区域轨道交通系统作为城市圈的核心交通骨架，凭借其多样化的轨道交通制式构筑起了一个综合的交通网络，该网络对推动城市圈的经济一体化发挥着不可或缺的重要作用。为了确保整个区域轨道交通网络的高效运行，针对当前系统中车流与客流协同优化的不足，以及系统运营的复杂性和不确定性带来的挑战，车流与客流的协同鲁棒优化显得尤为重要。针对这些现存问题，本文深入研究了区域轨道交通系统车流和客流协同鲁棒优化方法，主要研究内容如下：1）构建了基于Anylogic的区域轨道交通系统仿真模型。通过对区域轨道交通系统各个子系统的建模需求进行详细分析，设计了仿真模型的各个类和模块并在Anylogic中进行实现，为区域轨道交通系统车流和客流协同鲁棒优化方法的数学建模提供支撑，并为鲁棒优化方法的验证提供仿真环境。2）考虑客流不确定性，以区域轨道交通系统全局运能风险为优化目标，在客流分布已知的条件下，提出了一种基于均值方差理论的区域轨道交通系统车流和客流协同鲁棒优化方法。在客流分布未知的条件下，引入跳站策略作为优化变量以提升优化模型的优化效果，并提出了一种区域轨道交通系统跳站策略的鲁棒优化方法；进一步应用鲁棒跳站策略，提出了一种区域轨道交通系统车流和客流协同分布鲁棒优化方法，并引入先验信息对分布鲁棒优化方法进行了改进。3）考虑模型参数的不确定性，分析了优化模型中存在的各类模型参数的不确定性，基于全局灵敏度分析方法对各个不确定性参数进行灵敏度分析并进行参数筛选。以区域轨道交通系统全局运能风险和系统列车能耗为优化目标，在参数分布已知的条件下，提出了一种车流和客流协同随机优化方法，在参数分布未知的条件下，提出了一种车流和客流协同分布鲁棒优化方法。基于Anylogic仿真环境对本文提出的方法进行了验证。设计了包括单点突发客流、全线大客流和随机客流等多样化客流需求场景以及随机模型参数场景。探讨了在这些场景下，列车调度与客流引导协同鲁棒优化方法的应用效果，仿真结果表明，本文提出的多种鲁棒优化方法在不确定性场景下对系统全局运能风险和系统列车能耗有较好的优化效果和稳定性。
The regional rail transit system is a comprehensive transportation network composed of multiple rail transit modes within the urban circle, which plays an important role in the economic integration of the urban circle. To ensure the efficient operation of the entire transportation network, the collaborative optimization of train flow and passenger flow is particularly important. Meanwhile, considering the complexity and uncertainty of the system is crucial for enhancing the robustness of the system. Based on the above issues, this thesis studies the collaborative robust optimization methods of train scheduling and passenger guidance in regional rail transit systems, and the main research contents are as follows:1)An Anylogic-based simulation model of the regional rail transit system was developed. Through detailed analysis of the modeling requirements of each subsystem in the regional rail transit system, various classes and modules of the simulation model were designed and implemented in Anylogic, supporting the mathematical modeling of the collaborative optimization of train flow and passenger flow, and providing a simulation environment for the verification of? robust optimization methods.2)The uncertainty of passenger flow was considered. With global transport capacity risk as the optimization objective, under the condition of known passenger flow distribution, a collaborative robust optimization method based on mean-variance theory for train flow and passenger flow in the regional rail transit system was proposed. Under the condition of unknown passenger flow distribution, the skip-stop strategy was introduced as an optimization variable to improve the optimization model‘s performance, and a robust optimization method of the skip-stop strategy in the regional rail transit systems was proposed. Furthermore, by applying the robust skip-stop strategy, a collaborative distributionally robust optimization method for train flow and passenger flow in the regional rail transit system was proposed, and prior information was introduced to improve the distributionally robust optimization method.3)The uncertainty of model parameters was considered. The uncertainties of various model parameters in the optimization model were analyzed, and global sensitivity analysis methods was used to perform sensitivity analysis and parameter selection. Taking the global transport capacity risk and train energy consumption of the regional rail transit system as optimization objectives, under the condition of known parameter distribution, a stochastic optimization method for collaborative optimization of train flow and passenger flow was proposed, and under unknown parameter distribution, a distributionally robust optimization method for collaborative optimization of train flow and passenger flow was proposed.The proposed methods were verified within the Anylogic simulation environment. Various passenger demand scenarios, including single-point burst flows, full-line heavy flows, and random flows, as well as random model parameters scenarios, were designed. The application effects of the collaborative robust optimization method for train scheduling and passenger flow guidance under these scenarios were discussed, confirming the effectiveness and stability of the proposed robust optimization models in reducing global transport capacity risk and train energy consumption under uncertainty.
