Schur- and E-optimal two-level factorial designs
Schur-optimality is a very general class of optimality criteria that includes, as special cases, A- D- and E-optimality and Cheng Type 1 optimality. In this paper, Schur-optimal two-level factorial designs under a second-order model are derived for 3 and 5 factors for all numbers of runs where the model is estimable. In addition, orthogonal arrays of strength 4 (resolution V) with e added runs are shown to be E-optimal under a second-order model for e[less-than-or-equals, slant]10 and m=4 factors and for e[less-than-or-equals, slant]15 and m[greater-or-equal, slanted]5 factors. Corresponding results for third-order models are also given.
Year of publication: |
2008
|
---|---|
Authors: | Butler, Neil A. |
Published in: |
Statistics & Probability Letters. - Elsevier, ISSN 0167-7152. - Vol. 78.2008, 5, p. 518-527
|
Publisher: |
Elsevier |
Keywords: | Defining contrasts Fractional factorial Main effect Majorization Optimum design Regular design Two-factor interaction |
Saved in:
Online Resource
Saved in favorites
Similar items by person
-
Updating the forecast function of ARIMA models and the link with DLMs
Butler, Neil A., (1999)
-
Optimal additions to and deletions from two-level orthogonal arrays
Butler, Neil A., (2007)
-
Minimum G2-aberration properties of two-level foldover designs
Butler, Neil A., (2004)
- More ...