Near D-optimal designs
Since the
optimal added points for the minimal-point design based on 2k fractional factorial design with resolution III* and Plackett
and Burman design may not be useful in practice, we propose the near D-optimal designs by rounding optimal
added supports.
1.
Minimal-point
composite design based on 2k fractional factorial design with resolution III*
1.
Minimal-point
composite design with Plackett and Burman Design
three
factors、five factors、seven factors、eight factors
2.
Comparison
We use
relative efficiency to compare minimal-point design with near D-optimal designs, and the ratios are
shown in the last column of Table 7.
Factor |
Minimal-point design |
Near D-optimal design |
Relative efficiency |
3 |
0.6680 |
0.6680 , 0.6611 |
1 , 0.9896 |
4 |
0.7115 |
0.7113 |
0.9997 |
5 |
0.7580 |
0.7561 |
0.9975 |
6 |
0.7810 |
0.7814 |
1.0005 |
7 |
0.6868 |
0.6808 |
0.9913 |
8 |
0.6278 |
0.6224 , 0.6220 |
0.9914 , 0.9908 |
Table 7. The Peff values of minimal-point
designs and near-optimal designs.