A New Crowd Simulation Method by Using Multiagent on Cellular Automata

Tomoki Hamagami, Seiichi Koakutsu and Hironori Hirata
Chiba University

Abstract:

In this paper, we describe a new simulation method of crowd behavior,method that uses a two-layer model that consists of multiagent (MA) framework and cellular automata (CA). Features of our method are as follows: (1) Complicated crowd behavior is realized by autonomous actions of heterogeneous agents. (2) It is possible to separate autonomous action process and restrictions of physical interference. Based on the two-layer model, we developed a simulation system, and analyzed various behavior of crowd. In particular, the case of collision of counter crowd was minutely examined. Through some simulation experiments, the following interesting results were obtained. (1) When crowd consists of agents that had homogeneous property, crowd made ``whirlpools,'' ``waves,'' and ``holes,'' and agent's walking speed became slow. (2) When crowd consists of agents that had heterogeneous property, crowd formed ``files,'' and efficiently flowed. By using two-layer model, it has been realized to be able to simulate various patterns of combination of multiagent property and environmental structure.