RL1.6B Proof Sketch: SARSA is consistent with Bellman equation

Поделиться
HTML-код
  • Опубликовано: 23 ноя 2024
  • Online SARSA updates cause fluctuations of the estimated Q-values. Here we show that under some conditions the temporal mean of the Q-values solve the Bellman equation
  • НаукаНаука

Комментарии •