Information Technology Reference
In-Depth Information
20
15
Layer 1
10
Layer 2
Layer 3
5
0
0
5
10
15
20
25
30
Fig. 13.
Cross-section for example 3
Table 4.
The orthogonal test table for example 3
Results
λ
HMCR
PAR
NHS1
NHS2
1
1 (0.85)
1 (0.10)
1 (0.3)
1.4843
1.5419
2
1 (0.85)
2 (0.15)
2 (0.4)
1.5074
1.5462
3
1 (0.85)
3 (0.20)
3 (0.5)
1.5207
1.6225
4
1 (0.85)
4 (0.25)
4 (0.6)
1.5461
1.5227
5
2 (0.90)
1 (0.10)
2 (0.4)
1.5239
1.4814
6
2 (0.90)
2 (0.15)
1 (0.3)
1.4752
1.4872
7
2 (0.90)
3 (0.20)
4 (0.6)
1.5047
1.5345
8
2 (0.90)
4 (0.25)
3 (0.5)
1.5267
1.5215
9
3 (0.95)
1 (0.10)
3 (0.5)
1.4831
1.4681
10
3 (0.95)
2 (0.15)
4 (0.6)
1.4817
1.4923
11
3 (0.95)
3 (0.20)
1 (0.3)
1.4711
1.5107
12
3 (0.95)
4 (0.25)
2 (0.4)
1.5415
1.5385
13
4 (1.00)
1 (0.10)
4 (0.6)
1.4579
1.4779
14
4 (1.00)
2 (0.15)
3 (0.5)
1.4702
1.4970
15
4 (1.00)
3 (0.20)
2 (0.4)
1.4857
1.4870
16
4 (1.00)
4 (0.25)
1 (0.3)
1.5047
1.4828
to the analysis. It should be noted that the parameters used in the search algorithms
are not sensitive for the simple examples, but they can be sensitive or hypersensitive
for complicated examples. It is suggested that engineers should perform an orthogonal
analysis for complicated problems. The authors have also found similar behavior for
genetic algorithm and ant colony method. Actually, all stochastic global optimization
Search WWH ::
Custom Search