This repository includes various projects using different deep reinforcement learning algorithms and concepts performed on various OpenAI and Unity environments.
This folder contains various python notebooks in which different openAI environments like CartPole, MountainClimbing, Lunar lander etc have been solved using RL algorithms. Check out the notebooks for more details on that.
In this project, an agent was trained using a Deep-Q network to navigate in a square world full of yellow and blue bananas. The agent learnt to avoid the blue bananas and only grab the yellow ones.
In this project, robotic arms learnt how to align themselves in a particular target position by applying different value of torques to the two joints. The agent was trained using a DDPG model.
In this project, two agents were trained to play tennis with each other without letting the ball drop. This is an example of a multi-agent problem and hence was tackled using a multi-agent DDPG algorithm.