基于需求-响应的辅助公交运行优化研究

This project explores the following two new types of Demand-Responsive Paratransit that own a degree of flexibility between the conventional transit and private cars: (1) the feeder bus serving passengers depart from the same origin while arrive at different destinations in a particular area; (2) the flexible-route bus fulfilling one-off reserved trip demands. Before the optimization research, the project intends to acquire the characteristics of transit demands and the transit service bottlenecks in big cities by analyzing the smart card data and GPS data of the transit, and conducts routing/network design and optimal timetable development of feeder buses serving in a particular area. For the second type of paratransit, based on the newest achievements of the dial-a-ride problem and the vehicle routing problem, by considering multi-vehicle types, the vehicle routing optimization of this type of paratransit can be conducted; then by considering reality factors, such as time-varied reservation information and dynamic travel time of segments, the project aims at developing bus routing optimization methods and intelligent vehicle dispatch strategies that response the time-varied traffic information fast. By introducing the above two types of paratransit, the project intends to construct a new configuration of transit systems, to carry out some optimization researches by combining optimization theory, operational research, and artificial intelligence algorithms comprehensively, to improve the basic theory of the comprehensive urban transit system, and thus to lay the foundation of easing traffic jams and improving the whole level of transit service.

本课题重点探索两种灵活性介于常规公交与私家车之间的需求-响应型辅助公交形式：（1）服务于特定区域内具有不同起点、相同终点乘客的接运公交；（2）服务于一次性预约需求的灵活线路公交。优化研究前，课题拟通过分析公交和地铁IC卡及公交GPS数据得到大城市公交出行需求特征与公交服务瓶颈点，进而针对特定区域开展接运公交路径、线网优化及协同时刻表优化设计。针对第二类形式，通过借鉴拨号叫车服务与车辆路径优化等问题的研究成果，融入多车型变量，对车辆路径进行优化；之后考虑实时更新的预约需求及动态路段旅行时间等因素，探索能够快速响应动态信息的公交路径优化方法及智能化调度策略。课题试图通过引入以上两种辅助公交形式，构建一个新的城市公共交通系统框架，并综合运用优化理论、运筹学及人工智能算法，对其运行优化展开研究，旨在深化和完善城市综合公交系统基础理论，为改善交通拥堵、提升公交系统整体服务水平奠定基础。

本研究从多方式公共交通出行的多维度感知与数据提取入手，对个体公共交通出行链进行精确追踪与提取，并将出行链与公交车次进行匹配，以实现综合考虑乘客个体候车、在车、换乘、满载率等不同角度的复合公交网络评价与服务瓶颈点辨识。. 基于辨识结果，首先提出了配合地铁的地面公交线网调整方法；针对末端公交服务水平偏低、协同化程度不高的问题，提出接驳地铁的固定线路辅助公交路径优化方法，该方法以最小化乘客出行成本与企业运营成本之和为目标，同时嵌入站点优化布设算法与最优发车间隔确定算法，以得到接驳地铁的微循环公交优化开设方案（站点布设方案、优化路径方案、发车间隔与车队规模配备方案）；在此基础上，同时考虑车辆载荷能力与车队规模限制，提出了配合地铁的微循环公交协同时刻表优化编制方法，并设计多种算法实现给定车次数下的发车时刻表快速求解。. 最后，本研究将基于预约需求的灵活线路公交引入综合公交系统中，并对其静态路径生成与动态协同调度进行优化研究。在确定静态路径的基础上，考虑现实中路段旅行时间的实时变化，提出两种公交实时调度策略，即基于拥堵的车辆绕行策略与车辆替代策略，并详细介绍了两种策略的生成思路与算法流程，最后设计算例证明了两种策略的有效性。本研究相关成果可为掌握个体公交出行机理、支撑公交决策提供夯实的理论基础，也为推动公共交通一体化发展，实现综合交通系统中多方式公共交通的高效换乘提供技术支持。