Skip to content

Latest commit

 

History

History
37 lines (28 loc) · 2.13 KB

README.md

File metadata and controls

37 lines (28 loc) · 2.13 KB
Raw

Reinforcement Learning with MiniGrid

ShootingMiniGrid ShootingMiniGrid (no walls / static walls / random walls)

Description

This project explores reinforcement learning in 2D environments, using variations of the MiniGrid environment. It focuses on two types: those provided by MiniGrid and a custom variant called ShootingMiniGrid, which introduces shooting mechanics.

Project Structure

The project consists of several Python scripts, each corresponding to different aspects of reinforcement learning and environment simulation:

A) Basic LunarLander Environment from Gymnasium

  • lunar_learn.py: Script for training an agent in the Lunar Lander environment.
  • lunar_load.py: Load and render a trained model in Lunar Lander.

B) MiniGrid-Empty Environment

  • minigrid_empty_learn.py: Training an agent in the MiniGrid-Empty environment.
  • minigrid_empty_load.py: Load and render a trained agent in the MiniGrid-Empty environment.

C) Custom Basic Environment

  • custom_env.py: Definition of a custom MiniGrid environment.
  • custom_env_learn.py: Training an agent in the custom environment.
  • custom_env_load.py: Load and show a trained agent in the custom environment.

D) ShootingMiniGrid: MiniGrid Environment with Shooting Functionality

  • make_shooting_minigrid_env.py: Function to instantiate a ShootingMiniGrid environment.
  • manual_control_shooting.py: Manual control of the agent in ShootingMiniGrid.
  • shooting_minigrid_env.py / v1 / v2 / v3: Versions of the ShootingMiniGrid environment.
  • shooting_minigrid_learn.py / v3: Training the agent in ShootingMiniGrid.
  • shooting_minigrid_load.py: Load and render an agent in ShootingMiniGrid.

How to Use

Developed with Python 3.11. To run the scripts, it is necessary to have installed libraries such as Gymnasium, MiniGrid, Stable-Baselines3, among others. Dependencies can be installed using:

pip install -r requirements.txt