Function Approximation for Reinforcement Learning Controller for Energy from Spread Waves