Predictive and Prescriptive AI for Wildfire Suppression Optimization
A new research paper develops a predictive and prescriptive approach to optimize crew assignments and wildfire suppression. The problem involves discrete resource allocation with endogenous wildfire demand and non-linear dynamics. The model uses an integer optimization framework with crew assignments on a time-space-rest network and wildfire dynamics on a time-state network. A two-sided branch-and-price-and-cut algorithm is introduced, featuring column generation for fire suppression plans and crew routes, new cuts exploiting knapsack structure, and branching rules for non-linear dynamics. The study aims to improve decision-making during intense wildfire seasons when resources are scarce.
Key facts
- The paper is titled 'Predictive and Prescriptive AI toward Optimizing Wildfire Suppression'.
- It appears on arXiv with ID 2605.04510.
- The approach jointly optimizes crew assignments and wildfire suppression.
- The model includes endogenous wildfire demand and non-linear wildfire dynamics.
- An integer optimization model is formulated with time-space-rest and time-state networks.
- A two-sided branch-and-price-and-cut algorithm is developed.
- The algorithm uses column generation, knapsack cuts, and novel branching rules.
- The research addresses resource allocation during intense wildfire seasons.
Entities
Institutions
- arXiv