Pougala, J., Hillel, T., and Bierlaire, M.

From one-day to multiday activity scheduling: extending the OASIS framework

Speaker: Pougala Janody

STRC 2023

May 10, 2023

Applications of activity-based models for the estimation of transport demand have demonstrated to achieve greater behavioural realism than traditional trip-based models. However, state-of-the art models focus on single-day schedules as focal points to their estimations, thus ignoring fundamental dynamics that explain individual behaviour over longer periods of time. Several have highlighted the importance of multiday analyses in activity-travel contexts, which are still lacking in many state-of-the-art framework. In this paper, we present an extension of the OASIS framework (Pougala et al., 2022), an integrated model for the simulation of single day schedules, to include intrapersonal interactions influencing longer term decisions. We formulate the multiday problem as a multiobjective optimisation problem where each day d is associated with a utility Ud. We consider an activity-based set up where individuals maximise the total utility of their schedules over multiple days (e.g. week). We discuss implications and requirements of this formulation, and illustrate the methodology with an application on the MOBIS dataset (Molloy et al., 2021), an activity and travel survey capturing several weeks of behaviour for each respondent.