Tailor-Made Split-Plot Designs for Mixture and Process Variables

• Autores: Peter Goos, Alexander N. Donev
• Localización: Journal of quality technology: A quarterly journal of methods applications and related topics, ISSN 0022-4065, Vol. 39, Nº. 4, 2007, págs. 326-339
• Idioma: inglés
• Enlaces
  • Texto completo
• Resumen

  • A complication arising in the design of small experiments involving mixture or process variables is that these experiments are often conducted using split-plot designs that lead to correlated observations. It is shown how algorithmic search can be used to construct tailor-made split-plot mixture-process designs when there might be constraints on the mixture components. The D-optimality criterion is the main design criterion used. In addition, it is shown how to construct split-plot mixture-process variable designs when replication is required for estimating variance components.