minimaxdesign (version 0.1.5)

miniMaxPro: Compute minimax projection designs using clustering on constrained design regions

Description

miniMaxPro.

Usage

miniMaxPro(N,p,mMdes=NA, mMtol=1e-3*p,
          neval=1e5, itmax_refine=100, ...)

Arguments

N

Number of design points.

p

Dimension of design region.

mMdes

Minimax design from minimax().

mMtol

Tolerance for fill distance increase (actual increase may be slightly higher).

neval

Number of sample points for refinement.

itmax_refine

Maximum number of iterations for refinement.

...

Parameters for minimax().

Value

A list with two objects:

minimax

An N-by-p matrix for the minimax design.

miniMaxPro

An N-by-p matrix for the minimax projection design.

Examples

# NOT RUN {
#30-point miniMaxPro design on the hypercube [0,1]^6
D <- minimax(N=30,p=6)
D <- miniMaxPro(N=30,p=6,mMdes=D)
mMdist(D$minimax)$dist
mMdist(D$miniMaxPro)$dist #slightly higher fill distance
pairs(D$minimax,xlim=c(0,1),ylim=c(0,1),pch=16)
pairs(D$miniMaxPro,xlim=c(0,1),ylim=c(0,1),pch=16) #... but better projections

# }