When the aim of an experiment is the estimation of a Generalised Linear Model (GLM), standard designsfrom linear model theory may prove inadequate. This paper describes a flexible approach for findingdesigns for experiments to estimate GLMs through the use of D-optimality and a simulated annealingalgorithm. A variety of uncertainties in the model can be incorporated into the design search, includingthe form of the linear predictor, through use of a robust design selection criterion and a postulatedmodel space. New methods appropriate for screening experiments and the incorporation of correlationsbetween possible model parameters are described through examples. An updating formula for Doptimalityunder a GLM is presented which improves the computational efficiency of the search.
展开▼
