https://medium.com/@monocosmo77/how-multi-agent-reinforcement-learning-works-part3-artificial-intelligence-5c8ccd27077f