Machine Learning For Earthquake Emergency Evacuation: Site Selection And Neighborhood Navigation

Abstract

Many big cities are at risk of
earthquakes. Need good plans to help people evacuate quickly
and safely. Cities with a lot of people often don’t have emergency
centers making it hard for some areas to get help. This study
aims to create a plan for emergency evacuations during
earthquakes. It uses a computer model to find the locations for
emergency centers and help people navigate to them.
The city of Tehran with 8.7 million people and several active
fault lines is used as an example. Many neighborhoods in
Tehran lack evacuation facilities. The computer model uses data
about the city, its people and earthquake risks to find locations
for emergency centers.
Tehrans data was used to train the model along with data
from cities with similar risks and characteristics. The results
show that this approach makes it much easier for people to
evacuate. It reduces the distance people need to travel to reach
an emergency center and increases the number of centers per
person in each neighborhood.
A special app was also developed to help guide people to the
emergency center during an emergency. This app uses maps.
Routing algorithms to find the best route.
Combining computer-based site selection with navigation
makes emergency evacuation planning more efficient, accessible
and reliable. This framework can be used in cities, with similar
characteristics and can be updated with local data to support
disaster preparedness and urban planning.
Keywords: Earthquake Emergency Evacuation