Next: 3.1.17 Transformations
Up: 3.1 Design of Experiments
Previous: 3.1.15 Supplemantary Design
For higher order Factorial design the number of design points grows
rapidly. In the Three Level Factorial design all possible
combinations of the three discrete values of the parameter are used.
In Table 3.9 a comparison between the number of
experiments of a full Three Level Factorial design and other designs are shown.
The number of design points can be reduced by skipping some higher
order interactions between the input parameters.
Table 3.9:
Comparison of the number of sample points for
several types of experimental designs.
|
|
Sensitivity analysis design
|
|
|
Three level full factorial design
|
|
2 |
4 |
4 |
9 |
6 |
3 |
8 |
6 |
15 |
27 |
4 |
16 |
8 |
25 |
81 |
5 |
32 |
10 |
43 |
243 |
6 |
64 |
12 |
77 |
729 |
7 |
128 |
14 |
143 |
2187 |
8 |
256 |
16 |
273 |
6561 |
9 |
512 |
18 |
531 |
19683 |
10 |
1024 |
20 |
1045 |
59049 |
11 |
2048 |
22 |
2071 |
177147 |
12 |
4096 |
24 |
4121 |
531441 |
|
|
|
|
|
n |
2n |
|
|
3n |
Next: 3.1.17 Transformations
Up: 3.1 Design of Experiments
Previous: 3.1.15 Supplemantary Design
R. Plasun