An In-depth Benchmarking of Evolutionary and Swarm Intelligence Algorithms for Autoscaling Parameter Sweep Applications on Public Clouds
Table 7
Average makespan, average monetary cost, average number of task failures, and average L2-norm obtained by each variant of the autoscaler MOEA for applications MPG and GMS. For all metrics, lower values are better values. For each application and size, bold values are better than those obtained by MOEA-NSGA-III.
Application
Size
Autoscaler
Makespan
Cost
Task failures
L2-norm
MPG
30
MOEA-NSGA-III
13162.40
1.71
0.10
0.33
MOEA-SMPSO
20809.32
1.57
0.20
0.74
MOEA-SMS-EMOA
12944.68
2.74
0.00
0.57
MOEA-E-NSGA-III
12838.21
1.52
0.00
0.26
100
MOEA-NSGA-III
14971.13
4.44
0.00
0.16
MOEA-SMPSO
23190.13
10.46
1.40
0.73
MOEA-SMS-EMOA
23575.42
12.23
0.00
0.77
MOEA-E-NSGA-III
13339.48
3.78
0.00
0.06
300
MOEA-NSGA-III
38160.13
37.45
0.60
0.67
MOEA-SMPSO
31377.70
30.23
1.77
0.50
MOEA-SMS-EMOA
47792.46
63.50
0.57
0.88
MOEA-E-NSGA-III
31839.27
30.64
0.00
0.46
GMS
30
MOEA-NSGA-III
1074149.61
531.36
4.38
0.46
MOEA-SMPSO
1422601.27
456.40
4.29
0.49
MOEA-SMS-EMOA
1033579.47
680.52
5.43
0.62
MOEA-E-NSGA-III
1050575.97
406.47
2.77
0.30
100
MOEA-NSGA-III
1021845.42
1731.66
13.30
0.51
MOEA-SMPSO
1382518.24
1612.51
21.44
0.72
MOEA-SMS-EMOA
1079659.30
2287.06
6.41
0.58
MOEA-E-NSGA-III
915882.64
1635.15
6.07
0.38
300
MOEA-NSGA-III
1021592.20
5461.36
22.18
0.30
MOEA-SMPSO
1353654.86
5080.45
67.25
0.52
MOEA-SMS-EMOA
1355956.14
7660.57
42.00
0.56
MOEA-E-NSGA-III
918339.45
4613.51
14.64
0.19
indicate the best value. The baseline is in italics.