Abstract: Self-play theory is an important branch of reinforcement learning. Different levels of complexity in scenarios are also necessary to make reinforcement learning more sophisticated. In this ...