Influencing factors for exit selection in subway station evacuation.

• A social force model-based exit choice with pedestrian classification was proposed. • Simulations in the multiple exit evacuation scenarios were performed. • The standard deviation of pedestrian density in multiple exits reflects evacuation time. • The luggage-laden pedestrian has the most significant impact on the evacuation time. • A small number of wavering pedestrians improved the utilization efficiency of exits. As important urban transportation hubs, the subway stations with a massive flow of people require more evacuation exits. This study combined the exit choice model and the social force model with pedestrian classification. Three influencing factors were discussed in calculating the utility value: pedestrian density, distance from pedestrian to exits, and exit width. The effects of distance and density weight coefficient in the exit choice model on evacuation time are explored. The results show that with an increase in the density weight coefficient of the luggage-laden pedestrian, the evacuation time initially decreases and then increases. The evacuation time was shortened as the pedestrian density's standard deviation at exits decreased, and the evacuation time was extended otherwise. Evacuation time was compared based on different proportions of the pedestrian categories. The results show that the number of luggage-laden pedestrians is the main factor that leads to an increase in evacuation time. Scenarios with different exit widths are discussed. The evacuation time decreases with an increase in exit width, and the evacuation time tends to be stable when all exits have a width of 2.5 m. [ABSTRACT FROM AUTHOR]