On group evacuation behavior of subway station halls using an improved evolutionary game model

Abstract

This study is dedicated to examining the impact of stress on group evacuation behaviors in urban subway station fires, with the ultimate goal of improving evacuation efficiency and enhancing public safety. The authors have introduced an advanced evolutionary game model that combines “selfishness coefficient” and “stress threshold.” This introduces the concept of stress threshold and contributes to the exploration of stress propagation within groups. The simulation results indicate that the degree of selfishness in a crowd significantly affects evacuation efficiency. The emergence and propagation of stress emotions fundamentally represent a large-scale shift in individual evacuation behavior strategies within a group. At the onset of evacuation, the majority of the crowd follows an orderly and organized evacuation process, which occurs before the emergence of group stress. When the overall system benefit falls below the stress threshold, evacuees choose to crowd due to negative emotions and their pre-existing behavior strategies, leading the entire system into a “stress state”. Hence, effective strategies must simultaneously consider both physical methods and stress levels, lowering the threshold for orderly evacuation. For example, measures such as circular broadcasting and sending mobile messages can be employed to promptly inform evacuees about the on-site situation, and staff cab be deployed for guidance and comfort to improve people's sense of security. Analyzing evacuation patterns reveals factors influencing efficiency and individual behaviors, contributing insights to evacuation protocols and applications.