Transportation policy research assesses the effect of policy changes such as the imposition of a parking charge or the augmentation of passenger rail service, on individual travel behavior. Over the last few decades, Conditional Lo it Model has been strongly advocated among discrete choice models for transportation policy study due to its ease of estimation and realistic outcomes.

This dissertation analyzes the necessity of the entry of constraints in the indirect utility function. These functions are used in discrete choice models and form the basis for travel demand modeling. This dissertation proposes a hypothesis that the explicit entry of constraints in the systematic utility term will improve the specification of logit model because the constraints have significant effect on individual mode switching. A new model for incorporating constraints is developed based upon the hypothesis. Simulation and empirical study are combined to analyze the validity of this model's structure. The empirical evidence shows that explicit entry of constraints can effectively correct the errors caused by missing or ignoring of constraints in the indirect utility term.