Sliced Latin hypercube designs with arbitrary run sizes

TL;DR

This paper introduces a novel construction method for sliced Latin hypercube designs that accommodate arbitrary run sizes, along with an algorithm to reduce correlations, enhancing their flexibility and statistical properties for computer experiments.

Contribution

It presents the first construction of sliced Latin hypercube designs with arbitrary slice sizes and an algorithm to minimize correlations among points.

Findings

First construction of sliced Latin hypercube designs with arbitrary run sizes

Algorithm for reducing correlations in the designs

Enhanced flexibility for computer experiment designs

Abstract

Latin hypercube designs achieve optimal univariate stratifications and are useful for computer experiments. Sliced Latin hypercube designs are Latin hypercube designs that can be partitioned into smaller Latin hypercube designs. In this work, we give, to the best of our knowledge, the first construction of sliced Latin hypercube designs that allow arbitrarily chosen run sizes for the slices. We also provide an algorithm to reduce correlations of our proposed designs.