基于两阶段模型的空铁联运客流量预测方法——以沪蓉通道为例  

Prediction Method of Air-rail Transportation Passenger Flow Based on Two-stage Model:Taking Shanghai-Chengdu Corridor as an Example

作　　者：郭九霞[1] 田金玉 钟庆伟 陈曲 GUO Jiuxia;TIAN Jinyu;ZHONG Qingwei;CHEN Qu(School of Air Traffic Management,Civil Aviation Flight University of China,Guanghan 618307,Sichuan,China;Flight Planning Department,Civil Aviation Administration Operation Monitoring Center,Beijing 100710,China)

机构地区：[1]中国民用航空飞行学院空中交通管理学院,四川广汉618307 [2]民航局运行监控中心运行监控处,北京100710

出　　处：《科技和产业》2025年第5期82-87,共6页Science Technology and Industry

基　　金：四川省社科基金(SCJJ23ND186);中央高校基本科研业务费专项(PHD2023-041);民航教育人才类项目(MHJY2023010);中央高校教育教学改革专项(E2024024)。

摘　　要：随着国内综合交通立体枢纽的快速建设与发展,空铁联运模式给旅客出行带来了更多便利,准确地掌握空铁联运客流量对提升综合运输服务质量及保障运输安全至关重要。以城市群间空铁联运为背景,采用粒子群优化-随机森林(PSO-RF)模型和Logit模型相结合的两阶段模型,对出行路径客流量进行预测。以单向“空转铁”为实例。第1阶段,基于历史数据通过PSO-RF模型预测出行线路的日均民航客流量;第2阶段,对离港旅客进行行为调查,分析旅客换乘方式的特征及其选择行为,基于非集计理论构建旅客出行选择的二元Logit模型,并计算换乘高铁的分担率;最后,综合两阶段模型的结果,计算出行线路客流预测量。以上海至成都周边城市群构建算例,对所提出方法的有效性和可行性进行验证。结果表明,两阶段模型的准确率达到80.40%。With the rapid construction and development of comprehensive transportation hubs,the air-rail intermodal transportation model has brought more convenience to passengers’travel.Accurately understanding the air-rail intermodal passenger flow is crucial for improving the overall transportation service quality and ensuring transportation safety.A two-stage model combining particle swarm optimization-random forest model(PSO-RF)and Logit model is used to predict the passenger flow of travel paths in the context of air-rail intermodal transportation between urban agglomerations.Take one-way“Shifting from Railways to Aviation”as an example.In the first stage,the average daily civil aviation passenger flow of travel routes was predicted based on historical data by PSO-RF model.In the second stage,a behavioral survey was conducted through the airport outbound passengers to analyze the characteristics of passengers’transfer mode and choice behavior.Then,a Binary Logit model of passenger travel choices was constructed based on the disaggregate theory,and the high-speed rail transfer sharing rate was calculated.Finally,the results of the two-stage model were combined to calculate the travel route passenger flow forecasts.The effectiveness and feasibility of the proposed method by constructing a case study from Shanghai to Chengdu neighboring urban agglomeration were validated.The results indicate that the accuracy of the two-stage model reaches 80.40%.

关 键 词：综合运输 空铁联运 客流预测 粒子群优化-随机森林(PSO-RF)模型 LOGIT模型 

分 类 号：U15[交通运输工程]