Large supersaturated designs

K. M. Eskridge; S. G. Gilmour; R. Mead; N. A. Butler; D. A. Travnicek

doi:10.1080/00949650310001612436

Large supersaturated designs

K. M. Eskridge^*, S. G. Gilmour, R. Mead, N. A. Butler, D. A. Travnicek

^*Corresponding author for this work

Research output: Contribution to journal › Article › peer-review

29 Citations (Scopus)

Abstract

A supersaturated design (SSD) is an experimental plan, useful for evaluating the main effects of m factors with n experimental units when m > n - 1, each factor has two levels and when the first-order effects of only a few factors are expected to have dominant effects on the response. Use of these plans can be extremely cost-effective when it is necessary to screen hundreds or thousands of factors with a limited amount of resources. In this article we describe how to use cyclic balanced incomplete block designs and regular graph designs to construct E(s ²) optimal and near optimal SSDs when m is a multiple of n - 1. We also provide a table that can be used to construct these designs for screening thousands of factors. We also explain how to obtain SSDs when m is not a multiple of n - 1. Using the table and the approaches given in this paper, SSDs can be developed for designs with up to 24 runs and up to 12,190 factors.

Original language	English
Pages (from-to)	525-542
Number of pages	18
Journal	JOURNAL OF STATISTICAL COMPUTATION AND SIMULATION
Volume	74
Issue number	7
DOIs	https://doi.org/10.1080/00949650310001612436
Publication status	Published - Jul 2004

Keywords

Computer-aided design
Design of experiments
Incomplete block design
Regular graph design
Screening design

Access to Document

10.1080/00949650310001612436

Cite this

@article{c30fc7665b6a4a88a13d5ee386572f14,

title = "Large supersaturated designs",

abstract = "A supersaturated design (SSD) is an experimental plan, useful for evaluating the main effects of m factors with n experimental units when m > n - 1, each factor has two levels and when the first-order effects of only a few factors are expected to have dominant effects on the response. Use of these plans can be extremely cost-effective when it is necessary to screen hundreds or thousands of factors with a limited amount of resources. In this article we describe how to use cyclic balanced incomplete block designs and regular graph designs to construct E(s 2) optimal and near optimal SSDs when m is a multiple of n - 1. We also provide a table that can be used to construct these designs for screening thousands of factors. We also explain how to obtain SSDs when m is not a multiple of n - 1. Using the table and the approaches given in this paper, SSDs can be developed for designs with up to 24 runs and up to 12,190 factors.",

keywords = "Computer-aided design, Design of experiments, Incomplete block design, Regular graph design, Screening design",

author = "Eskridge, {K. M.} and Gilmour, {S. G.} and R. Mead and Butler, {N. A.} and Travnicek, {D. A.}",

year = "2004",

month = jul,

doi = "10.1080/00949650310001612436",

language = "English",

volume = "74",

pages = "525--542",

journal = "JOURNAL OF STATISTICAL COMPUTATION AND SIMULATION",

issn = "0094-9655",

publisher = "Taylor and Francis Ltd.",

number = "7",

}

TY - JOUR

T1 - Large supersaturated designs

AU - Eskridge, K. M.

AU - Gilmour, S. G.

AU - Mead, R.

AU - Butler, N. A.

AU - Travnicek, D. A.

PY - 2004/7

Y1 - 2004/7

N2 - A supersaturated design (SSD) is an experimental plan, useful for evaluating the main effects of m factors with n experimental units when m > n - 1, each factor has two levels and when the first-order effects of only a few factors are expected to have dominant effects on the response. Use of these plans can be extremely cost-effective when it is necessary to screen hundreds or thousands of factors with a limited amount of resources. In this article we describe how to use cyclic balanced incomplete block designs and regular graph designs to construct E(s 2) optimal and near optimal SSDs when m is a multiple of n - 1. We also provide a table that can be used to construct these designs for screening thousands of factors. We also explain how to obtain SSDs when m is not a multiple of n - 1. Using the table and the approaches given in this paper, SSDs can be developed for designs with up to 24 runs and up to 12,190 factors.

AB - A supersaturated design (SSD) is an experimental plan, useful for evaluating the main effects of m factors with n experimental units when m > n - 1, each factor has two levels and when the first-order effects of only a few factors are expected to have dominant effects on the response. Use of these plans can be extremely cost-effective when it is necessary to screen hundreds or thousands of factors with a limited amount of resources. In this article we describe how to use cyclic balanced incomplete block designs and regular graph designs to construct E(s 2) optimal and near optimal SSDs when m is a multiple of n - 1. We also provide a table that can be used to construct these designs for screening thousands of factors. We also explain how to obtain SSDs when m is not a multiple of n - 1. Using the table and the approaches given in this paper, SSDs can be developed for designs with up to 24 runs and up to 12,190 factors.

KW - Computer-aided design

KW - Design of experiments

KW - Incomplete block design

KW - Regular graph design

KW - Screening design

UR - http://www.scopus.com/inward/record.url?scp=2942620518&partnerID=8YFLogxK

U2 - 10.1080/00949650310001612436

DO - 10.1080/00949650310001612436

M3 - Article

AN - SCOPUS:2942620518

SN - 0094-9655

VL - 74

SP - 525

EP - 542

JO - JOURNAL OF STATISTICAL COMPUTATION AND SIMULATION

JF - JOURNAL OF STATISTICAL COMPUTATION AND SIMULATION

IS - 7

ER -

Large supersaturated designs

Abstract

Keywords

Access to Document

Other files and links

Fingerprint

Cite this