https://github.com/sainikhil0904/breakthrough_game_ai_agent
Breakthrough Game AI: Q-Learning-based intelligent player for an 8x8 strategic board game.
https://github.com/sainikhil0904/breakthrough_game_ai_agent
aritificial-intelligence chess-ai chess-game q-learning
Last synced: 7 months ago
JSON representation
Breakthrough Game AI: Q-Learning-based intelligent player for an 8x8 strategic board game.
- Host: GitHub
- URL: https://github.com/sainikhil0904/breakthrough_game_ai_agent
- Owner: SaiNikhil0904
- Created: 2023-11-17T15:32:22.000Z (almost 2 years ago)
- Default Branch: main
- Last Pushed: 2024-02-14T17:52:28.000Z (over 1 year ago)
- Last Synced: 2025-01-21T17:50:02.050Z (9 months ago)
- Topics: aritificial-intelligence, chess-ai, chess-game, q-learning
- Language: Jupyter Notebook
- Homepage:
- Size: 816 KB
- Stars: 1
- Watchers: 1
- Forks: 1
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Breakthrough_Game_AI_Agent
## Project Overview
Welcome to the Breakthrough Game project, a collaborative effort by Godavarthi Sai Nikhil and Anish Borkar. In this project, we have implemented an intelligent AI player using Q-Learning for the Breakthrough Game—a strategic board game that combines elements of chess and traditional war games. The project challenges players with intricate moves, pawn captures, and special maneuvers on an 8x8 square board.
## Team
- **Anish Borkar**
- **Godavarthi Sai Nikhil**
## Requirements
- **Programming Language:** Python
- **Libraries:** NumPy, Sys, IPython display
- **Development Environment:** VsCode, Google Colab, Jupyter Notebook
## Mentor
We extend our sincere gratitude to **Dr. Soharab Hossain Shaikh** for his guidance and mentorship throughout the development of our Breakthrough Game project.
## Purpose
The primary objective of this project is to create an AI player for the Breakthrough Game using Q-Learning. We aim to train the AI to make intelligent moves, strategically advancing towards the opponent's home row while considering pawn captures, special moves like en passant, and defensive strategies.
## Key Features
- **Table-Driven Approach:** We have implemented a structured decision-making process using predefined rules and strategies. This approach provides clarity in decision-making and facilitates easy scalability and modification of the AI's behavior.
- **Q-Learning:** Our intelligent agent utilizes the Q-Learning algorithm to learn and adapt its strategies based on observed rewards and penalties. This dynamic approach allows the AI to evolve over time and improve its decision-making capabilities.
- **Game Mechanics:** The project faithfully mimics the rules of the Breakthrough Game, involving pawn movement, captures, en passant moves, and an endgame focus on reaching the opponent's end of the board.
## Project Impact
Our intelligent AI player contributes to the field of reinforcement learning, showcasing adaptability and dynamic decision-making. The project includes a sensitivity analysis, providing insights into the impact of hyperparameters such as the discount factor, learning rate, and epsilon decay rate on the learning dynamics of the algorithm.
## Explore the Work
Feel free to explore our codebase, try different strategies, and understand how the Q-Learning algorithm evolves with each game. Check out our [Colab Notebook](https://colab.research.google.com/drive/1k20dgTXAw3rYJ4mMZ4lUyAzkJ1ZMDFEj?usp=sharing).
### Sample Images
*Fig 1.1: Initial Setup of the Breakthrough Game*
*Fig 1.2: Goal of the Breakthrough Game (Reaching the other end)*
## Contact Us
For inquiries, collaboration opportunities, or further information, please feel free to reach out to us:
- **Godavarthi Sai Nikhil:** nikhilgodavarthi9@gmail.com
- **Anish Borkar:** anishborkar73@gmail.com
Thank you for exploring our Breakthrough Game project! Join us in this journey of intelligent gameplay and reinforcement learning.