Quantum-enhanced reinforcement learning for control: a preliminary study