Comparer des méthodes

	Conception de Box-Behnken assistée par optimisation	Méthodologie de surface de réponse assistée par optimisation
Domaine	Plans d'expériences	Plans d'expériences
Famille	Process / pipeline	Process / pipeline
Année d'origine	1960 (BBD); optimization integration established 1980s–1990s	1951 (RSM); 1980 (desirability-function optimization formalized)
Auteur d'origine	Box & Behnken (design); Derringer & Suich (desirability optimization)	Derringer & Suich (desirability function); Box & Wilson (RSM foundation)
Type	Experimental design with post-modeling optimization	Hybrid experimental-optimization framework
Source fondatrice	Box, G. E. P., & Behnken, D. W. (1960). Some new three level designs for the study of quantitative variables. Technometrics, 2(4), 455–475.	Derringer, G., & Suich, R. (1980). Simultaneous optimization of several response variables. Journal of Quality Technology, 12(4), 214–219.
Alias	BBD with optimization, Box-Behnken design optimization, RSM-BBD optimization, Box-Behnken response optimization	OA-RSM, RSM with optimization, desirability-based RSM, multi-response RSM optimization
Apparentées	5	5
Résumé	Optimization-assisted Box-Behnken design (BBD) combines the Box-Behnken three-level experimental design with a formal optimization step to locate factor settings that maximize, minimize, or hit a target for one or more responses. BBD fits a second-order response surface model using fewer runs than a full factorial, and the optimization stage — typically via desirability functions or numerical search — then exploits that fitted model to identify the true optimum within the experimental region.	Optimization-assisted RSM couples a second-order response surface model with a mathematical optimization routine — most commonly Derringer and Suich's desirability function, but also genetic algorithms or gradient-based solvers — to locate the factor settings that simultaneously satisfy multiple quality or performance objectives. The result is a data-driven recommendation for optimal process or product conditions, supported by a polynomial model fitted to a structured experimental design.