A Bayesian design criterion for locating the optimum point on a response surface

Gilmour, S. G. and Mead, R. (2003) A Bayesian design criterion for locating the optimum point on a response surface. Statistics & Probability Letters, 64 (3). pp. 235-242. ISSN 0167-7152 doi: 10.1016/s0167-7152(03)00154-8

Abstract/Summary

Most factorial experiments in industrial research form one stage in a sequence of experiments and so considerable prior knowledge is often available from earlier stages. A Bayesian A-optimality criterion is proposed for choosing designs, when each stage in experimentation consists of a small number of runs and the objective is to optimise a response. Simple formulae for the weights are developed, some examples of the use of the design criterion are given and general recommendations are made. (C) 2003 Elsevier B.V. All rights reserved.