Explainable Reinforcement Learning for Next Generation Food Systems