In some place I read that one can solve the following problem
$$ \max_{x \in B} \mathbb{E}_{y}\left[ f(x,y) \right] $$
where $x$ is my choice vector and $y$ is a random vector, by simulating $N$ samples, for each sample maximizing the problem as if it was deterministic, but adding the constraint that the optimizer should be the same across samples.
Can someone provide a name for this method? or a reference? I think I read it in a book by Powell, but can't actually find it.