Skip to content

About

Reinforcement learning implementation of a Double Deep Q-Network agent trained to play Tic-Tac-Toe. Demonstrates experience replay, target networks, and stability improvements over vanilla DQN.

Resources

Stars

0 stars

Watchers

0 watching

Forks

 
 

Latest commit

 

History

6 Commits

Folders and files

NameName
Last commit message
Last commit date
 
 
 
 
 
 

Repository files navigation

Tic-Tac-Toe with Double Deep Q-Networks (Double DQN)

Overview

This repository features an implementation of the Double Deep Q-Network (Double DQN) reinforcement learning algorithm to master the game of Tic-Tac-Toe.

Implementation Details

The project utilizes Deep Reinforcement Learning to train an agent that can make optimal moves in a competitive environment. The use of Double DQN helps in reducing the overestimation of Q-values, leading to more stable training.

Key Features

  • Double DQN Agent: A neural-network-based agent trained using reinforcement learning.
  • Configurable Environment: A flexible game environment for training and evaluation.
  • Reproducible Experiments: Scripts and logs detailing the training progress and agent performance.

Getting Started

Setup

  1. Clone the repository:
    git clone https://github.com/Mrudula-itsjuzme/tic-tac-toe-double-dqn.git
  2. Install dependencies:
    pip install -r requirements.txt

Training the Agent

Execute the main entry script to begin the training process:

python main.py

Repository Structure

  • src/: Core implementation of the Double DQN agent and game logic.
  • docs/: Architectural notes and research findings.
  • FAI_proj_S3/: Directory containing project-specific assets and configurations.

Future Enhancements

  • Integration with multi-agent reinforcement learning (MARL).
  • Support for larger board sizes and complex game variants.
  • Implementation of a user interface for human-versus-AI gameplay.

About

Reinforcement learning implementation of a Double Deep Q-Network agent trained to play Tic-Tac-Toe. Demonstrates experience replay, target networks, and stability improvements over vanilla DQN.

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages