In a 2^k-factorial experiment with limited resources, when practitioners can identify the non-negligible effects and interactions beforehand, it is common to run an experiment with a saturated design that ensures the unbiased estimation of the non-negligible parameters of interest. We propose a method for the construction of D-optimal saturated designs for the mean, the main effects, and the second-order interactions of one factor with the remaining factors. In the process, we show the problem is just as hard as the Hadamard determinant problem.

References

• Chadjipantelis, T., Kounias, S., & Moyssiadis, C. (1987). The maximum determinant of 21×21 (+1, -1)-matrices and D-optimal designs. Journal of Statistical Planning and Inference, 16(2), 167–178. https://doi.org/10.1016/0378-3758(87)90066-8
• Cohn, J. H. E. (1989). On determinants with elements ±1, II. Bulletin of the London Mathematical Society, 21(1), 36–42. https://doi.org/10.1112/blms/21.1.36
• Cohn., J. H. E. (2000). Almost D-optimal designs. Utilitas Mathematica, 57, 121–128.
• Hedayat, A. S., & Pesotan, H. (1992). Two-level factorial designs for main-effects and selected two-factor interactions. Statistica Sinica, 2, 453–464.
• Hedayat, A. S., & Pesotan, H. (2007). Tools for constructing optimal two-level factorial designs for a linear model containing main effects and one two-factor interaction. Journal of Statistical Planning and Inference, 137(4), 1452–1463. https://doi.org/10.1016/j.jspi.2006.04.005
• Hedayat, A. S., & Zhu, H. (2011). An effective algorithm for searching for D-optimal saturated two-level factorial designs. Journal of Statistical Theory and Applications, 10(2), 209–227.
• Kiefer, J. (1959). Optimum experimental designs. Journal of the Royal Statistical Society, B21, 272–319.
• Metropolis, N. (1969). Spectra of determinant values in (0,1) matrices. In A. O. L. Atkin & B. J. Birch (Eds.), Computers in Number Theory: Proceedings of the Science Research Atlas Symposium No. 2 held at Oxford, 18–23 August 1969 (pp. 271–276). Academic Press.
• Orrick, W. P. (2005). The maximal {−1,1}-determinant of order 15. Metrika, 62(2-3), 195–219. https://doi.org/10.1007/s00184-005-0410-3
• Phoa, F. K. H., Wong, W. K., & Xu, H. (2009). The need of considering the interactions in the analysis of screening designs. Journal of Chemometrics, 23(10), 545–553. https://doi.org/10.1002/cem.v23:10
• Smith, Warren D. (1988). Studies in computational geometry motivated by mesh generation [Unpublished doctoral dissertation]. Princeton University.
• Vander Heyden, Y., Jimidar, M., Hund, E., Niemeijer, N., Peeters, R., Smeyers-Verbeke, J., Massart, D. L., & Hoogmartens, J. (1999). Determination of system suitability limits with a robustness test. Journal of Chromatography A, 845(1-2), 145–154. https://doi.org/10.1016/S0021-9673(99)00328-3
• Wald, A. (1943). On the efficient design of statistical investigations. The Annals of Mathematical Statistics, 14(2), 134–140. https://doi.org/10.1214/aoms/1177731454
• Živković, M. (2006). Classification of small (0,1) matrices. Linear Algebra and Its Applications, 414(1), 310–346. https://doi.org/10.1016/j.laa.2005.10.010

To cite this article: Francois K. Domagni, A. S. Hedayat & Bikas Kumar Sinha (18 Apr 2024): D-optimal saturated designs for main effects and interactions in 2k -factorial experiments, Statistical Theory and Related Fields, DOI: 10.1080/24754269.2024.2341983

To link to this article: https://doi.org/10.1080/24754269.2024.2341983