Linear Function Approximation optimal control
Q-Learning with LFA may diverge. [B,G]
Sarsa with LFA converges [NR,S].
Monte Carlo with LFA converges.
Previous slide
Next slide
Back to first slide
View graphic version