ABSTRACT
This chapter provides an overview of the development of different discrete choice models that occurred after the appearance of the multinomial logit (MNL), Nested Logit (NL), and multinomial probit (MNP) models. It presents an in-depth examination of models that fall within the class of Generalized Extreme Value (GEV) models, namely two-level models that belong to the generalized nested logit (GNL) family and multi-level models that belong to the Network GEV family. It also draws heavily on the work by Sethi, Koppelman and Coldren. The development of discrete choice models has been a very active area of research. Dozens of models have been developed, several that were specifically motivated by airline applications. Models that belong to the GEV class can relax the independence of irrelevant alternatives (IIA) assumptions by including covariance terms that are created by allocating alternatives to two or more nests. The chapter includes the product differentiation (PD) model, and the weighted nested logit (WNL) model.
