In contrast to the conventional deep Q-network (DQN) algorithm, the MDQN algorithm enables the experience of multi-agent to be input into a shared neural network to shorten the training time with the ...