Track #3: Wildfire (Both Agent and Task Openness)
- Web: MOASEI
- GitHub: oasys-mas/free-range-zoo
- Kaggle: Competition Configurations
University of Utah, Utah, USA
- Varun Raveendra
- Seongil Heo
- Yanxi Lin
The structure of the repository described below:
.
├── competition_configs # Environment configurations
│ └── wildfire
├── free_range_zoo # Package source
│ ├── envs # Environment implementations
│ ├── utils # Converters / environment abstract classes
│ └── wrappers # Model wrappers and utilities
├── tests # Tests
│ ├── free_range_zoo # Tests for the free_range_zoo package
│ │ ├── envs # environment utilities
│ │ └── utils # all package utilities
│ ├── profiles # Environment performance profiles
│ └── utils # Testing utilities
├── experiments # Experiments (**ours**)
│ ├── core.py # Core classes definitions (Graph, Actor, Critic, Network)
│ ├── evaluation.py # Evaluation scripts
│ ├── quick_start.py # Quick start guide and example scripts
│ ├── test.py # Test scripts for the baseline models
│ ├── train_a2c.py # Training script for A2C model
│ ├── train_gnn.py # Training script for PL model
│ └── utils.py # Utility functions
├── LICENSE # License file
├── poetry.lock # Poetry lock file
├── pyproject.toml # Package dependencies and package definition
├── README.md # Project documentation
└── setup.cfg # Setup configuration
For installation, please refer to the Installation Guide for detailed instructions on how to set up the environment and install the required dependencies.
python evaluation.py [OPTIONS] <output> <model> <config>
- output: Path to the directory where evaluation results and logs will be saved.
- model : Path to the directory containing the trained model.
- config: Path to the environment configuration (e.g., competition_configs/WS3.pkl).
Option | Description |
---|---|
-h, --help | Show help message and exit |
--cuda | Use CUDA (GPU) if available |
--threads THREADS | Number of threads to use |
--log_level LEVEL | Set logging level (DEBUG, INFO, WARNING, ERROR, CRITICAL) |
--seed SEED | Random seed for the evaluation process |
--dataset_seed DATASET_SEED | Random seed for initializing the environment configuration |
--testing_episodes N | Number of test episodes to run (parallel_envs) |
python run/evaluation.py ./output logging/250519_120000/model_a2c.h5 ./competition_config/wildfire/WS1.pkl --testing_episodes 100
This repository is licensed under the GNU Affero General Public License v3.0 (AGPL-3.0).
This project was developed by Team Markov Mayhem as part of the MOASEI 2025 competition.
It is based in part on the free-range-zoo repository by OASYS Labs, which is also licensed under the AGPL-3.0 license.
All code under the experiments/
directory was newly developed by the team in 2025.