ANDA, a new agent-based model for pedestrian dynamics

Over the past couple of years, we have given our best to develop a new branch of agent-based models for pedestrian dynamics, with a structure that would more closely follow the actual processes at play. In particular, we were eager to delineate the factors pertaining to decision-making and their mechanical consequences on the motion of pedestrians.

Our manuscript introducing this new modelling framework is finally out, in Transportation Research Part C. You can also get the permission to download the associated algorithm, written in C++, by filling out the form here. This model, dubbed ANDA, has been shown to reproduce an unprecedentedly broad range of crowd situations, often in a quantitative way. Even more, it can replicate salient features associated with digital distraction by smartphones, a growing concern for pedestrian safety.