基于乘客自适应路径变更的公交网络协同调度优化设计研究

Bus schedule coordination (BSC) aims to provide fast transfer opportunities for multi-legged trips, while disruption management is the matched strategies to reduce missed connection. Conventional studies focus on scheduling optimization based on fixed passenger route choice, while neglecting the adaptive rerouting behaviour in the event of transfer failure. This makes basis schedule parameters (e.g., frequency, timetable, bus size) and disruption management invalid (e.g., surplus slack time), which reduces the efficiency of BSC..To overcome the drawbacks, the study proposes a new mechanism that passengers with multi-legged trips could update their route in case of missed connection, and builds the relationship between the rerouting probability and schedule design through the transfer reliability. By using choice modelling and network equilibrium, the transfer demand is re-assigned between alternative routes and the downstream original route in the event of transfer failure, which is then integrated into the stochastic optimization of BSC via bi-level modelling. Combining discrete event-based simulation analysis, we investigate the interaction among regular schedule parameters, disruption management, passenger re-assignment and the network structure, aiming to form a behaviour-driven BSC modelling framework. The results would improve the accuracy and efficiency of BSC and reduce the side-effect of disruption management.

公交网络协同调度旨在为乘客联程出行提供快速换乘的机制，干扰管理是不确定环境下减少换乘失效的配套措施，现有研究基于乘客固定路径选择探讨运营参数优化，忽略了乘客在换乘失效下的自适应路径变更行为，不仅使其基本技术参数（发车频率、时刻表、车型等）失真，还产生无效的干扰管理配置（如过剩的缓冲时间），致使公交网络协同调度效率未能充分释放。.针对现有研究的不足，本项目提出联程出行乘客在换乘失效时更新路径的新机制，通过换乘可靠性等属性关联乘客路径变更概率与公交调度决策，运用方式选择和网络均衡理论，将换乘客流在备选路径与原路径下游进行重分配，建立集成乘客行为的公交网络协同调度双层规划模型，结合离散事件仿真技术，探讨基本技术参数、干扰管理配置、客流重分配模式、线网结构的相互作用，形成基于乘客行为预测反馈的公交网络协同调度设计方法体系。研究成果有助于提高网络协同调度的准确性和效率，并降低干扰管理的负面效应。

公交网络协同调度旨在提高线网效率和衔接性，是时刻表优化的重要一环。公交网络由枢纽节点通过线路的相互连接而成，线网形态对乘客路线选择和客流分布存在很大影响，在实际运营中，出行者通常会根据车辆延迟到达情况和线网形态，动态地调整自己的换乘计划，而以往研究仍遵循“先客流分配，后优化调度”的框架，忽略了换乘失效时乘客的路径变更行为，从而高估换乘失效的客流量，造成线网客流空间分布信息失真，不仅会影响基本技术参数的准确性，还将造成干扰管理配置过于保守，例如产生无效或过剩的缓冲时间，从而降低调度的效率。. 本项目提出了联程出行乘客在换乘失效时更新路径的新机制，通过换乘可靠性等属性关联乘客路径变更概率与公交调度决策，运用方式选择和网络均衡理论，建立集成乘客行为的公交网络协同调度双层规划模型，探讨基本技术参数、干扰管理配置、客流重分配模式、线网结构的相互作用，形成基于乘客行为预测反馈的公交网络协同调度设计方法体系。研究成果有助于提高网络协同调度的准确性和效率，并降低干扰管理的负面效应。. 依托本项目申请人累计发表论文27篇，含1篇ESI高被引论文，以第一作者在国际顶级期刊发表论文6篇（Transportation Research Part B发表1篇，Transportation Research Part E发表2篇，IEEE Transactions on Intelligent Transportation Systems发表2篇，Expert Systems With Applications 1篇）。以第一或通讯作者在主流期刊（Transportmetrica A/B, Communications in Nonlinear Science and Numerical Simulation, 中国公路学报, 自动化学报, 控制理论与应用等）发表论文多篇。做学科重要会议特邀报告1次。以第一发明人获授权发明专利和软件著作权各1项；申请专利8项。