ReinforcementLearning

A practical, hands-on introduction to Reinforcement Learning (RL) through applied projects. This repo walks you from core RL concepts → Atari agents → autonomous driving → custom environment design.

---

Repository Contents

• main.ipynb
  A foundational notebook that walks through RL basics:

  • Environment setup
  • Installation of Stable-Baselines
  • Exploration of OpenAI Gym spaces (observation & action spaces)
  • Training metrics visualization
  • TensorBoard integration
  • Performance tuning with callbacks
  • Customizing neural network architectures

• Project 1 – Project1-Atari games.ipynb

  • Algorithm: A2C (Advantage Actor-Critic)
  • Focus: Training an RL agent to play classic Atari games using Gym’s Atari environments.
  • Highlights: Preprocessing game frames, reward shaping, training/stability insights.

• Project 2 – Project2-Self Driving.ipynb

  • Algorithm: PPO (Proximal Policy Optimization)
  • Focus: Autonomous driving in a Box2D-based environment, using pyglet for rendering (e.g., CarRacing).
  • Highlights: Continuous control, PPO stability in physics-based environments, reward design, rendering integration.

• Project 3 – Project3-Custom Environment.ipynb

  • Algorithm: PPO
  • Focus: Building a custom Gym environment (a “shower temp control” simulation) to regulate temperature correctly.
  • Highlights: Environment creation, observation and action design, reward engineering, and agent training.

• Supporting Files:

  • cartpole_run.mp4: A demo video showcasing agent behavior.
  • PPO.zip: (Optional) Maybe contains pretrained models or training logs.

---

Getting Started

1. Clone the Repo

git clone https://github.com/anjaliy11/ReinforcementLearning.git
cd ReinforcementLearning