Mixture Designs

Points to Remember

Mixture design factors are components in a blend.
The levels of the components in a mixture must sum to one, so you need to make special considerations when designing and analyzing a mixture experiment. Refer to Cornell (1990) for details.
Mixtures can be pure (one factor) or a blend of factors.
You can use the interactive features of the prediction profiler and contour optimizer to find optimal mixtures.
You can underlay the standard error of the prediction as a shaded gradient so that you can find the minimum prediction error for your response.