reinforcement learning

have not been revised yet.