交通运输系统工程与信息

交通运输系统工程与信息 ›› 2026, Vol. 26 ›› Issue (2): 72-80.DOI: 10.16097/j.cnki.1009-6744.2026.02.007

• 综合交通运输体系 • 上一篇    下一篇

城市暴雨内涝下考虑地铁联合疏散的公交应急路线优化

阎桑慧宇1 ，马瑞2 ，李健*1   

  1. 1. 同济大学，道路与交通工程教育部重点实验室，上海201804；2.纽约大学，坦登工程学院，德州253100，美国
  • 收稿日期:2025-11-30 修回日期:2026-01-13 接受日期:2026-02-05 出版日期:2026-04-25 发布日期:2026-04-20
  • 作者简介:阎桑慧宇（1995—），女，河南新乡人，博士生。
  • 基金资助:
    国家重点研发计划(2023YFC3804004)。

Robust Optimization of Emergency Bus Route Design Under Urban Pluvial Flooding with Metro-Bus Joint Evacuation

YAN Sanghuiyu1, MA Rui2, LI Jian*1   

  1. 1. Key Laboratory of Road and Traffic Engineering of Ministry of Education, Tongji University, Shanghai 201804, China; 2. Tandon School of Engineering, New York University, New York TX 253100, USA
  • Received:2025-11-30 Revised:2026-01-13 Accepted:2026-02-05 Online:2026-04-25 Published:2026-04-20
  • Supported by:
    National Key Research and Development Program of China (2023YFC3804004)。

PDF

摘要： 气候变化导致城市暴雨内涝频发，引发道路积水，公交绕行、跳站或停运，对居民活动出行带来严重影响。本文针对强降雨导致的公交运营中断问题，提出一种考虑地铁-公交协同疏散的行程时间鲁棒优化建模方法，该方法基于路线行程时间预算不确定性集进行刻画。首先，基于多源数据构建了考虑多重因素的公交速度数据集；其次，利用基于循环结构的时序变分自编码器结合k-中心点聚类方法，得到有限个具有代表性的速度场景，并将速度场景结合路线长度转化为行程时间场景，对不确定性集合进行刻画；最后,构建基于路线预算不确定性的鲁棒优化模型，并将模型重构为线性模型。研究结果表明，在应急响应中引入地铁-公交协同可显著缩短乘客出行时间，与传统公交绕行策略相比，协同方案可节省约15%~20%的出行时间降幅；相较于公交区间运行策略，降幅进一步扩大至约50%~60%。所构建的基于路线的预算不确定性集合可有效反映路段尺度的暴雨扰动特征，其中，中等预算水平能在鲁棒性与成本之间取得较优平衡。本文提出的方法可为城市暴雨内涝下公交紧急疏散方案提供辅助决策支持。

关键词: 城市交通, 应急疏散优化, 鲁棒优化, 公交线路设计, 暴雨内涝

Abstract: Climate change has led to increasingly frequent urban pluvial flooding, resulting in road inundation, which subsequently forces buses to detour, skip stops, or suspend services, thereby severely disrupting residents' daily mobility. This study investigates travel time robust optimization for emergency bus route planning with metro-bus joint evacuation under severe rainfall-induced service interruptions. A robust optimization approach based on route-dependent budgeted uncertainty sets is proposed. A multi factor bus speed dataset is constructed using multi-source data. Representative speed scenarios are generated through a Recurrent Temporal Variational Autoencoder combined with k-medoids clustering, and these speed scenarios are converted into travel-time scenarios based on route length to characterize the uncertainty set. A robust optimization model with route-dependent budgeted uncertainty set is then formulated and reformulated into a linear model. The results indicate that incorporating metro-bus coordination into emergency operations can substantially reduce passenger travel time. Compared with conventional bus detour strategies, the coordinated scheme yields a travel time reduction of approximately 15%~20%, and compared with bus short-turn strategies, the reduction expands to approximately 50%~60%. The proposed route-dependent budgeted uncertainty sets capture rainfall-induced variability at the route scale, and it was found that moderate budget levels achieve a favorable balance between robustness and operational efficiency. The proposed framework can support decision-making for emergency bus evacuation planning under urban pluvial flooding.

Key words: urban transportation, emergency evacuation optimization, robust optimization, bus route design, pluvial flooding

中图分类号: