Compares SARSA, Q-learning, Expected SARSA and two variants on four cliff walking gridworlds, 500 runs each.
-
Updated
Nov 1, 2026 - Jupyter Notebook
Compares SARSA, Q-learning, Expected SARSA and two variants on four cliff walking gridworlds, 500 runs each.
🦞 Zero-dependency classic RL environments & tabular agents in pure Python (bandits, GridWorld, FrozenLake, CliffWalking, Q-Learning/SARSA, DP). 零依赖纯Python经典强化学习工具箱。
带自动验证的强化学习动手练习:表格 SARSA、蒙特卡洛 vs TD 自举(偏差与方差)、神经网络 SARSA(CartPole)|王树森《深度强化学习》第 5 章配套 | Hands-on RL exercises with an auto-checker: SARSA, Monte Carlo vs TD bootstrapping, neural SARSA on CartPole
深度强化学习动手练习:TD、Q 学习、SARSA、同策略 vs 异策略、DQN(配套王树森《深度强化学习》第 4 章,自动验证)| Hands-on deep RL exercises with an auto-checker: TD learning, Q-learning, SARSA, on-policy vs off-policy, DQN on CartPole (PyTorch + Gymnasium).
To associate your repository with the cliff-walking topic, visit your repo's landing page and select "manage topics."