Citation

Alizadeh H, Farooq B, Morency C, Saunier N. Transp. Res. F Traffic Psychol. Behav. 2019; 64: 171-183.

Copyright

(Copyright © 2019, Elsevier Publishing)

DOI

10.1016/j.trf.2019.05.009

PMID

unavailable

Abstract

Previous route choice studies have mostly focused on the effect of observable factors, such as route attributes and socio-economic characteristics, on individuals' decisions. However, route choice decisions might not be exclusively dependent on these observable variables, but also on latent variables, which cannot be directly observed and measured. Also, the latent behavioral heterogeneity among the population has mostly been ignored by assuming that all the individuals in the sample population have similar attitudes, perceptions, and lifestyles. In this paper, we present a comprehensive framework to explicitly incorporate latent behavioral constructs as well as segment heterogeneity based on a probabilistic segmentation of the population. We apply the proposed framework to compare the route choice behavior of frequent versus occasional drivers. An Integrated Choice and Latent Variable model is used to bring in the role of the underlying behavioral constructs, while a Latent Class model accounts for taste heterogeneity across the two segments of our sample population. An Extended Path-Size Logit model is adopted as the choice model component and a Metropolis-Hastings based algorithm is used to generate route alternatives. Data is collected through a web-based survey designed to collect behavioral data on drivers' route choices, using psychometric indicators and behavioral questions on respondents' perceptions and attitudes.

RESULTS confirm that the inclusion of latent variables and latent heterogeneity across population segments significantly improve the explanatory power of the choice model, and illustrate how the route choice behavior of frequent car users is different from that of occasional ones.

Language: en

Keywords

Behavioral heterogeneity; Integrated choice and latent variable models; Latent class models; Revealed preference data; Route choice modelling; Web-based survey