Table 6 Comparison of design results using several ROM methods (after 200 iterations)
Method | \(\varvec{N_{b}}\) | \(\varvec{\hat{e}_{rb}}\) | Calls to ROM | CPU running time \(\varvec{(/s)}\) | Optimal compliance \(\varvec{(c^{*})}\) | Relative error of \(\varvec{c^{*}(\%)}\) |
|---|---|---|---|---|---|---|
Reference | – | – | 0/300 | 23406.28 | 45.02 | - |
SF-ROM | 4 | 0.05 | 224/300 | 21388.52 | 45.02 | 0 |
MF-ROM | 4 | 0.05 | 225/300 | 19682.72 | 45.02 | 0 |
4 | 0.5 | 260/300 | 19578.27 | 45.07 | 0.11 | |
Adap-ROM | 4 | \(0\rightarrow 0.05\) \(\rightarrow 0.5 \) | 233/300 | 19655.47 | 45.01 | 0.02 |
Bi-ROM | 4 | 0.5 | 260/300 | 18663.07 | 45.09 | 0.16 |
6 | 0.1 | 281/300 | 18663.07 | 45.09 | 0.16 | |
AdapBi-ROM | 4 | \(0\rightarrow 0.05\) \(\rightarrow 0.5 \) | 233/300 | 18896.49 | 45.00 | 0.04 |
6 | \(0\rightarrow 0.01\) \(\rightarrow 0.1 \) | 254/300 | 18896.49 | 45.00 | 0.04 |