The design of experiments for discriminating between two rival models

Abstract

Sequential experimental designs are developed for discriminating between two rival regression models, which need not be linear in the parameters. Nonsequential procedures are also derived and numerical examples given of both classes of design.