This repository is used for the final project of IANNwTF 2022/2023. We aim to train Multi Agents with Reinforcement Learning to play Atari's Pong.

Even though we implemented a compiling code using a DQN for both a Single-Agent Pong and a Multi-Agent Pong, our Agents are not learning in either scenario. A brief introduction to Deep Reinforcement Learning with Deep-Q-Networks in the Single- and Multi-Agent case, a description of our code as well as a closer analysis of our results and further suggestions may be found in our Documentation.

Video Presentation: https://myshare.uni-osnabrueck.de/f/6f66b225ff8f4e8b8a01/