In Central Composite Design (CCD), you specify a design type. Choices are:
Face-Centered
Rotatable
VIF-Optimality
Optimality
Auto Defined
A Rotatable (spherical) design is preferred because the prediction variance is the same for any two locations that are the same distance from the design center. However, there are other criteria to consider for an optimal design setup. The following two criteria are commonly considered in setting up an optimal design using the design matrix.
The degree of non-orthogonality of regression terms can inflate the variance of model coefficients.
The position of sample points in the design can be influential based on their position with respect to others of the input variables in a subset of the entire set of observations.
An optimal CCD design should minimize both the degree of non-orthogonality of term coefficients and the number of sample points having abnormal influence. In minimizing the degree of non-orthogonality, the Variation Inflation Factor (VIF) of regression terms is used. For a VIF-Optimality design, the maximum VIF of the regression terms is to be minimized, and the minimum value is 1.0. In minimizing the opportunity of influential sample points, the leverage value of each sample points is used. Leverages are the diagonal elements of the Hat matrix, which is a function of the design matrix. For a G-Optimality design, the maximum leverage value of sample points is to be minimized.
For a VIF-Optimality design, the alpha value/level is selected such that the maximum VIF is minimized. Likewise, for a G-Optimality design, the alpha value/level is selected such that the maximum leverage is minimized. The rotatable design is found to be a poor design in terms of VIF- and G-efficiencies.
For an optimal CCD, the alpha value/level is selected such that both the maximum VIF and the maximum leverage are the minimum possible. For an Auto Defined design, the alpha value is selected from either the VIF-Optimality or G-Optimality design that meets the criteria. Because it is a multi-objective optimization problem, in many cases, there is no unique alpha value such that both criteria reach their minimum. However, the alpha value is evaluated such that one criterion reaches minimum while the other approaches minimum.
For the current Auto Defined setup (except for a problem with five variables that uses a G-Optimality design) all other multi-variable problems use VIF-Optimality. In some cases, despite the fact that Auto Defined provides an optimal alpha meeting the criteria, this design might not give as good of a response surface as anticipated due to the nature of the physical data used for fitting in the regression process. In this case, you should try other design types that might give a better response surface approximation.
Note: You can set any design type for CCD as the default in the Options dialog box. On the Design of Experiments tab, set Design Type to the type to want to use as the default. For more information, see Design Exploration Options.
It is a good practice to always verify some selected points on the response surface with an actual simulation evaluation to determine its validity of use for further analyses. In some cases, a good response surface does not mean a good representation of an underlying physics problem. The response surface is generated according to the predetermined sampling points in the design space, which sometimes misses capturing an unexpected change in some regions of the design space. In this case, try using an enhanced design—a Rotatable or Face-Centered design type with Template Type set to Enhanced. For an enhanced DOE, a mini CCD is appended to a standard CCD design, where a second alpha value is added and set to half the alpha value of the standard CCD. The mini CCD is set up so that the rotatability and symmetry of the CCD design are maintained. The purposes of the appended mini CCD are to capture any drastic changes in the design space and to provide a better response surface fit.
Note: Alternatively, you can try to enrich the DOE by changing the selection for Design of Experiments Type from Central Composite Design to Custom + Sampling and then specifying the Total Number of Samples. For more information, see Custom + Sampling.
The location of the generated design points for the deterministic
method is based on a central composite design. If N is the number of input parameters, then a central composite design
consists of:
One center point.
2*N axis point located at the -α and +α position on
each axis of the selected input parameters.
2(N-f)
factorial points located at the -1 and +1 positions along the diagonals of the input
parameter space.
The fraction f of the factorial design
and the resulting number of design points are given in the following
table:
Number of Generated Design Points as a Function of the Number of Input Parameters
| Number of Input Parameters | Factorial Number f | Number of Design Points |
|---|---|---|
| 1 | 0 | 5 |
| 2 | 0 | 9 |
| 3 | 0 | 15 |
| 4 | 0 | 25 |
| 5 | 1 | 27 |
| 6 | 1 | 45 |
| 7 | 1 | 79 |
| 8 | 2 | 81 |
| 9 | 2 | 147 |
| 10 | 3 | 149 |
| 11 | 4 | 151 |
| 12 | 4 | 281 |
| 13 | 5 | 283 |
| 14 | 6 | 285 |
| 15 | 7 | 287 |
| 16 | 8 | 289 |
| 17 | 9 | 291 |
| 18 | 9 | 549 |
| 19 | 10 | 551 |
| 20 | 11 | 553 |