Solving MIMO Env Beamforming problem between two Base Stations (BS), in radio communications using reinforcement learning

Millimeter waves is one of the key innovations of 5G communications, that can help achieve high data rates using massive Multiple-Input-Multiple-Output (MIMO) radio units and Beamforming technologies. There have only been very traditional approaches in applying radio beamforming which includes complete scanning or scanning at multiple levels. With the increase in complexities to radio communications through 5G technologies, this implementation is an attempt to solve Beamforming problem using Machine Learning especially, Reinforcement Learning.

This is a custom environment developed for MIMO RL application using OpenAI gym interface. This directory contains implementation of the environment developeed between two BS.

This directory contains the main function of various tests carried on the developed custom environment and Reinforcement learning solution for Beamforming problem. The implementations in this directory has both Q-table learning implementations as well as deep Qlearning.

run python setup.py before executing the test code. This helps installing the necessary packages required for this environment.