The Kinetic Model Of Multi-agent Crowd Dynamics

In this paper,we try to simulate the dynamic characteristics and rules of multi-agent particle population in different types of regions by using the kinetic model.The four basic scenarios of group interactions are unidirectional flow,bidirectional flow,merging flow and cross flow.According to the physical layout of doors and walls,the moving bounded areas can be simplified as the following types:straight corridor type areas,T-shaped areas and crossroads type areas.Kinetic modeling in mesoscopic scale is used to model the different types of interactions that affect particle motion using game theory tools.The interaction between particles is mainly divided into two parts.One is the particle-particle interaction,which shows that particles tend to follow the same type of group and find areas with smaller local density to avoid crowd in the process of movement;Another is the interaction between the particles and the geometry of the region,manifested by the urgency of the exit and the tendency to avoid obstacles and walls.In the simulation process,we use vector field to represent the interaction of geometric types of particles.Because of the different distributions of obstacles,walls and exits in different types of areas,different vector fields are obtained.In this paper,the operator splitting method is used to decompose the problem into pure horizontal flow problem and the problem involving interaction.Through numerical simulations,the self-organizing behaviors of several particle populations are obtained.