Abstract: Aiming at the poor local optimization ability and easy premature convergence of the particle swarm optimization (PSO) algorithm in evacuation simulation, a crowd evacuation model based on the improved PSO algorithm is proposed. On the basis of the original model, factors such as the forward driving force, exclusion driving force and panic degree of evacuees are introduced. According to the environmental factors and crowd movement, evacuees' direction is dynamically adjusted, and tends to the ideal exit under the optimal exit selection mechanism. The parametric simulation experiment of evacuation efficiency is carried out for the physical factors such as the exit location. The experimental results show that the evacuation model is effective, and can avoid local congestion, provides reasonable reference to the architectural design and management.

Key words: evacuation simulation, PSO algorithm, driving force, panic level, export choice