| description abstract | As urban populations continue to grow rapidly, the demand for efficient public transportation systems has increased. Consequently, metro stations have become indispensable components of urban infrastructure. Despite their critical role in urban transportation networks, metro stations are facing various risks and challenges, such as natural disasters, technical failures, and operational disruptions. Unlike other indoor or open spaces, metro stations are enclosed and confined, which presents unique challenges for emergency evacuation. However, past research has shortcomings in fully understanding the evacuation process within metro stations and in developing appropriate strategies to enhance their emergency response capabilities. This study proposes a metamodeling-based simulation optimization approach to provide comprehensive evacuation analysis and adaptable evacuation strategies for metro stations. The feasibility of this approach is demonstrated through a case study based on Wan Chai Station in Hong Kong. The primary contribution of this study lies in providing an integrated framework that combines agent-based simulation, machine learning, and multiobjective optimization to model metro station evacuation scenarios and optimize evacuation performance. This framework aids in analyzing the impact of evacuation influencing factors and capturing the relationship between these factors and evacuation performance, ultimately providing optimal evacuation strategies. In practice, this method can swiftly and flexibly generate optimized evacuation plans and actions within a reasonable budget while considering the specific characteristics of a metro station and potential emergency scenarios. The entire research enhances the emergency preparedness of metro stations and ensures passenger safety. | |
