🧠 Rainforcement — Deep Reinforcement Learning Exercises

Welcome to Rainforcement, a structured and interactive collection of Jupyter notebooks designed to guide learners through the core concepts of Reinforcement Learning (RL). This project emphasizes building an intuitive understanding of decision-making algorithms used in games and AI systems by implementing them from scratch.

📚 What's Inside

This repository includes four hands-on exercises that gradually increase in complexity, covering from basic planning methods to advanced deep reinforcement learning.

---

📘 exercise-one-step-lookahead.ipynb

Learn the foundation of reinforcement learning through one-step lookahead value iteration, a classic algorithm that evaluates all possible actions for each state. You’ll:

• Implement a simple Markov Decision Process (MDP)
• Understand the Bellman equation and how policies evolve
• Evaluate and improve policy until convergence

---

📗 exercise-n-step-lookahead.ipynb

Build upon the previous notebook by extending the concept to n-step lookahead methods. This enhances strategic planning over multiple future actions:

• Develop recursive planning strategies
• Learn how future rewards impact current decision making
• Implement rollouts and simulations for value estimation

---

🎮 exercise-play-the-game.ipynb

Now it's time to play and test the environment using agents you’ve built! In this notebook, you:

• Interact with a grid-based game environment
• Visualize agent decisions based on current policy
• Evaluate the effectiveness of different planning strategies

---

🧠 exercise-deep-reinforcement-learning.ipynb

Step into modern AI with Deep Q-Learning. This is where function approximation meets reinforcement learning:

• Train a neural network to approximate Q-values
• Implement experience replay and epsilon-greedy strategies
• Understand target networks and stability techniques

---

🛠️ Technologies Used

• Python 3
• NumPy
• OpenAI Gym (optional)
• Matplotlib
• PyTorch / TensorFlow (for deep learning notebook)

🚀 Getting Started

Clone the repository and open notebooks using Jupyter:

git clone https://github.com/rusiru-erandaka/Rainfrocement_Learning_and_GameAI.git
cd Rainfrocement_Learning_and_GameAI

🌱 Ideal For

• Students studying AI or machine learning
• Anyone learning reinforcement learning through implementation
• Developers wanting to understand planning algorithms

📄 License

This project is licensed under the MIT License. Feel free to use and modify it for your own learning or teaching purposes.