Abstract
Mixed Logit model is developed recently as a strong tool which can approximate arbitrarily any random utility models. However, the flexibility and the applicability have not been well examined. This paper aims to clarify the attributes and summarize the relationship between the Mixed Logit model and other developed methods. We test its flexibility from the following aspects: 1) error structure among alternatives, 2) unobserved heterogeneity and 3) random coefficient model. Some simulation data or actual travel data are used for the model estimations. The results showed that the Mixed Logit model has high flexibility. Finally, we summarize its applicability and the further topics to be studied.
