Latin hypercube sampling (LHS), a stratified-random procedure, provides an efficient way of sampling variables from their distributions.