Reinforcement learning algorithms

shape