SOAs_8level {SOAs}

Function to create 8-level SOAs according to Shi and Tang 2020

Description

creates strength 3 or 3+ SOAs with 8-level factors in 2^k runs, k at least 4. These SOAs have at least some more balance than guaranteed by strength 3.

Usage

SOAs_8level(
  n,
  m = NULL,
  constr = "ShiTang_alphabeta",
  noptim.rounds = 1,
  noptim.repeats = 1,
  optimize = TRUE,
  dmethod = "manhattan",
  p = 50
)

Arguments

n

run size of the SOA; power of 2, at least 16

m

number of colums; at most 5n/16 for constr="ShiTang_alpha" (exception: only 9 for n=32), at most n/4 for constr="ShiTang_alphabeta"; for m=NULL, defaults are m=5n/16 and m=n/4-1, respectively; the latter yields strength 3+.

constr

construction method. Must be one of "ShiTang_alphabeta", "ShiTang_alpha". See Details section

noptim.rounds

the number of optimization rounds for each independent restart

noptim.repeats

the number of independent restarts of optimizations with noptim.rounds rounds each

optimize

logical: should space filling be optimized by level permutations?

dmethod

distance method for phi_p, "manhattan" (default) or "euclidean"

p

p for phi_p (the larger, the closer to maximin distance)

Details

The construction is implemented as described in Groemping (2023a).

The 8-level SOAs created by this construction have strength 3 and at least the additional property alpha, which means that all pairs of columns achieve perfect 4x4 balance, if consecutive level pairs (01, 23, 45, 67) are collapsed.

The "ShiTang_alphabeta" construction additionally yields perfect 4x2x2 balance, if one column is collapsed to 4 levels, while two further columns are collapsed to 2 levels (0123 vs 4567). with m = n/4 columns, the "ShiTang_alphabeta" construction has a single pair of correlated columns, all other columns are uncorrelated, due to a modification of Shi and Tang's column allocation that was proposed in Groemping (2023a).

For m <= n/4 - 1, the "ShiTang_alphabeta" construction also yields perfect balance for 8x2 projections in 2D (i.e. if one original column with another column collapsed to two levels). Thus, it yields all strength 4 properties in 2D and 3D, which is called strength 3+. Furthermore, Groemping (2023a) proposed an improved choice of columns for matrix C that implies orthogonal columns in this case.

Value

matrix of class SOA with the attributes that are listed below. All attributes can be accessed using function attributes, or individual attributes can be accessed using function attr. These are the attributes:

type

the type of array (SOA or OSOA)

strength

character string that gives the strength

phi_p

the phi_p value (smaller=better)

optimized

logical indicating whether optimization was applied

permpick

matrix that lists the id numbers of the permutations used

perms2pickfrom

optional element, when optimization was conducted: the overall permutation list to which the numbers in permlist refer

call

the call that created the object

Author(s)

Ulrike Groemping

References

For full detail, see SOAs-package.

Groemping (2023a)
Shi and Tang (2020)
Weng (2014)

Examples

## use with optimization for actually using such designs
## n/4 - 1 = 7 columns, strength 3+
SOAs_8level(32, optimize=FALSE)

## n/4 = 8 columns, strength 3 with alpha and beta
SOAs_8level(32, m=8, optimize=FALSE)

## 9 columns (special case n=32), strength 3 with alpha
SOAs_8level(32, constr="ShiTang_alpha", optimize=FALSE)

## 5*n/16 = 5 columns, strength 3 with alpha
SOAs_8level(16, constr="ShiTang_alpha", optimize=FALSE)