地铁作为一种重要的公共交通方式，其规模在近几十年间快速发展。随着城市化的推进，不同地域之间有更加频繁的人员流动。部分线路客流呈现明显的潮汐特征。此外，许多地铁线路运营收入小于运营成本，依赖政府财政支持。因此，构建一个改善乘客体验、降低列车运行成本的时刻表优化方法至关重要。然而，目前关于潮汐客流场景下列车时刻表优化方法的研究仍有不足。 本文探究了潮汐客流场景下地铁单条线路的非对称时刻表优化问题。在车数有限、考虑再生能量利用的情况下，通过优化列车发车间隔、大小交路计划、长短车计划，提高乘客服务水平，降低列车能耗。本文的主要工作内容如下： （1）为了应对潮汐客流，建立了包括时刻表框架、列车数量方程、客流行为方程在内的非对称时刻表模型，设计了人均等待时间、车公里数两个优化目标，构建了一个混合整数优化问题。在案例分析中，探究了客流分布、车数限制对优化结果的影响，分析了影响乘客体验、列车能耗的重要因素。 （2）为了构建更灵活的调度策略，建立了考虑长短车、大小交路调度策略的综合优化方法。在案例分析中，阐释了长短车、大小交路调度策略的节能原理。结果表明长短车与大小交路策略可以显著提高地铁列车能效。 （3）为了提高再生能量利用率，建立了再生制动能量利用模型、超级电容储能装置利用模型，设计了列车能耗的评价方法，构建了考虑再生能量的非对称时刻表优化问题。实验表明，在应用再生制动能量利用技术、超级电容储能技术后，基准方案的能耗水平有显著的下降。 （4）为了提高收敛速度，本文分析了时刻表模型的结构，结合梯度下降与启发式搜索方法，构建了针对时刻表优化问题的求解算法。通过数值实验，证明了相较于传统的MOPSO、NSGA-II进化算法，该算法收敛速度更快、解的分布更好。这对实时调度具有重要意义。
As an important public transportation, urban rail transit has been developing rapidly in recent decades. The urbanization process has resulted in more passenger flows between different geographical areas. Thus tidal flow of passengers has become more usual in urban rail transit. In addition, many metro lines show high levels of energy consumption and rely on government financial support. Therefore, it is crucial to construct a timetable optimization method that improves passenger experience and reduces train operation costs. However, there is still a lack of asymmetric timetable optimization methods that consider regenerative energy utilization in tidal passenger flow scenarios. In this thesis, we investigate the asymmetric timetable optimization problem for a single line of urban rail transit in a tidal flow scenario. With limited number of rolling stocks and the use of regenerative energy, the passenger experience and the energy efficiency are improved by optimizing the train headway, fleet size, and full-length and short-turn services. The main work of this paper is as follows. (1) In order to cope with tidal passenger flow, an asymmetric schedule model including assymetric schedule framework, train number equation, and passenger flow behavior equation is established, and two optimization objectives of waiting time per capita and train mileage are designed to construct a mixed integer optimization problem. In the case study, the effects of passenger flow distribution and train number limitation on the timetable are explored, and the important factors affecting passenger experience and train energy consumption are analyzed. (2) In order to construct a more flexible dispatching strategy, a comprehensive optimization method considering fleet size, and full-length and short-turn services is established. The energy-saving principles fleet size, and full-length and short-turn services are analyzed through numerical experimental. And the results show that fleet size, and full-length and short-turn services can significantly improve the energy efficiency. (3) In order to improve regenerative energy utilization, a regenerative braking energy utilization model and a super capacitor energy storage device utilization model are established. The evaluation method of train energy consumption is designed, and an asymmetric schedule optimization problem considering regenerative energy is constructed. The experiments show that the energy consumption level of the benchmark scheme is significantly reduced after applying regenerative braking energy utilization technology and supercapacitor energy storage technology. (4) To improve the convergence speed, we analyzes the structure of the timetable model and constructs an algorithm by combining gradient descent and heuristic search methods. Through numerical experiments, it is demonstrated that the algorithm converges faster and has better solution distribution compared with the traditional MOPSO and NSGA-II evolutionary algorithms. This has important implications for real-time scheduling.
