This codebase is based on PyMARL and contains the implementation of the SOP-CG algorithm.

Tasks can be found in src/envs. To run experiments on MACO benchmark:

python src/main.py --config=sopcg --env-config=pursuit with construction='tree' use_action_repr=False

To run experiment on Tag:

python src/main.py --config=sopcg_vs_vdn --env-config=tag with construction='tree' use_action_repr=False

To run experiments on SMAC benchmark:

python src/main.py --config=sopcg --env-config=sc2 with env_args.map_name='10m_vs_11m' construction='tree' use_action_repr=True

The hyperparameter construction is used to control the graph class, where tree represents $\mathcal{G}_T$, and matching represents $\mathcal{G}_P$. By default, it is set to tree. The hyperparameter use_action_repr is set to False by default, and setting it to True would use the action representation learning technique.

The requirements.txt file can be used to install the necessary packages into a virtual environment.

You can save the learnt models to disk by setting save_model = True, which is set to False by default. The frequency of saving models can be adjusted using save_model_interval configuration. Models will be saved in the result directory, under the folder named models. The directory corresponding to each run will contain models saved throughout the training process, each of which is named by the number of timesteps passed since the learning process starts.

Learnt models can be loaded using the checkpoint_path parameter, after which the learning will proceed from the corresponding timestep.

save_replay option allows saving replays of models which are loaded using checkpoint_path. Once the model is successfully loaded, test_nepisode number of episodes are run on the test mode and a .SC2Replay file is saved in the Replay directory of StarCraft II. Please make sure to use the episode runner if you wish to save a replay, i.e., runner=episode. The name of the saved replay file starts with the given env_args.save_replay_prefix (map_name if empty), followed by the current timestamp.