Studying the dynamics of crowd panic propagation during emergency evacuation

Abstract

Casualties during emergency evacuations are often attributed to people’s panic-driven extreme behaviors rather than the accidents themselves. The propagation of panic is influenced by various factors. Based on the susceptible–infectious–recovered–susceptible (SIRS) model, a system dynamics (SD) model was developed using AnyLogic software to investigate the spread of panic emotions within a population. A case study focused on hospital emergency evacuations was conducted, wherein factors influencing panic propagation were divided into individual and group levels. The population was classified into three categories—staff, caregivers, and patients—and the effect of the ratio of these categories on evacuation efficiency was examined. Based on these classifications, an evacuation simulation experiment was conducted to examine the effects of panic emotions on evacuation efficiency. Results indicate that optimal hospital evacuation efficiency is achieved with a staff:caregiver:patient ratio of 2:2:1. The overall evacuation process is significantly impacted by panic, resulting in a 64 % increase in evacuation times when panic propagation is considered compared to scenarios where it is not. Furthermore, the initial 10 s following a disaster were identified as crucial for managing severe panic. Valuable insights for improving emergency evacuation management are provided by this study.