[RL/Value-based]Q-Learning

强化学习(Reinforcement Learning)Building BlocksValue-based RL

Off Policy,学习最优$Q$

$$Q(s,a)\leftarrow Q(s,a)+\alpha[r+\gamma \max_{a'} Q(s',a')-Q(s,a)]$$

思想：直接逼近最优动作值函数$Q^\star(s,a)$

Backward [RL/Value-based]Double Q-Learning [RL/Value-based]SARSA Forward