Optimal (Custom) Response Surface Design

The optimal response surface design’s main use is to cope with multi-linear constraints on the factor settings. The factors can be both numeric and categoric. Numeric factor’s type can be changed from continuous to discrete. This allows the experimenter to limit the numeric factors to specific settings, but still fit a continuous polynomial.

On the first screen, choose the number of factors. Enter the factor names, units, and value for the low and high limits. Click Next to choose the designed for model, number of blocks, and number of runs.

The designed for model and number of blocks determine how many runs are required. Five lack of fit points and five replicates are recommended to improve the capabilities of the design and are included by default.

The required model points are the minimum number of runs necessary to estimate the coefficients in the designed for model. Additional model points will improve the precision of these estimates.

Block names (if required) are entered after the clicking Next.

Click Next to enter the response information and Finish to build and evaluate the design.