INFORMATION-BASED EVACUATION EXPERIMENT AND ITS CELLULAR AUTOMATON SIMULATION

Abstract

The evacuation process under emergency is studied by means of experiments and simulations, focusing on the influence of the environment information. A revised cellular automaton model in which environment information is considered as "static information" (building structure, spatial distance, etc.) and "dynamic information" (sounds of fire alarm, etc.) is introduced. Two scenarios, including evacuation with and without visibility in a classroom, are studied to investigate the different influence of the two kinds of information on human behavior. The experimental and simulation results demonstrate that: (1) to intensify the spatial distance information can reduce the evacuation time; (2) the spatial distance is not the only decisive factor especially in evacuation without visibility because the sound information, which is ignorable in evacuation with visibility, is playing a more important role under this condition; (3) the intensity of static information can reflect evacuees' familiarity of the environment; (4) the model can reproduce the experiments well, and the simulation method is useful for further study of the crowd movement simulation.