Isovolumetric adaptations to space-filling design of experiments

Abstract A brief review of methods in design experiments and criteria to determine space-filling properties a set samples is given. Subsequently, the so-called curse dimensionality sampling reviewed used as motivation for proposal an adaptation strata creation process Latin hypercube based on idea nested same-sized hypervolumes. The proposed approach places closer space boundaries, where higher dimensions majority volume located. same introduced Monte Carlo considering affordable number a-posteriori transformation. Both ideas are studied different algorithms compared using distance-based criteria. new then enables more efficient optimization especially high-dimensional problems, i.e. problems with high variables.