Manually Control Agents

To quickly test the functionality of Crew Wildfire environment without using API costs, you can choose to manually control each agent from the algorithm side.

Once CREW and the additional wildfire-specific instructions are complete and the docker container is running, navigate to the CREW Wildfire directory and activate the crew conda environment. Note that you do not need to set up an OpenAI Access or save API keys.

cd crew-algorithms/crew_algorithms/wildfire_alg
conda activate crew

Then run the following command:

python algorithms\\MANUAL\\__main__.py envs.level=Cut_Trees_Sparse_small envs.seed=483  envs.max_steps=20

You may change the level to any level of your choosing, as well as the seed and the max environment steps.

The list of all preset levels is available here.