Research On The Influence Of Individual Heterogeneity On Passengers’ Choice Of Transfer Behaviors

The high-speed rail station serves as a meeting point for urban external and internal traffic.Its concept is to provide passengers with "zero distance" transfer conditions.Passengers’ preferences for transfer modes determine the internal layout of passenger terminals.In recent years,with the increase in the number of ride-hailing users,high-speed rail stations have added ride-hailing interchange points.However,the size and selection preferences of ride-hailing users are uncertain.At the same time,in order to promote public travel and reduce traffic congestion in the city,it is also important to understand and analyze the preferences of passengers with available private vehicles leaving the station.For a long time,research on choice behavior has focused on the hypothesis of homogeneous rational people,while ignoring the impact of individual heterogeneity.Therefore,this article takes high-speed rail passengers as the research object and explores the passenger’s transfer preference by constructing multiple logit models that do not consider heterogeneity and latent class models that consider heterogeneity.Calculate willingness to pay in order to provide a reference for market segmentation and fare formulation.First of all,analyze the factors affecting passenger transfer.From the perspectives of passenger attributes,travel attributes,and transfer mode attributes,select gender,age,occupation,income,education,travel purpose,travel distance,luggage load,peer care,and travel costs,waiting and walking time,and in-vehicle time are the main variables.Secondly,the questionnaire was selected based on the different design scenarios of selecting limbs.Using the D-optimal design method,two design options,"disposable private car transfer options" and "non-disposable private car transfer options" were obtained,with 18 choices for each design.Based on the responses of 495 passengers,the effects of the choice distribution and travel attributes of the two designs were compared.Finally,a multiple logit models without considering heterogeneity are used as a comparison group,and compared with the latent class models considering heterogeneity from the parameter estimation results,significance,and model suitability indicatorsThe results show that the latent class models refine passengers with disposable private car transfer options away from the station into four categories,and passengers with non-disposable private car transfer options into three categories,giving classification probabilities and passenger transfer choice habits and individual characteristics.However,multiple Logit models can only reflect average preferences.From the perspective of model fitting goodness,the AIC and BIC indicators of the latent class models are better than the multiple logit models,and the former fits better.From the perspective of willingness to pay,the latent class models provide the payment preference of each type of passengers for each transfer mode under different travel attributes,which can provide more accurate reference recommendations for market segmentation and fare formulation.