Abstraction in Reinforcement Learning door Sertan Girgin