Function Approximations for Reinforcement Learning Controller for Wave Energy Converters