A reimplementation of the canonical *Hard Maze* deception benchmark used in quality-diversity and neuro-evolution research.

Installation

pip install gymnasium_hardmaze

Usage

The Gymnasium interface allows to initialize and interact with the Hardmaze environment as follows:

import gymnasium as gym

env = gym.make("HardMaze-v0", render_mode="human")

observation, info = env.reset(seed=42)
for _ in range(1000):
   action = policy(observation)  # User-defined policy function
   observation, reward, terminated, truncated, info = env.step(action)

   if terminated or truncated:
      observation, info = env.reset()
env.close()

Citation

To cite this project please use:

@software{gymnasium_hardmaze,
  author = {Stefano Palmieri},
  title = {Maze Simulator: A Gymnasium-compatible Implementation of hardmaze environment},
  url = {https://github.com/Teaspoon-AI/gymnasium-hardmaze},
  year = {2025},
}