Landing Rocket Lander with DDSM

This is an implementation of the Rocket Lander Environment using Deep Deterministic Selective Memory.

The Author for this environment is arex18

Results

The agent starts to land regularily after 500 episodes. You can see the agent in action here

How to run

To run the simulation follow the steps below:

Download or clone the repository https://github.com/FitMachineLearning/rocket-lander
Install the necessary dependencies. Usually pip install
run the project with the following command python Main_Sim.py

Prerequisites

List of libraries needed to run the project: (some, e.g. cvxpy require other pre-requisites). Windows users head to the life-saving list of [Windows Python Extension Libraries](Unofficial Windows Binaries for Python Extension Packages) to install cvxpy and any other failing pip installs.

tensorflow
matplotlib
gym
numpy
Box2D
logging
pyglet
cvxpy
abc
concurrent

python pip install PATH_TO_YOUR_DOWNLOADED_LIBRARY (ending in whl)

AUtho: Who is Michel Aka

Michel is an AI researcher and a graduate from University of Montreal who currently works in the Healthcare industry.

References: https://gym.openai.com/envs/LunarLander-v2/

License