Rule abstraction and transfer in reinforcement learning by decision tree | IEEE Conference Publication | IEEE Xplore