A new dynamic cellular automaton model for pedestrian evacuation process in stadium

This paper studies the pedestrian evacuation problem in stadium with or without obstacles using cellular automaton model. We give a multi-agent individual decision-making framework, in which the action direction of each pedestrian (called agent) is determined by the distance of the agent to the exits, the number and density of agents and obstacles within the view field of the agent. Different from the existing results, we divide all the pedestrians in the stadium into four classes: young male, young female, old male and old female. In evacuation process, the weighting that affects individual decision-making between each class of agents is different. In the simulation, we consider two cases: (I) two exits are both unclogged, and (II) one of the two exits is totally clogged. In both cases, two typical scenes are simulated: (a) there is no obstacle and the agents are distributed randomly, and (b) there are obstacles and the agents are put regularly. In addition, we further analyze the effect of the exit width to the evacuation efficiency. The simulation can reproduce exactly the actual pedestrian evacuation process. Therefore, the proposed method might be useful to assess public buildings design.