Abstract
We propose an agent-based simulation system for evacuations after earthquakes and tsunami. We focus on an improved model of communication among agents, by including a model of speech intelligibility in crowds, and the use of voice to change the evacuation-start behavior. This allows the model to represent the effects of evacuees calling to each other, and of automated emergency broadcast systems. To validate our model, we simulate hypothetical scenarios using data from the Great Hanshin-Awaji Earthquake and the Great East Japan Earthquake and tsunami as the initial conditions. Evacuation start and completion rates from the simulation were similar to those observed in the original events, based on data from post-disaster surveys, and showed the model’s ability to serve as a policy modeling tool.
