Model-robust experimental designs for the fractional polynomial response surface models

Abstract

Fractional polynomial response surface models are polynomial models whose powers are restricted to a small predefined set of rational numbers. Very often these models can give a good a fit to the data and much more plausible behavior between design points than the polynomial models. In this paper, we propose a one-stage and two-stage design strategy for obtaining designs under model uncertainty for these nonlinear class of models.

Keywords: nonlinear, model-robust, lack of fit, nesting strategy, support points, locally optimal designs.