Effectiveness Verification of Evacuation Guidance Including Underground Passages Using Multi-objective Optimization

In this study, we approach the optimization problem of evacuation guidance assuming major terminal stations in Tokyo using crowd simulation and optimization algorithms. We propose a method to optimize multiple indicators with various guidance variables, including underground passages, and evaluate multiple scenarios obtained through optimization. The results of experiments using three algorithms: Random search, NSGA-II, and MOTPE, showed that MOTPE can be used to search for high-quality solutions quickly. Additionally, scenarios with guidance shortened the total evacuation time and reduced the congestion levels compared to scenario without guidance.