Definition of nested experimental design sets for Multi-Fidelity Cokriging models

Share:

Description

Build an object of class S3 ("NestDesign") defining nested experimental design sets used to build multi-fidelity Cokriging models.

Usage

1
	NestedDesign(x, nlevel , indices = NULL, n = NULL)

Arguments

x

a matrix representing the experimental design set of the code level 1.

nlevel

an integer representing the number of code levels.

indices

a list of vectors. The ith element of the list contains the indices of the points in the experimental design set of the level i-1 constituting the experimental design set of the level i. If indices = NULL they are randomly sampled according to the number of observations defining in n.

n

a vector containing the number of observations at level k=2,...,nlevel. It is not taking into account if indices is different from NULL.

Details

The procedure does not change the experimental design set of the highest code level. During the procedure, the points of D_{k-1} the closest to those of D_k with k=2,...,nlevel are removed and they are replaced by the points of D_k . Thus, the length of the final D_{k-1} could be larger than the one of the initial D_{k-1}. (see "MuFicokm")

Value

an object of class ("NestDesign") representing a nested experimental design set.

Author(s)

Loic Le Gratiet

References

LE GRATIET, L. & GARNIER, J. (2012), Recursive co-kriging model for Design of Computer Experiments with multiple levels of fidelity, arXiv:1210.0686.

See Also

MuFicokm, ExtractNestDesign, NestedDesignBuild

Examples

1
2
3
4
5
6
7
#-- Nested Experimental design sets
	dimension <- 3
	nD1 <- 100
	nD2 <- 50		
	nD3 <- 20
	set.seed(1);D1 <- matrix(runif(n=nD1*dimension, 0,1),ncol=dimension)
	NestDesign <- NestedDesign(D1, nlevel=3 , n = c(nD2,nD3))