mazebot

This is a demo I created when I was learning the basics of reinforcement learning. It uses the Q-learning algorithm to train a bot to navigate a maze automatically.

Usage

version: Python 3.11

For macOS users, if you encounter ModuleNotFoundError: No module named 'tkinter' when running the code, you need to execute the corresponding commands for your version.

$ brew install [email protected]

Install dependencies.

$ pip3 install -r requirements.txt

Run main.py

$ python3 main.py

Black represents traps, brown represents walls, green represents the destination, red represents the starting point, and gray represents the accessible path.👇🏻 (ps: If the bot falls into a trap, it will be returned to the starting point.)

The bot will become smarter as it tries different moves.

Contributing

Contributions must be available on a separately named branch based on the latest version of the main branch.