Moving Clusters within a Memetic Algorithm for Graph Partitioning
Table 4
Comparison of KL-MA and CMPA on bipartitioning.
Graph
Best known1
KL-MA
CMPA
Best2
Ave2
CPU3 (Gen4)
Best2
Ave2
CPU3 (Gen4)
G500.2.5
49
49
51.34
0.32 (595)
49
51.22
0.34 (558)
G500.05
218
218
218.65
0.59 (700)
218
218.45
0.64 (682)
G500.10
626
626
627.55
1.02 (681)
626
627.46
1.10 (672)
G500.20
1744
1744
1745.83
1.93 (694)
1744
1745.62
2.09 (687)
G1000.2.5
93
93
97.26
0.96 (745)
93
97.08
1.04 (722)
G1000.05
445
445
451.50
2.07 (884)
445
450.92
2.22 (860)
G1000.10
1362
1362
1366.68
4.07 (969)
1362
1366.18
4.45 (977)
G1000.20
3382
3382
3385.61
7.42 (971)
3382
3385.26
8.13 (987)
U500.05
2
2
4.96
0.55 (793)
2
3.94
0.66 (803)
U500.10
26
26
26.31
0.45 (459)
26
26.04
0.57 (465)
U500.20
178
178
178.06
0.42 (300)
178
178.00
0.53 (271)
U500.40
412
412
412.00
0.34 (190)
412
412.00
0.37 (102)
U1000.05
1
1
11.67
1.71 (1077)
1
8.81
2.12 (1068)
U1000.10
39
39
46.27
1.51 (645)
39
44.15
1.95 (653)
U1000.20
222
222
222.19
1.13 (347)
222
222.07
1.51 (327)
U1000.40
737
737
737.00
1.20 (243)
737
737.00
1.77 (205)
Best result from the literature. 2Best and average results from 1,000 runs. 3CPU seconds on a Pentium IV 2.8 GHz. 4Average number of generations over 1,000 runs.