Are you an EPFL student looking for a semester project?
Work with us on data science and visualisation projects, and deploy your project as an app on top of Graph Search.
In practice, most operational activity-based models have focused on single-day analyses. This common simplifying assumption significantly limits the models' behavioural realism, as they cannot adequately capture the dynamics and processes involved in the scheduling of activities over multiple days. Decisions taken daily are affected by both habits built over time and forward-looking behaviour, where individuals decide based on the expected outcomes of future decisions. A person's activity/travel planning behaviour depends on their behaviour on other days of the week and that there are two main components to this dynamic behaviour: 1. same-day and next-day substitution effects for activities and trips, and 2. latent propensity to engage in some activities or choose a specific transportation mode. In this semester project, we investigate the existence of these components and how to integrate them into a multiday activity scheduler (OASIS). More specifically, we will attempt to model these unobserved influences in the context of a latent choice model.
Daniel Kuhn, Andreas Krause, Yifan Hu, Jie Wang