A Queueing Framework and Analysis of Observed One-Day Travel and Activity Patterns

Recognizing that travel is a demand derived from the need to participate in activities, planners and engineers are increasingly analyzing observed data on travel and activity decisions over varying timeframes. This paper proposes analyzing and modeling activity scheduling from the perspective an M/G/1 queuing system with vacations and service-in-random-order (SIRO). From this perspective, individuals behave like a service facilities, completing arriving activities based on a perceived stress, or the anticipated activity workloads relative to the individual’s own abilities to complete activities. Vacations or “rest periods” are initiated for activities, with fixed start and stop times, such as medical appointments, school classes or other activities. This approach is motivated by interest in evaluating the impacts of latent activities that have been generated, but not completed within an observational timeframe. Observed one-day activity schedules from a conventional travel survey are analyzed under the assumptions of an M/G/1 queuing system with vacations to illustrate some of the insights gained from this perspective. Overall, the length of queues and their composition vary by socio-demographic attributes. Consistently in-home activities comprise a significant proportion of activity queues, daily, with infrequent errands, such as shopping for durables, making up a smaller proportion. Retirees and homemakers show the highest utilization or busy proportions, relative to students and employed individuals. This paper concludes with a discussion of implications of this modeling approach for investigating stress, underscoring the necessity of accounting for dynamic activity decisions.