Text this: Reinforcement learning and approximate dynamic programming for feedback control /