Analysis Of The Travel Behavior Of Ride-Hailing Users Based On Nested Choice Models

In the era of Internet + sharing economy,online ride-hailing is favored by a large number of travelers with its characteristics of convenience,comfort and economy,and has gradually become an indispensable part of urban transportation.The emergence and development of online ride-hailing is having a profound impact on the urban transportation system.Therefore,the in-depth understanding and guidance of the travel behavior of online ride-hailing users may become one of the important focal points to promote the coordinated development of the urban transportation system.This study in 2018 in Dalian network about car users as the survey object,the network about car users travel choice SP design,and through the RP survey collection respondents personal social and economic attributes,relying on the python program compiled respectively based on travel mode and travel time cross nest selection model(Cross-nested Logit,CNL),based on the travel time nest selection model(Nested Logit,NL)and NL model based on the way of travel.Through comparative analysis of the similarities and differences among the three,the CNL model with high goodness of fit and strong explanatory power was selected for sensitivity analysis,including the elasticity analysis of the characteristic variables of online ride-hailing and the prediction of sharing rate change.On this basis,the pollutant emission simulation and evaluation were combined with the estimation results.the result shows that:(1)The CNL model has stronger explanatory ability and higher goodness of fit than the two NL models,which can more accurately capture the differences and correlation between the selected limbs and have better explanatory ability for the travel behavior of online ride-hailing users.(2)in terms of travel behavior,choose limb attributes and personal economic and social attributes have significant influence on travel choice behavior,including travel time,travel cost,net about car premium and discount is an important factor affecting the network about car use,the higher the personal income level,the younger the age,the purpose of travel strict time requirements travelers tend to choose network about car travel.(3)The sensitivity analysis based on the CNL model shows that the sensitivity of travelers to the travel cost,travel time,premium and discount of online ride-hailing services decreases successively.On this basis,the pollutant emission is measured by five modes of transportation: online taxi,taxi,private car,subway and bus.