Development of generalized cost model for private cartrip makers under traffic information

Abstract

To formulate rational improvement proposal under traffic information, it is necessary to develop a generalized cost (GC) model of trip makers' route selection process under the effect of traffic information. The paper demonstrates a methodology for developing GC models for private car trip makers, when travel time is projected as instantaneous traffic information along with its variation level on a roadside variable message sign (VMS) board as a measure of influencing trip makers' route choice behavior. As an empirical study, two competitive traffic corridors in Kolkata Metro City, India, are considered. A stated preference (choice based) experiment is carried out and both multinomial logit (MNL) and random parameter logit (RPL) techniques are employed for the estimation of coefficients of the relevant attributes. Considering a priori constrained T-distribution of random parameters, two different RPL models are explored: one without accounting heterogeneity and the other with accounting heterogeneity around the mean of random parameter(s).