接驳地铁的社区公交微循环系统优化研究

发布时间：2018-09-03 20:25

【摘要】：地铁作为城市公共交通的骨干,有着快速、可靠性强等地面公交无法比拟的优点,目前已成为大城市居民的主要出行方式之一。而社区公交具有机动、灵活的特点,可弥补地铁线网可达性较差、无法提供门对门服务的缺点,在公共交通系统中起着不可替代的微循环作用。本文通过对社区公交路径、协同时刻表以及基于需求-响应的灵活公交等方面的优化研究,旨在为与地铁相接驳的优化的社区公交微循环系统设计提供理论支持,目的是方便乘客到达地铁站点,减小乘客出行成本并兼顾公交企业利益,实现综合交通系统中多方式公共交通的高效换乘。论文主要工作和创新点总结如下：(1)构建了两类社区公交路径优化模型。第一类模型基于现实路网,同时针对路段定义了公交潜在需求指标,并以最大化潜在需求为目标建立模型,同时考虑线路旅行时间约束；第二类模型基于半现实路网,并以最小化总成本(乘客出行成本与企业成本)为目标建立路径优化模型,模型中嵌套了一种站点布设的启发式算法与最优发车间隔求解算法。对于两类模型的求解,首先设计了深度优先搜索(Depth-first Search, DFS)算法对所有可行解进行遍历：之后设计了一种改进的遗传算法(Genetic Algorithm, GA)对两类问题进行求解,并分别通过相关实例与算例验证两种算法。通过对比两种算法的求解结果与运行时间,证实了GA在求解该问题上具备可行性与高效性。同时也深入分析了线路长度和最大允许步行距离对相关成本及发车间隔的影响。(2)在车次数与车队规模给定的前提下,以乘客出行成本(计划延迟成本与换乘成本)为目标函数建立协同时刻表优化模型,并同时考虑两类约束条件,即车辆载荷能力约束与车队规模约束。针对前者,将其视为软约束并在原目标函数基础上追加超载惩罚函数：对于后者,设计了一种考虑车队规模约束的时刻表生成机制。在求解方面,首先利用GA进行求解；之后设计了一种基于Frank-Wolfe算法并结合发车时间调整的求解算法(Frank-Wolfe Algorithm combined with a Heuristic Algorithm of Shifting Departure Times, FW-SDT)。通过相关算例与实例分析证实了两种算法在求解该问题上均具备可行性,而数据实验及相关灵敏度分析表明FW-SDT在求解效率、准确性及稳定性方面均优于GA。(3)考虑了一种基于需求-响应型(Demand-Responsive Transit, DRT)的更加灵活的公交形式,并将其引入与地铁相接驳的社区公交系统设计中,对其优化运行进行研究。在建模方面,以最小化总成本(运营成本与乘客在车成本)为目标建立模型,同时考虑服务时间窗、乘客在车时间、车辆载荷以及车辆最大走行时间等现实约束。在求解方面,设计了禁忌搜索算法(Tabu Search, TS)与基于可变邻域的模拟退火算法(Variable Neighborhood Search based Simulated Annealing, VNS-SA)。为验证两种算法,设计了基于现实路网的数值实验。为在优化结果与计算效率之间进行合理权衡,在该数值实验下分别应用不同算法及内部算法组合进行计算,并对相关结果进行了对比分析。
[Abstract]:Subway, as the backbone of urban public transport, has many advantages, such as high speed and high reliability, and has become one of the main travel modes of urban residents. Community bus has the characteristics of mobility and flexibility, which can make up for the shortcomings of poor accessibility of subway line network and unable to provide door-to-door service in public transport system. This paper aims to provide theoretical support for the design of optimized microcirculation system of community bus connecting with subway, aiming at facilitating passengers to reach subway stations and reducing passengers by optimizing community bus routes, cooperative timetables and flexible bus based on demand-response. The main work and innovations of this paper are summarized as follows: (1) Two types of community bus routing optimization models are constructed. The first model is based on the real road network, and the potential demand index is defined for the road segment, and the maximization is achieved. The second model is based on the semi-realistic road network and aims at minimizing the total cost (passenger travel cost and enterprise cost). In the model, a heuristic algorithm for site layout and an optimal heuristic interval algorithm are nested. First, the Depth-first Search (DFS) algorithm is designed to traverse all feasible solutions. Then an improved genetic algorithm (GA) is designed to solve the two types of problems, and the two algorithms are verified by relevant examples and examples. The results show that GA is feasible and efficient in solving this problem. At the same time, the effects of line length and maximum allowable walking distance on the related costs and departure intervals are analyzed in depth. (2) The passenger travel costs (planned delay costs and transfer costs) are taken into account when the number of vehicles and the size of the fleet are given. A cooperative timetable optimization model is established based on the scaling function, and two kinds of constraints, i.e. vehicle load capacity constraints and vehicle size constraints, are considered simultaneously. Firstly, GA is used to solve the problem. Then, a Frank-Wolfe algorithm combined with a Heuristic algorithm of Shifting Departure Times (FW-SDT) is designed to solve the problem. The two algorithms are verified by relevant examples and case analysis. The data experiment and sensitivity analysis show that FW-SDT is superior to GA in solving efficiency, accuracy and stability. (3) Considering a more flexible form of public transport based on Demand-Responsive Transit (DRT), FW-SDT is introduced into the design of community public transport system connected with subway. In the aspect of modeling, the objective is to minimize the total cost (operation cost and passenger in-vehicle cost), and the practical constraints such as service time window, passenger in-vehicle time, vehicle load and vehicle maximum travel time are considered. Variable Neighborhood Search based Simulated Annealing (VNS-SA). To verify the two algorithms, a numerical experiment based on real road network is designed. In order to make a reasonable trade-off between the optimization results and the computational efficiency, different algorithms and combinations of internal algorithms are applied to the numerical experiment, and the results are compared with each other. The related results are compared and analyzed.
【学位授予单位】：北京交通大学
【学位级别】：博士
【学位授予年份】：2015
【分类号】：U491.17

【相似文献】