|
Title
|
|
|
|
Enumeration and multicriteria selection of orthogonal minimally aliased response surface designs
| |
|
Author
|
|
|
|
| |
|
Abstract
|
|
|
|
| Response surface designs (RSDs) are a core component of the response surface methodology, which is widely used in the context of product and process optimization. In this contribution, we consider three-level RSDs, which can be viewed as matrices with entries equal to . Each column of an RSD corresponds to a factor and each row to an experimental test. We define a new family of orthogonal RSDs, for which there is no aliasing between the main effects and the second-order effects (two-factor interactions and quadratic effects). Using integer programming techniques, we construct a database of 55,531 such RSDs for 3-7 factors. We name these designs orthogonal minimally aliased RSDs (or OMARS designs). Each design in the catalog is extensively characterized in terms of efficiency, power, fourth-order correlations, fraction of design space plots, projection capabilities, etc. We identify interesting designs and investigate trade-offs between different quality criteria. Finally, we present a multiattribute decision algorithm to select designs from the catalog. An important result of our study is that we discovered some novel and interesting designs that challenge standard RSDs. |
| |
|
Language
|
|
|
|
English
| |
|
Source (journal)
|
|
|
|
Technometrics : a journal of statistics for the physical, chemical, and engineering sciences. - Washington, D.C., 1959, currens
| |
|
Publication
|
|
|
|
Washington, D.C.
:
2020
| |
|
ISSN
|
|
|
|
0040-1706
[print]
1537-2723
[online]
| |
|
DOI
|
|
|
|
10.1080/00401706.2018.1549103
| |
|
Volume/pages
|
|
|
|
62
:1
(2020)
, p. 21-36
| |
|
ISI
|
|
|
|
000510334100003
| |
|
Full text (Publisher's DOI)
|
|
|
|
| |
|
Full text (publisher's version - intranet only)
|
|
|
|
| |
|