How reasonable is the typical week modeling time frame approach in activity-based travel demand modeling?: Case study on shopping activity, trip using multi-week trip diary survey data

Abstract

This paper investigates the potential of using a ‘typical week’ time frame in activity-based travel demand modeling. This paper concentrates on a specific activity/trip: shopping, using a 6week travel diary survey, MobiDrive. Each week of the MobiDrive survey is considered as a random week and models are compared. Weekly shopping trips are modeled using a multivariate ordered probability modeling approach. Each model has two major components: a deterministic and a stochastic component. The deterministic component accommodates various variables and the stochastic component captures the interday correlations within the random week. The estimated parameters of the deterministic component reveal different behavioural patterns and the independent nature of the individual days within a week. On the other hand the interday correlation patterns of the random weeks reveal that a typical week does represent a minimum cycle of shopping activity behaviour.