This model serves as the baseline for the Wildfire Resource Management environment, trained and tested on task 0 with difficulty 4 using the Proximal Policy Optimization (PPO) algorithm.
Environment: Wildfire Resource Management
Task: 0
Difficulty: 4
Algorithm: PPO
Episode Length: 500
Training max_steps: 450000
Testing max_steps: 45000
Train & Test Scripts
Download the Environment
Evaluation results
- Cumulative Reward on hivex-wildfire-resource-managementself-reported180.52263412475585 +/- 90.05154933560763
- Collective Performance on hivex-wildfire-resource-managementself-reported73.34980735778808 +/- 33.30569260296172
- Individual Performance on hivex-wildfire-resource-managementself-reported36.9462173461914 +/- 16.558159505779127
- Reward for Moving Resources to Neighbours on hivex-wildfire-resource-managementself-reported88.78441047668457 +/- 41.79921581399254
- Reward for Moving Resources to Self on hivex-wildfire-resource-managementself-reported2.90758473277092 +/- 2.9947906681759133