Reinforcement Learning code applied to the Lunar Lander problem using MATLAB.
This repository contains the code for the TU Delft project course "Bio-Inspired Intelligence and Learning for Aerospace Applications - AE4350" under the folder "Code".
The subfolder "LunarLander_DQN_nominal" contains all the necessary files to run the problem in the nominal setup. The problem is run from "mainLunarLander.m".
The folder "LunarLander_DQN_sensitivity_analysis" contains all the necessary files to run the sensitivity/robustness analyses using one of "mainLunarLander_xxxx.m" codes.
The Reinforcement Learning Toolbox and Deep Learning Toolbox are needed for the scripts to run.
"Untrained_agent_crash.avi" and "Trained_agent_landing.avi" are sample animations that show, in a graphical way, the landing trajectory before and after Agent training. They were generated with the provided code.