Two-stage stochastic optimization of passenger evacuation routes in metro stations considering stampede incidents

Stampede incidents frequently occur in densely populated metro stations during emergencies, and if not promptly addressed, they can cause more severe cascading effects, resulting in passenger injuries and even deaths. To address this issue, this study proposes a two-stage stochastic optimization model that incorporates the occurrence of unpredictable stampedes to effectively plan individualized evacuation routes for passengers. The model aims to minimize evacuation time by optimizing both the initial evacuation routes of passengers and the subsequent routes after stampede incidents, thereby mitigating passenger congestion in bottleneck areas and providing passengers with alternative destinations. By introducing auxiliary variables, the model is reformulated to reduce the constraints. The model is tested under various stampede scenarios. The results indicate that the proposed optimization strategy can significantly reduce evacuation time, as well as the density in bottleneck area of the metro station, regardless of the severity of stampede incidents. Via comprehensive sensitivity analysis, the critical factors having a significant impact on the evacuation process have been identified. This finding is expected to provide valuable insights for regulatory agencies in the effective implementation of evacuation strategies.