Dyna-Q Learning

Algorithm convergence

Property guaranteeing that Dyna-Q's value estimates converge to the optimal values under certain conditions of an exact model and infinite visits.

← Tillbaka