Blocking Factorial Experiment with Unequal Number of Replications for Factor Levels
The blocking of factorial designs become a complex task when there are many factors with multiple levels. Often, in this case, incomplete blocks are needed, as in some agronomy experiments. This paper examines a situation in which complete (orthogonal) block designs are impossible for the factorial design with no initial evident of combinatorics to design finding. One of the interesting problems we deal with is unequal number of replicates of factor levels, this is mainly due to cost restrains of practical experimentation. The presented example involves a garlic vernalization study with a factorial structure of 2^{2} × 3 × 4. We developed a tailor-made search algorithm to find a locally optimal (or near-optimal) design using an interchange strategy suited to the irregular factorial structure. The resulting design has interesting combinatorial properties that could not have been envisaged in advance. The methods used in this paper are not new to design theory, but are considerably underused in agronomical trials. The design itself is unprecedented in horticultural practice.
Authors
- Felipe O. Silva (ORCID: https://orcid.org/0000-0002-3715-7023)
- Júlio Sílvio de Sousa Bueno Filho (ORCID: https://orcid.org/0000-0002-3243-0959)
- SAKA Adisa Jamiu (ORCID: https://orcid.org/0000-0003-0276-5579)
Institutions
- Universidade Federal de Lavras (BR)
- Obafemi Awolowo University (NG)
Publication Details
- Journal
- Brazilian Journal of Biometrics
- Published
- 2026-11-13
- DOI
- https://doi.org/10.28951/bjb.v44i4.1017
- Primary Topic
- Optimal Experimental Design Methods
- Type
- article
- Field-Weighted Citation Impact
- 0.00
Funders
- Tertiary Education Trust Fund
- Universidade Federal de Lavras