Supersaturated split-plot designs for industrial experimentation
Kalliopi Mylona (King's College London)
Supersaturated split-plot experiments combine two classes of designs which are important for industrial experimentation; screening designs and designs with restrictions on randomisation due to hard-to-change factors or two-stage processes. Although such designs are prevalent in industry, the literature on them is limited. We propose an optimal design approach and present Bayesian optimality criteria to find these designs. The analysis of supersaturated split-plot designs is complicated by the correlation of columns in the model matrix and the estimation of two variance components. We propose a novel analysis method for responses from these designs, which includes empirical Bayes and coordinate descent. Industrial examples from materials and pharmaceutical sciences are used to demonstrate new approaches to both the design and analysis of such supersaturated split-plot experiments.
Joint work with Emily Matthews and Dave Woods.