引发的拥堵问题，本文提出一种提高多线路运行效率的公交车辆调度与泊位设置联合优化方法。以车辆成本和乘客等待时间成本的加权和最小，以及重叠区间泊位数量最小为目标，综合考虑时刻表、公交运行和分离式站点泊位分配等约束，建立多目标联合优化模型，并通过鲁棒优化方法重构模型。设计混合自适应大邻域搜索算法(MODE-ALNS)求解，多目标差分进化算法(MODE)负责全局搜索和生成初始解，自适应大邻域搜索算法(ALNS)负责对多目标差分进化算法生成的解进行局部优化。通过两者结合，实现全局搜索与局部优化的有机结合，提升解的质量。最后，以哈尔滨市6条重叠运行的公交线路为例进行案例分析。结果表明，相较现状方案，优化后的方案在车辆成本和乘客等待时间成本上减少了3.17%和7.19%，重叠区间内泊位总数的优化效果可达16.67%，能同时顾及运营商和乘客利益。与确定性模型相比，鲁棒优化模型能以更低的总成本和更稳定的泊位数量应对不同的扰动场景，验证了模型的优越性。

关键词: 城市交通, 联合优化, 自适应大邻域搜索算法, 区间重叠, 泊位优化, 鲁棒优化

Abstract: To address the congestion resulting from multiple overlapping bus routes, this paper proposes a joint optimization method for bus vehicle scheduling and bay allocation to enhance the operational efficiency of multi-route. It is established with the goals of minimizing both the weighted sum costs of vehicle and passenger waiting time, and the number of bays in the overlapping sections. The model, which considers constraints such as timetables, bus operations, and berth allocation at separated stops, is reformulated using a robust optimization approach. A hybrid adaptive large neighborhood search algorithm (MODE-ALNS) is designed for the solution. The multi-objective differential evolution (MODE) algorithm is responsible for global search and generating initial solutions, while the adaptive large neighborhood search (ALNS) algorithm is responsible for locally optimizing the solutions generated through the multi-objective differential evolution algorithm. By combining the two, an organic integration of global and local searches is achieved to improve the solution quality. A case study of six overlapping bus routes in Harbin shows that the optimized plan reduces vehicle costs and passenger waiting time costs by 3.17% and 7.19% respectively, compared to the current scenario. It also achieves a 16.67% reduction in the total number of bays within the overlapping section, benefiting both operators and passengers. The robust optimization model proves its superiority over the deterministic model by its ability to cope with various perturbation scenarios with a lower total cost and a more stable number of berths.

Key words: urban traffic, joint optimization, adaptive large neighborhood search algorithm, overlapping interval, berth optimization, robust optimization

中图分类号: