SARSA
An on-policy temporal-difference (TD) control algorithm that updates the action-value function Q(s,a) using the action actually taken, so that the behavior policy and the target policy are the same.
Timeline (last 90 days)
In environments with strong exploration and high risk, Q-learning is more aggressive and SARSA is more conservative, and the two may converge to different policies.
SARSA is an on-policy temporal-difference learning algorithm that updates value estimates based on the sequence of actions actually taken by the agent.
Q-Learning 是 off-policy 的 TD 控制方法，直接学习最优动作价值函数；SARSA 是 on-policy 方法，更新时使用智能体实际选择的下一个动作
Q-Learning 与 SARSA 的核心区别在于更新目标中是取 max（Q-Learning）还是取实际执行的动作值（SARSA），SARSA 在悬崖行走等环境中会学习到更保守但更安全的策略
All Facts (4)
In environments with strong exploration and high risk, Q-learning is more aggressive and SARSA is more conservative, and the two may converge to different policies.
50%UnverifiedSARSA is an on-policy temporal-difference learning algorithm that updates value estimates based on the sequence of actions actually taken by the agent.
50%UnverifiedQ-Learning 是 off-policy 的 TD 控制方法，直接学习最优动作价值函数；SARSA 是 on-policy 方法，更新时使用智能体实际选择的下一个动作
50%UnverifiedQ-Learning 与 SARSA 的核心区别在于更新目标中是取 max（Q-Learning）还是取实际执行的动作值（SARSA），SARSA 在悬崖行走等环境中会学习到更保守但更安全的策略
50%