Solving the Rubik's Cube with Deep Reinforcement Learning and Search