Utils

PettingZoo has some utilities to help make simple interactions with the environment trivial to implement. Utilities which are designed to help make environments easier to develop are in the developer documentation.

Average Total Reward

The average total reward for an environment, as presented in the documentation, is summed over all agents over all steps in the episode, averaged over episodes.

This value is important for establishing the simplest possible baseline: the random policy.

from pettingzoo.utils import average_total_reward
from pettingzoo.butterfly import pistonball_v6
env = pistonball_v6.env()
average_total_reward(env, max_episodes=100, max_steps=10000000000)

Where max_episodes and max_steps both limit the total number of evaluations (when the first is hit evaluation stops)

Observation Saving

If the agents in a game make observations that are images then the observations can be saved to an image file. This function takes in the environment, along with a specified agent. If no agent is specified, then the current selected agent for the environment is chosen. If all_agents is passed in as True, then the observations of all agents in the environment is saved. By default, the images are saved to the current working directory in a folder matching the environment name. The saved image will match the name of the observing agent. If save_dir is passed in, a new folder is created where images will be saved to. This function can be called during training/evaluation if desired, which is why environments have to be reset before it can be used.

from pettingzoo.utils import save_observation
from pettingzoo.butterfly import pistonball_v6
env = pistonball_v6.env()
env.reset()
save_observation(env, agent=None, all_agents=False, save_dir=os.getcwd())