Fitting S-shaped activity utility functions based on stated-preference data

Abstract

The theory of the S-shaped utility functions that was developed by Joh et al. (2003) within the framework of their activity-based transport demand model Aurora, will be used in the present research as the underlying framework of a dynamic activity-based rescheduling approach. The aim of this paper is to examine the impact of different activity characteristics and situational attributes, namely activity duration, history, discretionary time, time of day, location and accompanying persons, on an individual’s rescheduling decisions. In order to study the effect of these variables, different stated preference scenarios were included in a dynamic Internet-based survey that was provisionally tested on a small-scale pilot sample. The data gathered by the survey were used to estimate the fraction-utilities of the various attributes by means of a binary logit model.