Analysis of impact of group walking patterns on pedestrian evacuation

An understanding of the walking patterns of groups of pedestrians in an evacuation is critical for the establishment of policies, procedures, and organizational structures to respond effectively to emergencies. Groups of pedestrians compose a crowd in which pedestrian motions are significantly constrained to maintain cohesion. On the basis of behavior theory, this paper proposes a multiagent model for the simulation of crowds of pedestrians. The main innovative aspect of this model is the genuine representation of the patterns of movement of groups of pedestrians. Patterns of movement consisting of the line-abreast pattern, the chain pattern, and the mixed pattern were investigated, and their influences on evacuations were evaluated quantitatively by taking into account the discrepant densities, disparate distributions of the proportions of pedestrian groups of different sizes, and heterogeneous velocities of groups of pedestrians. The simulation results show that the walking patterns of groups of pedestrians have a significant influence on the dynamics of pedestrian evacuation. The chain pattern was safer when the time of evacuation under high-density conditions was considered, and the mixed pattern had a better performance under moderate-density conditions. Moreover, the influence of patterns of movement was distinct with different distributions of pedestrian groups of different sizes; the chain pattern had the highest evacuation efficiency among the three patterns of pedestrian movement. In addition, a homogeneous velocity condition had a higher evacuation efficiency than a heterogeneous velocity condition. Thus, a chain pattern with a homogeneous velocity is recommended as the optimal pattern of movement in pedestrian evacuations when the safety and efficiency of plans and designs for the evacuation of pedestrian traffic with the different patterns of movement are considered.