K. Mylona, S. G. Gilmour, P. Goos
We present a novel approach to design blocked and split-plot experiments. Textbooks on response surface methodology generally stress that experiments should allow for pure-error estimation and should involve replicated treatments. Moreover, sometimes the experimental runs cannot be performed under homogeneous circumstances, in which case the experiment is blocked. On other occasions, the experiment involves hard-to-change factors or two-stage processes. In that case, we have a split-plot experiment. This complicates the design and the analysis of the experiment. For the analysis, a mixed regression model with two variance components is required. The key feature of our approach is that it ensures that the two variance components can be estimated from pure error, in addition to a precise estimation of the response surface model. Our approach involves a new Bayesian compound D-optimal design criterion which pays attention to both the variance components and the fixed treatment effects.
Palabras clave / Keywords: model-independent variance component estimates, restricted maximum likelihood (REML), restricted randomization, treatment model
Programado
Sesión GT07-1 Diseño de Experimentos: Advances in Experimental Design (OED-1). Organizadores: Víctor Casero-Alonso y Jesús López-Fidalgo
30 de mayo de 2018 10:50
Sala 6
