本文介绍了强化学习中，通过蒙特卡洛方法进行策略评估和控制。章节涵盖从每步更新到首次访问的蒙特卡洛策略评估，以及动作值的估计。此外，讨论了蒙特卡洛控制方法，包括在线策略和离线策略的控制，以寻找近似最优策略。
第五章 蒙特卡洛方法
Chapter 5 Monte Carlo Methods
Monte Carlo methods require only experience——sample sequences of states, actions, and rewards from on-line or simulated interaction with an environment.
- episode-by-episode（每个episode更新一次）
- learn directly from raw experience without a model of the environment’s dynamics
5.1 Monte Carlo Policy Evaluation
- An obvious way to estimate it from experience, then, is simply to average the returns observed after visits to that state. As more returns are observed, the average should converge to the expected value.
- Each occurrence of state s in
最低0.47元/天 解锁文章
683
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
