Deep Learning and back-propagation have been successfully used to perform centralized training with communication protocols among multiple agents in a cooperative Multi-Agent Deep Reinforcement Learning (MARL) environment. In ...