This paper proposes implementing fuzzy logic to improve upon the decision-making and resource allocation during a wildland fire. The problem is based on previous work of implementing neural dynamic programming in the theater-missile defense problem. The scenario was modified to the parameters of a wildland fire and extended to include multiple layers of defense. Three key areas were studied to evaluate the performance of the algorithm: sensitivity to engagement probability, uncertainty analysis, and scalability. The control methodologies were critiqued by the remaining health of the assets and the execution time. The neuro-fuzzy dynamic programming showed improved results in the uncertainty cases while being robust to system complexity.
展开▼