Abstract: Forest management entails planning silviculture activities, often optimized through Mixed Integer Programming (MIP) considering objectives like carbon stock, biodiversity, and soil erosion constraints. Our innovation addresses fire differently: simulating its spread rather than integrating it directly into the MIP model. Initially, the MIP model is solved. Then, a worst-case fire spread simulation identifies paths with excessively high rates. Each unacceptable path adds a constraint to the MIP. This iterative process continues until no unacceptable paths remain, ensuring fire spread rates never exceed a set threshold.
