Learning Near-Optimal Policies with Bellman-Residual Minimization Based Fitted Policy Iteration and a Single Sample Path.

Andrs Antos, Csaba Szepesvri, Rmi Munos

Browse the full COLT paper archive.