Transformation robust experimental design with application to some problems in chemistry

In this paper we consider the selection of an appropriate experimental design when the exact form of the error distribution is unknown. The goal of error-robust design is to design an experiment so that the 'ill-effects' resulting from a lack of knowledge of the error structure will be minimal. Numerical algorithms for computer construction of error-robust designs are developed and the method is illustrated in connection with the design of experiments for nonlinear modeling of chemical reactions.