Finding Optimal Strategies for Fighting Wildfire With Reinforcement Learning

July 2025 - July 2026

We are developing a large scale simulation of the process of fighting wildfires in Alberta using a variety of real time, historical, and simulated geospatial data to simulate the process of fighting a spreading wildfire under different weather, geographical, and crew conditions. Our aim is to use Reinforcement Learning to find the best strategies to fight wildfire while minimizing operational cost and damages.