Symbolic method for deriving policy in reinforcement learning | IEEE Conference Publication | IEEE Xplore