Latin hypercubes and space-filling designs (Q2799900)

This chapter focuses on a general approach to planning computer experiments ``which seeks design points that fill a bounded design region as uniformly as possible''. Such designs are called space-filling (s-f) designs. The most commonly used class of space-filling designs for computer experiments is that of Latin hypercube designs. However a random Latin hypercube design may not be a good choise in sense of some optimality criteria such as maximin distance and orthogonality. ``As an alternative to the use (s-f) designs, one could choose designs that perform well with respect to some model-dependent criteria -- the minimum integrate mean square error and the maximum entropy''.NEWLINENEWLINEFor the entire collection see [Zbl 1327.62001].