Optimum designs versus orthogonal arrays for main effects and two-factor interactions
article
Designs with full estimation capacity permit estimation of all main effects and all two-factor interactions. By allowing correlation among the effects, the run size of such designs can be smaller than required for a resolution of 5. To construct a design, one can either use commercial software for designs with optimized D-efficiencies or a catalog of orthogonal arrays. In the context of a wood construction experiment, we discuss how to choose between these approaches. We enumerate mixed-level and multilevel resolution-4 designs with run size up to 72 and with the maximum number of factors compatible with a full estimation capacity. Algorithmically constructed benchmark designs were generated with commercial software. Our study results in a list of recommended designs.
TNO Identifier
427717
ISSN
00224065
Source
Journal of Quality Technology, 42(2), pp. 197-208.
Pages
197-208
Files
To receive the publication files, please send an e-mail request to TNO Repository.