Two Statistics and Machine Learning Toolbox™ algorithms generate D-optimal designs:
In practice, you may want to add runs to a completed experiment to learn more about a process and estimate additional model coefficients.
In many experimental settings, certain factors and their covariates are constrained to a fixed set of levels or combinations of levels.
Categorical factors take values in a discrete set of levels.
The row-exchange algorithm exchanges rows of an initial design matrix X with rows from a design matrix C evaluated at a candidate set of feasible treatments.
Traditional experimental designs (Full Factorial Designs (Statistics and Machine Learning Toolbox), Fractional Factorial Designs (Statistics and Machine Learning Toolbox), and Response Surface Designs (Statistics and Machine Learning Toolbox)) are appropriate for calibrating linear models in experimental settings where factors are relatively unconstrained in the region of interest.