Tabular reinforcement learning for reward robust, explainable crop rotation policies matching deep reinforcement learning performance