Modern Transportation and Metallurgical Materials

This study focuses on the optimization of emergency evacuation efficiency in rail transit stations, with an emphasis on exploring the influence of heterogeneous behaviors of partnered passengers on the evacuation process. Based on the theory of social force model, AnyLogic simulation platform is used to construct the evacuation model of a typical urban rail station in Chongqing, and the influence of the partnering ratio and the waiting and turning back behaviors of the partnering pedestrians on the evacuation efficiency is systematically analyzed through the control variable method. It is found that the proportion of companionship is significantly negatively correlated with the evacuation efficiency, and the evacuation time is prolonged by 27.5% under the proportion of 50% companionship. The influence of companion pedestrians' waiting behavior grows significantly with the expansion of evacuation scale. Different spatial locations of the folding back behavior produces differentiated flow conflicts, in which the gate area is narrowed due to the narrow channel, resulting in the reduction of the evacuation efficiency by 37.4%. Targeted recommendations are made based on the station's emergency evacuation plan and facility layout.