Abstract: With an increasing number of agents, the classic multi-agent deep deterministic policy gradient (MADDPG) algorithm often suffers from an input-dimension explosion in the critic caused by ...