Lunar-Lander-with-Reinforcement-Learning

This project was done as capstone project in Reinforcement learning specialization from university of Alberta.

• In this project, reinforcement learning is used to learn to land a rocket on a landing pad by controlling rocket engine actions.

Dependencies :

1. Numpy
2. Matplotlib
3. Gym
4. Pytorch

GYM installing instructions

• To install the gym which is from openAI run the following code for windows10 pip install gym pip install box2d

• And use pip3 in case of ubuntu

• In case you are using conda environment first install pip in conda evironment and use the same above mentioned commands.

Files Description

In the codes folder following files are there :
1.Actor_critic.py : It is the utility script for actor-critic algorithm.
2.DQNetwork.py : It is the utility script for Deep Q-Network algorithm.
3.Policy_gradient.py : It is the utility script for policy gradient algorithm.
4.utils1/2.py : It is the utilty script containing plotting functions.
5.Lunar_Lander.ipynb : It is notebook containing all the 3 models executed for 300 episodes.

License & copyright

© Rahul Kumar
Licensed under the MIT License