This model serves as the baseline for the Aerial Wildfire Suppression environment, trained and tested on task 1 with difficulty 9 using the Proximal Policy Optimization (PPO) algorithm.

Environment: Aerial Wildfire Suppression
Task: 1
Difficulty: 9
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-1-difficulty-9

Aerial Wildfire Suppression

Collection

90 items • Updated Mar 20, 2025

Paper for hivex-research/hivex-AWS-PPO-baseline-task-1-difficulty-9

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

0.38333334103226663 +/- 0.2709092371191193
Extinguishing Trees on hivex-aerial-wildfire-suppression
self-reported

17.783333230018616 +/- 23.626856569133906
Extinguishing Trees Reward on hivex-aerial-wildfire-suppression
self-reported

889.1666702270508 +/- 1181.3428509988885
Fire Out on hivex-aerial-wildfire-suppression
self-reported

0.10000000298023223 +/- 0.13679711768822556
Fire too Close to City on hivex-aerial-wildfire-suppression
self-reported

0.925 +/- 0.24468024246479642
Preparing Trees on hivex-aerial-wildfire-suppression
self-reported

669.3166633605957 +/- 541.9104371075354
Preparing Trees Reward on hivex-aerial-wildfire-suppression
self-reported

669.3166633605957 +/- 541.9104371075354
Water Drop on hivex-aerial-wildfire-suppression
self-reported

25.191666746139525 +/- 15.035657045429138
Water Pickup on hivex-aerial-wildfire-suppression
self-reported

25.01666667461395 +/- 15.04835278769134
Cumulative Reward on hivex-aerial-wildfire-suppression
self-reported

1912.962504005432 +/- 2485.098888848231