2023 Volume 78 Issue 5 Pages I_1-I_20
This paper reviews recent methodological developments in discrete choice analysis and discusses related issues for their application to transport modeling and infrastructure planning. We mainly focus on two research directions: (1) user heterogeneity and (2) concepts of perturbed utility. For heterogeneity, we summarize some mixture models derived from mixed multinomial logit models and semi-nonparametric structures, especially for empirical applications of estimating the distributions of users’ marginal valuation of attributes. Regarding perturbed utility, generalized entropy, convex conjugate, and demand inversion would be the key concepts for comprehensively understanding the various aspects of discrete choices, including bounded rationality. Finally, we argue possible future research directions of choice modeling.
