Estimation of Travelers' Values of Time Using a State-Preference Survey with Variable Pricing Options

摘要

This study analyzed data from a stated preference survey of Houston travelers faced with numerous mode choices, including value pricing options. The study: (a) examined the possibility of using a genetic algorithm to estimate mode choice models while removing the need of making the IIA assumption, (b) estimated nested logit models, (c) attempted to estimate random parameter logit models, and (d) estimated numerous multinomial logit models. After comparing different specifications and optimization techniques (namely the genetic algorithm and Newton-Raphson method in the econometric software, Limited Dependent LIMDEP) the multinomial logit model estimation using LIMDEP was found to be more efficient because of easy estimation and a much lower time requirement for estimation. Hence a multinomial logit model was used for estimating the values of travel time savings (VTTS) and the penalty for changing travel schedule for different groups of travelers. The values estimated for the penalty for changing travel schedule were not statistically significant and were therefore not used. The values estimated for travel time savings were significant and comparable to those obtained in previous studies.