When you say it is much more efficient, do you mean more efficient compared to VBSA only, or is it also more efficient than the Method of Morris? Surely if you want to make clear distinctions between either end of the variogram function, then there will have to be several values of h, perhaps 5 or more. If this is then used on a multidimensional surface, perhaps due to 10 factors, the number of pairs will skyrocket, will they not? Surely the computation of these points to produce these pairs will need as many simulations, in terms of order of magnitude, as the Variance-based approaches? Thank you.
When you say it is much more efficient, do you mean more efficient compared to VBSA only, or is it also more efficient than the Method of Morris?
Surely if you want to make clear distinctions between either end of the variogram function, then there will have to be several values of h, perhaps 5 or more. If this is then used on a multidimensional surface, perhaps due to 10 factors, the number of pairs will skyrocket, will they not? Surely the computation of these points to produce these pairs will need as many simulations, in terms of order of magnitude, as the Variance-based approaches?
Thank you.