基于AFC数据的耦合时空特征下地铁高峰客流预测研究  

A Study on the Prediction of Metro Peak Passenger Flow Based on AFC Data with Coupled Temporal and Spatial Features

作　　者：宋丽梅 SONG Li-mei(Yangling Vocational and Technical College,Yangling,Shaanxi 712100,China)

机构地区：[1]杨凌职业技术学院,陕西杨凌712100

出　　处：《杨凌职业技术学院学报》2023年第1期14-17,共4页Journal of Yangling Vocational & Technical College

基　　金：杨凌职业技术学院2021年院内基金项目“城市轨道交通网络高峰客流拥挤管控研究”(ZK21-38);陕西省职业技术教育学会2022年度教育教学改革研究课题(2022SZX380);杨凌职业技术学院2022年人文社会科学(教改)研究基金项目(GJ22035)。

摘　　要：高峰客流拥挤是大城市轨道交通面临的主要问题,表现为客运服务水平降低,存在客流安全问题。高峰客流中主要客流为以工作出行的客流,这类客流具有出行稳定、强度高等特征,从而出现高峰客流拥挤问题。本文利用自动售检票数据所包含的乘客出行内在关系,构建基于规则的乘客工作出行客流辨识算法;从不同方面挖掘工作出行客流特征,以西安地铁1号线为例,分析了地铁客流的耦合时空特征,发现当高峰客流拥挤时,其呈周期性重复特征因素包含产生位置、时间及持续时间等,从而提取了影响地铁客流变化主要因素,包括节日、非节日、时间段、站点和天气,从而提高地铁客流预测的准确性。Peak passenger congestion is the main problem faced by rail transit in large cities,which reduces the level of passenger service and leads to passenger flow safety problems.The main passenger flow in the peak passenger flow is the passenger flow of work travel,which has the characteristics of stable travel and high intensity,resulting in peak passenger congestion problems.This paper uses the intrinsic relationship of passenger travel contained in the automatic ticket vending data to construct a rule-based passenger flow identification algorithm for passenger work travel.Taking Xi’an Metro Line 1 as an example,the coupled spatiotemporal characteristics of subway passenger flow are analyzed,and it is found that when the peak passenger flow is crowded,its periodic repeating characteristic factors include the location,time and duration,etc.,so as to extract the main factors affecting the change of subway passenger flow,including festivals,non-festivals,time periods,stations and weather,so as to improve the accuracy of subway passenger flow forecasting.

关 键 词：城市轨道交通 高峰客流拥挤 客流辨识算法 客流预测 

分 类 号：U231.2[交通运输工程—道路与铁道工程]