Table 3 : Speed-up(x-times) for each partitioned algorithm compared to non-partitioned COBYLA
From: Partition-based optimization model for generative anatomy modeling language (POM-GAML)
# of Cons affected | k-means-10 | k-means-5 | CNM | DBSCAN-10 | DBSCAN-5 |
|---|---|---|---|---|---|
1500 | 73.17 | 108.64 | 97.18 | 38.15 | 82.11 |
2000 | 159.55 | 498.42 | 47.57 | 91.92 | 474.40 |
2500 | 123.63 | 449.27 | 36.99 | 64.96 | 567.79 |
3000 | 181.20 | 1036.72 | 78.81 | 46.24 | 849.14 |
3500 | 118.93 | 1610.41 | 62.89 | 15.32 | 1248.50 |
4000 | 86.77 | 1068.89 | 578.53 | 12.95 | 927.99 |
4500 | 231.31 | 2218.13 | 701.81 | 18.38 | 477.99 |
5000 | 510.19 | 2677.52 | 601.19 | 13.18 | 2689.72 |