RL Algorithms from scratch using Open AI Gym
Overview
In this project, I implemented and analyzed 9 reinforcement learning algorithms from scratch—ranging from tabular methods like Q-Learning and SARSA to deep and actor–critic approaches such as DQN, DDQN, Actor-Critic, A2C, PPO, DDPG, and TD3.
For each algorithm, I:
Derived the mathematical foundations and update equations
Implemented the model from scratch in Python (PyTorch + NumPy)
Tested and benchmarked across multiple OpenAI Gym environments
Analyzed learning curves, policy behavior, and algorithmic trade-offs
This project deepened my understanding of reinforcement learning principles, bridging theoretical derivations with practical implementations and highlighting the evolution from simple value-based agents to sophisticated actor–critic frameworks.
Tools
Python
Open AI Gym
Semester
7
Grade
A+
RL Algorithms from scratch using Open AI Gym
Overview
In this project, I implemented and analyzed 9 reinforcement learning algorithms from scratch—ranging from tabular methods like Q-Learning and SARSA to deep and actor–critic approaches such as DQN, DDQN, Actor-Critic, A2C, PPO, DDPG, and TD3.
For each algorithm, I:
Derived the mathematical foundations and update equations
Implemented the model from scratch in Python (PyTorch + NumPy)
Tested and benchmarked across multiple OpenAI Gym environments
Analyzed learning curves, policy behavior, and algorithmic trade-offs
This project deepened my understanding of reinforcement learning principles, bridging theoretical derivations with practical implementations and highlighting the evolution from simple value-based agents to sophisticated actor–critic frameworks.
Tools
Python
Open AI Gym
Semester
7
Grade
A+
