摘要
为更加精准刻画动态随机客流与列车时刻表编制间的复杂匹配过程,综合考虑乘客动态随机到达过程、列车容量限制、车内拥挤度对乘客上下车速率影响、停站时间动态调整和列车运行安全防护等重要运营特征,构建城市轨道交通列车运行离散仿真模型。在此基础上,建立以最小化乘客平均候车时间为目标的随机非线性优化模型,并设计一种基于仿真的有限差分随机逼近算法,用于优化城市轨道交通发车方案。以国内某条线路的实际运营数据为例验证仿真模型及优化方法,结果表明所构建模型及算法具有较好的优化效果和计算效率,优化后的列车实绩运行时刻能更好地适应客流需求的动态随机性,且在不增加发车频次的前提下有效降低乘客平均候车时间,提升城市轨道交通运营管理水平。
To more accurately depict the complex matching process between dynamic random passenger flow and train capacity,this paper comprehensively considered important operational characteristics such as passenger dynamic random arrival process,train capacity constraints,carriage congestion effects on passenger boarding and alighting rates,dynamic station dwell time adjustments,and train operation safety protection,and constructed a discrete simulation model for the operation of urban rail transit trains.Based on this,a random nonlinear optimization model was built to minimize average passenger waiting time,and a simulation-based finite difference stochastic approximation algorithm was designed to optimize the departure plan of trains.Taking the actual operational data of an urban rail transit line in China as an example to verify the simulation model and optimization methods,the results show that the constructed model and algorithm have effective optimization and computational efficiency.The optimized actual train operation time can better adapt to the dynamic randomness of passenger demand,significantly reducing the average passenger waiting time without increasing the frequency of train services,which is of great value for further improving the quality of urban rail transit operation and management.
作者
张雨洁
闫海峰
骆泳吉
朱蕾
施润宇
ZHANG Yujie;YAN Haifeng;LUO Yongji;ZHU Lei;SHI Runyu(School of Transportation and Logistics,Southwest Jiaotong University,Chengdu 611756,Sichuan,China)
出处
《铁道运输与经济》
北大核心
2024年第5期161-170,共10页
Railway Transport and Economy
基金
国家自然科学基金项目(72101218)
四川省科技计划资助项目(2023NSFSC1033,2023NSFSC0906)。
关键词
城市轨道交通
动态随机客流
列车时刻表
系统仿真
有限差分算法
Urban Rail Transit
Dynamic Random Passenger Flow
Train Timetable
System Simulation
Finite Difference Algorithm
