A modified social force model with different categories of pedestrians for subway station evacuation

Abstract

With the extensive construction of rail tunnels, the subway has gradually become the most widely used means of public transport. As important transportation hubs, the evacuation capacity of subway stations is especially significant. An accurate prediction model can efficiently reduce the evacuation time, and the social force (SF) model is one of the most widely used evacuation models. Almost all past research based on the SF model has focused on a single model for pedestrians. However, in this research, based on different force models, a crowd is divided into three categories that have different kinds of movement based on different force models. The first category of pedestrians, namely luggage-laden pedestrians, is characterized by a lower velocity and large radii. The second category of pedestrians is ordinary pedestrians. The final category of pedestrians comprises panic pedestrians, who choose an angle at which to overtake the pedestrians in front of them. Simulations are performed based on different proportions of these three categories of pedestrians, and it is found that with the increase of the proportions of luggage-laden and panic pedestrians, the evacuation time is also increased; this phenomenon could not be observed with the original SF model. The performance of the modified SF model is then tested in a passage with a bottleneck, and it is found that the interaction between panic pedestrians and luggage-laden pedestrians greatly increases the difficulty of evacuation. The number of luggage-laden pedestrians is found to be the controllable factor that leads to an increase of the evacuation time at the evacuation bottlenecks. Therefore, solutions like warning signs at evacuation bottlenecks should be implemented.