https://singularityfeed.com/understanding-multi-agent-reinforcement-learning-marl/