The precision of estimating the parameters of a statistical model depends to a large extent on the design used in empirical studies. The design specifies the levels of the predictors that influence the response. An optimal design minimizes the number of experimental runs."
