Project Name: Reinforcement Learning Tic-Tac-Toe Agent

Description:

This project implements a Tic-Tac-Toe game using reinforcement learning (Q-learning) to train an AI agent to play against a human player. The AI learns through interactions with the environment and gradually improves its gameplay strategy.

Components:

• TicTacToe Class: Manages the game environment, tracks the game state, and checks for wins, draws, available moves, etc.
• QLearningAgent Class: Implements the Q-learning algorithm for the AI agent. Manages Q-values, chooses actions based on exploration-exploitation strategy, and updates Q-values based on rewards received.
• play_game Function: Runs a single game between the AI agent and the environment, facilitating learning and updating Q-values accordingly.
• Main Function (main): Orchestrates the training process for the AI agent by running multiple episodes of the game. After training, it allows a user to play against the trained AI.

Files Included:

• tictactoe.py: Contains the main code implementing the game logic, Q-learning agent, and gameplay functionalities.
• README.md: Provides an overview, instructions, and details about the project.

Usage:

Training the AI:

1. Run the main() function in tictactoe.py to initiate the training process for the AI agent.
2. Adjust the num_episodes variable to control the number of training episodes.

Playing against the AI:

1. After training, the user can play against the trained AI by following the prompts displayed in the console.
2. Run the main() function and interact with the command-line interface to make moves against the AI.

Requirements:

• Python 3.x
• NumPy library

Setup Instructions:

1. Clone this repository to your local machine.
2. Ensure Python 3.x is installed.
3. Create a virtual environment: python3 -m venv venv_name
4. Activate the virtual environment:
   • For Windows: venv_name\Scripts\activate
   • For macOS and Linux: source venv_name/bin/activate
5. Install the required dependencies (numpy): pip install numpy
6. Run python tictactoe.py to start the program.

Additional Notes:

• Feel free to explore and modify the code as needed. Contributions, suggestions, and improvements are welcome.
• If you encounter any issues or have questions, please create an issue in the repository.

Author:

Youssef Baghrous

License:

This project is licensed under the MIT License.