Mean-field game analysis of crowd evacuation using the Cristiani-Santo-Menci method.

Abstract

In this paper, we investigate the crowd evacuation from a room on the basis of the mean field game theory. In particular, effects of the predictability of pedestrians on the evacuation process are considered by applying the Cristiani-Santo-Mensi method as a solver of the Hamilton-Jacobi-Bellman equation coupled with the Fokker-Planck equation. Some numerical tests of the crowd evacuation from a room are implemented in order to investigate effects of the predictability, mass diffusion, interactive force and domain, and form of the running cost function on the evacuation process. Finally, we investigate effects of the predictability on the evacuation process, when two exits are opened and closed, alternatively.