Investigating the influence of assigning a higher priority to scheduling work and school activities in the activity-based models on the simulated travel/activity patterns

Two dynamic, gap-based activity scheduling models are tested by applying a short-run microsimulation approach to replicate workers’ travel/activity patterns over a 1-week time period. In the first model, a two-level work episode scheduling model is applied to schedule weekly work episodes (Dianat et al. in Transp Res Rec 2664:59–68, 2017. https://​doi.​org/​10.​3141/​2664-07). This includes joint choices of working or not on each day and work episode duration and start time in case of choosing to work. Assigning higher priority to scheduling work episodes, and assuming night sleep to be pre-determined, provides a weekly “skeleton schedule”. Non-work/school (NWS) episodes are then generated and scheduled in the available gaps as a joint choice of activity type and destination followed by a continuous time expenditure choice. The second model applies the same mathematical framework as the NWS model for scheduling all activity types including work/school, considering only night sleep as the pre-determined skeleton schedule. This exercise allows us to study the impact of assigning a higher priority to scheduling work/school activities on complete out-of-home travel/activity pattern prediction, compared to the alternative hypothesis, which is scheduling all the activities simultaneously. Comparing the simulation outcomes of the two models with the observed dataset reveals that organizing NWS episodes around the schedule skeleton not only is behaviorally more representative but also increases the accuracy of the predicted NWS episodes’ patterns. Moreover, applying the work scheduling model results in a more accurate prediction of the weekly work schedule compared to the second model.