高铁枢纽与城市轨道交通换乘流线仿真与优化  

Simulation and optimization of pedestrian transfer flow in high-speed railway hubs and urban rail transit systems

作　　者：程龙[1,2] 宁哲 薛小钰 张霁扬 刘志鹏 CHENG Long;NING Zhe;XUE Xiaoyu;ZHANG Jiyang;LIU Zhipeng(School of Transportation,Southeast University,Nanjing 211189,China;Laboratory of Transport Industry of Com-prehensive Transportation Theory(Nanjing Modern Multimodal Transportation Laboratory),Ministry of Transport,Nanjing 211135,China;BYD Auto Co.,Ltd.,Xi’an 710100,China;China Electronics System Technology Cooperation,Beijing 100040,China;Planning and Natural Resources Bureau of Wenjiang District,Chengdu 611130,China)

机构地区：[1]东南大学交通学院,南京211189 [2]综合交通运输理论交通运输行业重点实验室(南京现代综合交通实验室),南京211135 [3]比亚迪汽车有限公司,西安710100 [4]中电云计算技术有限公司,北京100040 [5]成都市温江区规划和自然资源局,成都611130

出　　处：《北京交通大学学报》2024年第4期43-52,共10页JOURNAL OF BEIJING JIAOTONG UNIVERSITY

基　　金：国家自然科学基金(52072066,52372301,U20A20330);综合交通运输理论交通运输行业重点实验室(南京现代综合交通实验室)开放课题资助课题(MTF2023006);东南大学至善青年学者研究基金(2242023R40027)。

摘　　要：针对高铁枢纽与城市轨道交通换乘流线优化问题,以南京南站作为研究对象,建立全过程仿真模型,识别换乘瓶颈,采用梯度提升决策树法(Gradient Boosting Decision Tree,GBDT)确定参数重要度,提出改善方案.首先,分解换乘过程的行人与设施流线,分析高速铁路到达客流与城市轨道交通客流的分布特征.其次,使用AnyLogic软件建立高铁枢纽换乘城市轨道交通的全过程仿真模型,分析现状仿真结果,识别空间瓶颈.然后,设计不同优化类型下的措施参数及调整范围,形成不同参数组合方案.采取梯度提升决策树算法,识别不同措施参数的相对重要度,并据此确定改善措施的优先级.最后,依据措施的优先级,确定不同类型下的优化组合方案,选择机器学习全局可解释性方法对其进行优化效果分析,为不同场景下的服务改善提出建议.研究结果表明:换乘瓶颈主要集中于楼/扶梯通道设施以及闸机、售检票机等服务设施处;乘客换乘城轨的购票比例对平均换乘时间和单位时间最大换乘人数均起到重要影响,对于平均换乘时间,城轨自动售票机数量、购票时间、城轨进站服务时间、城轨进站闸机数量的影响相对较大,对于单位时间最大换乘人数,购票时间与城轨进站闸机服务时间的影响相对较大.为提高高铁枢纽换乘效率,建议推广电子客票和多种支付方式,优化购票及检票设施.This study addresses the issues related to the optimization of transfer flows between highspeed rail hubs and urban rail transit,focusing on Nanjing South Railway Station as the case study.A full-process simulation model is developed to identify transfer bottlenecks,using the Gradient Boosting Decision Tree(GBDT) to determine parameter importance and propose improvement measures.Firstly,the pedestrian and facility flows of the transfer process are decomposed to analyze the distribution characteristics of passenger flows from high-speed rail to urban rail transit.Secondly,AnyLogic is employed to create a full-process simulation model of the transfer at high-speed rail hubs,analyzing the current simulation results to identify spatial bottlenecks.Then various optimization measures are designed under different types,adjusting parameters to formulate combination schemes.The GBDT algorithm is used to ascertain the relative importance of different measures and parameters,establishing the priority of improvement measures.Finally,based on this prioritization,different types of optimization combination schemes are determined,with machine learning interpretability methods applied to analyze their effectiveness,providing recommendations for service enhancements across different scenarios.Results show that transfer bottlenecks are primarily located at stair/escalator passageways and service facilities such as gates and ticket machines.The proportion of urban rail ticket purchases significantly influences both average transfer time and the maximum number of transfers per unit time.For average transfer time,the number of automatic ticket machines,ticket purchase time,urban rail entry service time,and the number of entry gates have a relatively large impact.For maximum transfers per unit time,ticket purchase time and entry gate service time are most influential.To improve transfer efficiency at high-speed railway hubs,it is recommended to promote electronic tickets and various payment methods,and optimize ticketing and checking

关 键 词：综合交通系统 交通枢纽 换乘全过程优化 AnyLogic仿真 梯度提升决策树 

分 类 号：U115[交通运输工程]