Estimation of Travel Choice Models with Randomly Distributed Values of Time

The value of time is a key concept in transport planning in terms of the economic valuation of travel time savings and the relative importance of time versus cost in travel forecasting models. A standard method for deriving values of time is to use the tradeoff ratio implied by the time and cost coefficients estimated in travel choice models. In actual choice situations, it is impossible to observe all the factors that affect the relative importance of time and cost. Thus, a method for estimating discrete travel choice models was derived and demonstrated with a randomly distributed value of time. In the case studies considered, significant improvements in model fit were obtained when distributed values of time were allowed. In prediction, more realistic responses were found by using the distributed value of time model than by using models with a fixed value of time.