This model serves as the baseline for the Aerial Wildfire Suppression environment, trained and tested on task 2 with difficulty 6 using the Proximal Policy Optimization (PPO) algorithm.

Environment: Aerial Wildfire Suppression
Task: 2
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

Collection including hivex-research/hivex-AWS-PPO-baseline-task-2-difficulty-6

Aerial Wildfire Suppression

Collection

90 items • Updated Mar 20, 2025

Paper for hivex-research/hivex-AWS-PPO-baseline-task-2-difficulty-6

HIVEX: A High-Impact Environment Suite for Multi-Agent Research (extended version)

Paper • 2501.04180 • Published Jan 7, 2025

Evaluation results

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

0.11666667014360428 +/- 0.18809603682131015
Extinguishing Trees on hivex-aerial-wildfire-suppression
self-reported

26.57500003799796 +/- 36.448371998578736
Extinguishing Trees Reward on hivex-aerial-wildfire-suppression
self-reported

132.87500058710575 +/- 182.24186375120422
Fire Out on hivex-aerial-wildfire-suppression
self-reported

0.3166666701436043 +/- 0.4114586821666224
Fire too Close to City on hivex-aerial-wildfire-suppression
self-reported

0.95 +/- 0.15389675281277315
Preparing Trees on hivex-aerial-wildfire-suppression
self-reported

637.158332157135 +/- 502.3677062623285
Preparing Trees Reward on hivex-aerial-wildfire-suppression
self-reported

3185.791680145264 +/- 2511.8385615224347
Water Drop on hivex-aerial-wildfire-suppression
self-reported

81.89166564941407 +/- 40.19878442392626
Water Pickup on hivex-aerial-wildfire-suppression
self-reported

81.56666650772095 +/- 40.211261838762134
Cumulative Reward on hivex-aerial-wildfire-suppression
self-reported

3751.6249725341795 +/- 3124.340359471712