Value of time determination for the city of Alexandria based on a disaggregate binary mode choice model

Abstract

In the travel demand modeling field, mode choice is the most important decision that affects the resulted road congestion. The behavioral nature of the disaggregate models and the associated advantages of such models over aggregate models have led to their extensive use. This paper proposes a framework to determine the value of time (VoT) for the city of Alexandria through calibrating a disaggregate linear-in parameter utility-based binary logit mode choice model of the city. The mode attributes (travel time and travel cost) along with traveler attributes (car ownership and income) were selected as the utility attributes of the basic model formulation which included 5 models. Three additional alternative utility formulations based on the transformation of the mode attributes including relative travel cost (cost divided by income) and log (travel time) and the combination of the two transformations together were introduced. The parameter estimation procedure was based on the likelihood maximization technique and was performed in EXCEL. Out of 20 models estimated, only 2 models are considered successful in terms of the parameters estimates correct signs and the magnitude of their significance (t-statistics value). The determination of the VoT serves also in the model validation. The best two models estimated the value of time at LE 11.30/hr and LE 14.50/hr with a relative error of +3.7% and +33.0%, respectively, of the hourly salary of LE 10.9/hr. The proposed two models prove to be sensitive to trip time and income levels as factors affecting the choice mechanism. The sensitivity analysis was performed and proved the model with higher relative error is marginally more robust.