Small, balanced, efficient, optimal, and near rotatable response surface designs for factorial experiments asymmetrical in some quantitative, qualitative factors

摘要

We present a structured method of constructing cost-efficient response surface designs (RSDs) as compared to the replicated central composite designs (RCCDs), that are useful for modelling and optimization of the experiments asymmetric in some qualitative, quantitative factors with at least two unrestricted quantitative factors while the remaining take two or three levels. We demonstrate the method by designing various experimental situations in 3 to 6 factors, and the analysis competency of the RSDs by analysing a published data set from an optimization experiment. The structural and prediction properties make our RSDs a good alternative to the known RSDs.