Nowadays split-plot designs play a crucial role in technological fields, both for their flexibility when applying a robust design approach and in relation to the modelling step, by considering mixed Response Surface models and/or the class of Generalized Linear Mixed Models (GLMMs). In this paper, a split-plot design is studied in a process optimization scenario involving several response variables, a multi-response situation, where two optimization methods are compared. More precisely, by considering a real case study related to the improvement of a measurement process of a Numerical Control machine for measuring dental implants, the optimization is carried out with the Pareto front approach and then compared with an analytical optimization method obtained starting from the definition of a risk function. In the final discussion advantages and disadvantages of application for both methods are evaluated.
Split-Plot Designs and Multi-Response Process Optimization: a Comparison Between Two Approaches / Rossella Berni, Lorenzo Piattoli, Christine M. Anderson-Cook, Lu Lu. - In: STATISTICA APPLICATA. - ISSN 2038-5587. - ELETTRONICO. - 34:(2022), pp. 1-16. [10.26398/IJAS.0034-004]
Split-Plot Designs and Multi-Response Process Optimization: a Comparison Between Two Approaches
Rossella Berni
Methodology
;
2022
Abstract
Nowadays split-plot designs play a crucial role in technological fields, both for their flexibility when applying a robust design approach and in relation to the modelling step, by considering mixed Response Surface models and/or the class of Generalized Linear Mixed Models (GLMMs). In this paper, a split-plot design is studied in a process optimization scenario involving several response variables, a multi-response situation, where two optimization methods are compared. More precisely, by considering a real case study related to the improvement of a measurement process of a Numerical Control machine for measuring dental implants, the optimization is carried out with the Pareto front approach and then compared with an analytical optimization method obtained starting from the definition of a risk function. In the final discussion advantages and disadvantages of application for both methods are evaluated.I documenti in FLORE sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.