When can we improve on sample average approximation for stochastic optimization?

Open in new window