This model serves as the baseline for the Aerial Wildfire Suppression environment, trained and tested on task 2 with difficulty 9 using the Proximal Policy Optimization (PPO) algorithm.

Environment: Aerial Wildfire Suppression
Task: 2
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-2-difficulty-9

Aerial Wildfire Suppression

Collection

90 items • Updated Mar 20, 2025

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

0.09166666939854622 +/- 0.18317377972526547
Extinguishing Trees on hivex-aerial-wildfire-suppression
self-reported

8.949999886751176 +/- 12.562137901097202
Extinguishing Trees Reward on hivex-aerial-wildfire-suppression
self-reported

44.75000040531158 +/- 62.81068995864626
Fire Out on hivex-aerial-wildfire-suppression
self-reported

0.3500000022351742 +/- 0.4114586818171751
Fire too Close to City on hivex-aerial-wildfire-suppression
self-reported

0.975 +/- 0.11180339887498947
Preparing Trees on hivex-aerial-wildfire-suppression
self-reported

652.12500230968 +/- 512.4888000138135
Preparing Trees Reward on hivex-aerial-wildfire-suppression
self-reported

3260.6250272393227 +/- 2562.444032266141
Water Drop on hivex-aerial-wildfire-suppression
self-reported

65.37499959468842 +/- 34.89231000298373
Water Pickup on hivex-aerial-wildfire-suppression
self-reported

64.99166650772095 +/- 34.92349427276053
Cumulative Reward on hivex-aerial-wildfire-suppression
self-reported

3682.374974441528 +/- 2239.4285326262125