Robustness Of Two Factor Response Surface Designs Within Split-Plot Structure Against Missing Observations

In this paper we presents Minimaxloss criterion for fixed variance ratio, based on D-,G- and A-efficiency, and which accounts within-whole plot and sub plot singlemissing observations, for constructing two factor robust designs with in split-plot response surface designs structure that is robust to a single missing observation at different designs points for fixed variance ratio. A missing observation make the outputs of Response Surface Modal is difficult. Then comparison of the loss in the data due to a single missing values by numerically computations for A-optimality, D- optimality and G-optimality. In these situations the effect of missing observations will also be studied on the basis of graphical tools like, Variance Dispersion Graph (VDG).