Albatross IV: from single day to multi time horizon travel demand forecasting

After years of silence, this paper informs the travel behavior community about progress in the development of the new version of the Albatross. It upgrades previous versions by including new mobility options such as electric cars, car sharing, and Mobility as a Service, and adds social networks to the synthesized population. More fundamentally, it transforms a single day forecasting system into a multiday dynamic forecasting system, in which long-term dynamics are linked to short term dynamics in activity-travel patterns. To that end, a dynamic synthetic population generator is developed, modules predicting behavioral change due to life trajectory events are added and an improved scheduler is implemented. In addition, the scope of the system is extended, linking activity travel participation and scheduling decisions to weather conditions and simulating energy and emissions consequences of activity travel behavior. Finally, Albatross IV uses parallel computing technology, significantly reducing computing times of the system. The paper systematically outlines the motivation behind each change/addition to the system, discusses the modeling approach and positions the changes in the literature in transportation research.