城市轨道交通系统的高效运营对于缓解城市交通压力、提升居民出行效率具有重要作用。随着城市化进程的加速，轨道交通系统面临着客流需求的快速增长，这对系统的运营效率和服务质量提出了更高的要求。本文针对城市轨道交通系统在客流需求迅速增长背景下的运营挑战，开展了列车时刻表优化研究，旨在通过科学的方法提升系统的整体运营效率和服务水平。在客流分析方面，本文深入分析上海市轨道交通客流数据，识别了车站客流与断面客流的时空分布特征。通过计算进站客流与断面客流，详细分析了客流的不均衡性，并探讨了不同车站类型和周边用地性质对客流分布的影响。这些分析为优化工作提供了重要的数据支持和理论基础。在短时客流预测模型的研究中，本文创新地结合了ARIMA和LSTM两种不同技术路线的预测模型，构建了一种ARIMA-LSTM组合预测模型。该模型充分利用了两种模型的优势，通过在上海地铁1号线的实际算例进行预测，对比了ARIMA、LSTM以及ARIMA-LSTM组合模型的预测效果。通过多尺度误差（RMSE、MAPE、MAE）等指标检验和相关系数（R2）检验，实验结果表明ARIMA-LSTM组合模型在预测精度和泛化能力上均优于单独的ARIMA和LSTM模型，证明了组合方法在城市轨道交通客流预测场景下的优越性。本文提出了一个多目标优化模型，旨在最小化乘客平均等待时间和列车能耗。该模型综合考虑了非成对发车、满载率、乘客候车时间等多个因素，并通过NSGA-II算法进行求解。在解决多目标优化问题的过程中，除了传统的遗传操作，本文引入了快速非支配排序和拥挤度计算的关键步骤，以实现对初始种群的有效评估和选择。不仅有助于保留优秀基因，而且确保了种群的多样性和解的均衡性。在实证分析中，本文以上海地铁1号线为案例，运用所建立的优化模型和元启发式算法进行求解，对生成的Pareto最优解集进行了详细分析。优化结果表明，相较于现状时刻表，优化后的时刻表在乘客平均等待时间上减少了21.4%，在列车能耗上降低了9.3%，有效提升了服务质量并降低了运营成本。此外，本文还对比了安排不同车次对优化目标的影响，优化方案为列车不成对开行方案，实验结果表明有助于降低满载率，从而显著提升乘客的乘车体验。

The efficient operation of urban rail transit systems plays an important role in alleviating urban traffic pressure and enhancing the travelling efficiency of residents. With the acceleration of urbanisation, the rail transit system faces a rapid growth in passenger demand, which puts higher demands on the operational efficiency and service quality of the system. This study addresses the operational challenges of urban rail transit systems in the context of rapid growth in passenger demand, and carries out a study on train schedule optimisation, aiming to improve the overall operational efficiency and service level of the system through scientific methods.In terms of passenger flow analysis, this study analyses the passenger flow data of Shanghai‘s rail transit in depth and identifies the spatial and temporal distribution characteristics of the passenger flow at stations and the cross-section passenger flow. By calculating the inbound passenger flow and cross-section passenger flow, this paper analyses the imbalance of passenger flow in detail, and explores the impact of different station types and the nature of the surrounding land on the distribution of passenger flow. These analyses provide important data support and theoretical basis for the optimisation work.In the study of short-term passenger flow prediction model, this paper innovatively combines two different technical lines of prediction models, ARIMA and LSTM, and constructs a combined ARIMA-LSTM prediction model. The model makes full use of the advantages of the two models, and compares the prediction effects of ARIMA, LSTM and ARIMA-LSTM combination model by predicting the actual cases in Shanghai Metro Line 1. Through the indicators such as multi-scale error test (RMSE, MAPE, MAE) and correlation coefficient (R2) test, the experimental results show that the combined ARIMA-LSTM model is superior to the separate ARIMA and LSTM models in terms of prediction accuracy and generalisation ability, which proves the superiority of the combined method in urban rail transit passenger flow prediction scenarios.In this paper, we propose a multi-objective optimisation model that aims to minimise the average passenger waiting time and train energy consumption. The model integrates several factors such as unpaired train departure, full load rate, and passenger waiting time, and is solved by the NSGA-II algorithm. In the process of solving the multi-objective optimisation problem, in addition to the traditional genetic operations, this paper introduces the key steps of fast non-dominated sorting and congestion calculation to achieve effective evaluation and selection of the initial population. It not only helps to retain the excellent genes, but also ensures the diversity of the population and the equilibrium of the solution.In the empirical analysis, this paper takes Shanghai Metro Line 1 as a case study, and uses the established optimisation model and meta-heuristic algorithm to solve the problem, and the generated Pareto optimal solution set is analysed in detail. The optimisation results show that compared with the status quo timetable, the optimised timetable reduces the average waiting time of passengers by 21.4%, and reduces the energy consumption of trains by 9.3%, which effectively improves the service quality and reduces the operating costs. In addition, this paper also compares the impact of different train schedules on the optimisation objective, and the optimised scheme of not running trains in pairs is shown to help reduce the full load rate, thus significantly improving the passenger experience.