Citation

Dianat L, Habib KMN, Miller EJ. Transp. Res. A Policy Pract. 2020; 133: 337-352.

Copyright

(Copyright © 2020, Elsevier Publishing)

DOI

10.1016/j.tra.2020.01.017

PMID

unavailable

Abstract

A dynamic, gap-based activity scheduling model is developed for predicting out-of-home non-work/school (NWS) episodes over a day. In the developed model, work/school, and night sleep are assumed to be pre-determined, thereby providing a daily "skeleton schedule". NWS episodes are then simultaneously generated and scheduled in the available gaps as a joint activity type and destination choice, followed by a continuous time expenditure choice. The model is built on a subset of the Transportation Tomorrow Survey (TTS) collected in the Greater Toronto and Hamilton Area (GTHA) in 2001. The developed model is validated on another sample from the TTS 2001 and is also applied to forecast individuals' schedules for the years 2006 and 2011, for which observed TTS data are also available. This study, which is rarely conducted in the literature, examines the model's capability to replicate the base year schedule and predict the activity patterns of the future years, which is the ultimate purpose of any travel demand model. Simulation outcomes of the three years follow similar trends to each other. Replication of the base year's schedule is more accurate than the future years; however, there are no significant changes in the accuracy of the outcomes of the model's application on all the three years.

Language: en

Keywords

Activity-based modelling; Microsimulation; Non-work/school activity scheduling; Skeleton schedule; Travel demand forecasting