Reinforcement Learning Policy Approximation by Behavior Trees