This model serves as the baseline for the Aerial Wildfire Suppression environment, trained and tested on task 0 with difficulty 6 using the Proximal Policy Optimization (PPO) algorithm.

Environment: Aerial Wildfire Suppression
Task: 0
Difficulty: 6
Algorithm: PPO
Episode Length: 3000
Training max_steps: 1800000
Testing max_steps: 180000

Train & Test Scripts
Download the Environment

Downloads last month: -; Downloads are not tracked for this model. How to track

Video Preview

Reinforcement Learning

Evaluation results

Crash Count on hivex-aerial-wildfire-suppression
self-reported

0.09166666939854622 +/- 0.13759632868100613
Extinguishing Trees on hivex-aerial-wildfire-suppression
self-reported

32.016666914522645 +/- 54.38789961151889
Extinguishing Trees Reward on hivex-aerial-wildfire-suppression
self-reported

160.08333310484886 +/- 271.9394937062471
Fire Out on hivex-aerial-wildfire-suppression
self-reported

0.21666666865348816 +/- 0.3381234780238663
Fire too Close to City on hivex-aerial-wildfire-suppression
self-reported

0.8 +/- 0.34027852368936023
Preparing Trees on hivex-aerial-wildfire-suppression
self-reported

613.7666711807251 +/- 594.0818245118544
Preparing Trees Reward on hivex-aerial-wildfire-suppression
self-reported

613.7666711807251 +/- 594.0818245118544
Water Drop on hivex-aerial-wildfire-suppression
self-reported

56.541665983200076 +/- 29.057508493771863
Water Pickup on hivex-aerial-wildfire-suppression
self-reported

56.166666412353514 +/- 29.054395046718312
Cumulative Reward on hivex-aerial-wildfire-suppression
self-reported

885.5350048065186 +/- 571.6299167493522

View on Papers With Code