Title: Adaptive sequential experiment methodology for response surface optimisation

Authors: Adel Alaeddini, Kai Yang

Addresses: Department of Industrial and Manufacturing Engineering, Wayne State University, Detroit, MI 48201, USA. ' Department of Industrial and Manufacturing Engineering, Wayne State University, Detroit, MI 48201, USA

Abstract: Traditional response surface methodology (RSM) designs such as central composite design (CCD) use fixed factor settings to fit appropriate mathematical models. Despite their widespread applications, such approaches are not effective in terms of the number of required experiments and model precision. The number of required experiments shows its importance when economical aspects of each extra experiment are considered. Also, model precision reveals its significance when the problem deals with large factor space. Considering these facts, we introduce the methodology of adaptive sequential experiments on two-dimensional space which conducts fewer experiments on non-optimal region and more experiments on the optimal region. This methodology is based on a simple strategy: using previous experiments information for determining the factor setting of new experiments and shrinking the factor space to smaller region toward the optimal point. We show the effectiveness of these strategies by comparing the proposed approach with CCD on twelve numerical examples.

Keywords: adaptive sequential experiments; response surface methodology; RSM; response surface optimisation; central composite design; CCD; mathematical modelling.

DOI: 10.1504/IJQET.2009.030500

International Journal of Quality Engineering and Technology, 2009 Vol.1 No.1, pp.40 - 61

Published online: 18 Dec 2009 *

Full-text access for editors Access for subscribers Purchase this article Comment on this article