基于面板混合Logit模型的中长途高速客运方式分担率预测

针对高铁和民航在中长途高速客运中的方式分担率变化问题，考虑到不同旅客群体间的异质性及旅客在不同情景下选择的相关性，以旅客性别、年龄、月收入等为基本属性，程前程后费用和时间、旅程费用和时间作为特性变量，建立了面板混合Logit模型。利用NLOGIT软件，基于北京—南京客运通道旅客出行意向调查数据，分别对传统Logit模型、横截面混合Logit模型和面板混合Logit模型进行参数估计。模型估计结果表明,面板混合Logit模型具有更好的行为解释能力和预测精度。基于面板混合Logit模型，使用仿真方法预测了高铁和民航的方式分担率随旅程费用和时间的变化趋势，得到高铁分担率随旅程费用差的增大而增加、随旅程时间差的增大而减小，民航则相反，并且旅程费用差对方式分担率的影响比旅程时间差更加显著。

Prediction of the mode share ratio ofmiddle-and-long distance high-speedpassenger transport based on the Panel MixedLogit model

To solve the problem of mode share changeof high-speed railways and civil aviation in middle-and-long distance passengertransport, taking the heterogeneity among passenger groups and the relevance ofpassenger choices in different scenarios into consideration, the Panel MixedLogit(PML) model was established, which taking gender, age, and monthly incomeof passengers as basic property, the cost and time before ,after and during thetrip as specific property. NLOGIT was used to estimate the parameters ofTraditional Logit, Cross-sectional Mixed Logit, and Panel Mixed Logit modelsbased on the stated preference data of passenger in Beijing-Nanjing transportcorridor. Model estimation results show that the PML model has better behaviorinterpretation capabilities and prediction accuracy. Based on PML model, thesimulation method was used to predict the trend of mode share of high-speedrailways and civil aviation along with the changes of trip cost and time. Theresults show that the high-speed railways mode share ratio increases with the increaseof trip cost difference value and decreases with the increase of trip timedifference value; which is the opposite for civil aviation. It also shows thatthe trip cost difference value has a more significant effect on the mode sharethan the trip time difference value.