B1386
Title: Exploring composite design: Investigating alternatives in response surface methodology
Authors: Despina Athanasaki - RMIT (Australia) [presenting]
Abstract: In the realm of experimental designs, the D-value serves as an established measure for assessing the minimum detect table effect. Additionally, there has been a growing popularity of DefinitiveScreening Designs (DSDs) in recent years, especially for investigating second-order effects in response surface methodology (RSMs). Two alternative composite design approaches are introduced, aimed at providing more efficient designs based on the D-value criterion. Orthogonal matrices are utilised to create new composite designs and augment them with definitive screening designs. In the process, an axial component is also incorporated using either orthogonal designs or block orthogonal designs, as demonstrated in the previous research. Furthermore, new, improved design matrices are also introduced, derived from existing construction methods. Importantly, all the designs presented demonstrate enhancements in the D-value criterion for the full second-order model when compared to any other known design from the existing literature.
