A richer set of points for the DOE usually provides a more accurate response surface. However, when you select a response surface type other than Genetic Aggregation, which is the default, you generally need to specify how many additional points are required, which can be difficult without actually generating the response surface.
The first step to solving this problem is to set Response Surface Type to Kriging. This response surface type determines the accuracy of the response surface as well as the number of points that are required to increase the accuracy. With this method, you can set a manual or automatic refinement type. Manual refinement allows you to control the number of points to compute.
Sparse Grid is yet another of the response surface types available. This adaptive response surface is driven by the accuracy that you request. It automatically refines the matrix of design points where the gradient of the output parameters is higher to increase the accuracy of the response surface. To use this feature, you set Response Surface Type to Sparse Grid and Design of Experiments Type to Sparse Grid Initialization.
While no manual refinement is available when Sparse Grid is selected, manual refinement is available for other response surface types. With manual refinement, you can enter specific points into the set of existing design points used to calculate the response surface.
For more information, see Response Surface Types.