This study introduces multiagent conversation models utilizing generative AI to overcome the challenges of traditional rule-based multiagent simulations (MAS) and proposes an appropriate evacuation instruction method through simulations considering psychological factors and information transmission during disasters. Using an earthquake as the disaster scenario, the correlation between agent conversations and evacuation behaviors was evaluated, and the optimal evacuation instructions were verified. Additionally, the evacuation instruction methods were analyzed, and the improvement of prediction accuracy was verified. This research deepens the understanding of evacuation behavior during disasters and contributes to the development of realistic disaster response simulations, offering insights for improving actual evacuation instruction methods.
