Data assimilation and agent-based modelling: towards the incorporation of categorical agent parameters

This paper explores the use of a particle filter—a data assimilation method—to incorporate real-time into an agent-based model. We apply method to simulation real pedestrians moving through concourse Grand Central Terminal in New York City (USA). The results show that filter does not perform well due (i) unpredictable behaviour some and (ii) because optimise categorical agent parameters are characteristic this type This problem only arises experiments real-world pedestrian movement data, rather than simulated, hypothetical as is more common. point potential solution involves resampling variables particle, such locations agents space, but keeps other agents’ choice destination. research illustrates importance including provides proof concept for application improved obstacles solutions discussed have important implications future work focused on building large-scale models.