Foraging Commons

Harvesting health restores an agent but depletes a shared resource. A cleanup room restores the commons at a personal opportunity cost. Tests tragedy-of-the-commons dynamics and whether some agents take on the cleanup role.

import comrad

env = comrad.parallel_env("foraging_commons")

Gymnasium id	comrad/foraging_commons-v0
Agents	2 by default (supports 2–4)
Capability tested	Commons dilemma, public-goods provision
Episode limit	5250 steps
Observation	RGB image (72, 128, 3) per agent; measurement vector available with additional_input=True
Action space	Tuple(Discrete(3), Discrete(3), Discrete(3))
Default shared_reward	0.0: each agent receives its own reward
Score (info["true_objective"])	Survival time (episode steps)

Actions

Each agent sends one choice per head; index 0 is always "do nothing".

1. move: none / forward / backward
2. strafe: none / right / left
3. turn: none / left / right

Reward

Per agent: a small bonus for each own harvest and each cleanup (cleanup requires leaving and re-entering the zone); -1 on death.

The score in info["true_objective"] is what the benchmark reports. The shaped reward is only a training signal.

Variants

• None.