An Agent-Based Model for Simulating Human-Like Crowd in Dense Places

Crowd simulation has been becoming an efficient tool to study the crowd behaviour and its movement. Compared with macroscopic model, microscopic model is able to generate a fine grain simulation result. This paper proposes a agent-based model for crowd simulation for dense area. The simulation model considers how agent selects a goal as its moving destination, how avoids collision with neighboured agents, how leads or follow a group of agents, and how avoids high dense area to reduce its traveling time. Simulation results show the proposed agent-based model is able to simulate agent navigating and move around dense simulation environment. The simulation performance is also efficient.