112
Voted
JMLR
2002
15 years 23 hour ago
2002
We consider a finite-state Markov decision problem and establish the convergence of a special case of optimistic policy iteration that involves Monte Carlo estimation of Q-values,...