Virtual Machine Replica Placement Using a Multiobjective Genetic Algorithm

Table 10

The LSD test regarding utilization of first-fit, next-fit, best-fit, PSO, NSGA-III, and MOEA/D algorithms.

	Algorithm 1	Algorithm 2	Diff.	lwr.ci	upr.ci	Sig.
Number of PMs	NSGA-III	Best-fit	−1016.00	−1377.00	−655.00	0.000051
	NSGA-III	First-fit	−67.67	−428.66	293.33	0.69017
	NSGA-III	MOEA/D	−172.67	−533.66	188.33	0.31789
	NSGA-III	Next-fit	−1219.33	−1580.33	−858.34	0.0000087
	PSO	NSGA-III	381.17	20.17	742.16	0.04016
Performance degrading	NSGA-III	Best-fit	−541.50	−658.41	−424.59	3.20E − 07
	NSGA-III	First-fit	−508.83	−625.75	−391.92	6.30E − 07
	NSGA-III	MOEA/D	−148.83	−265.75	−31.92	0.0168
	NSGA-III	Next-fit	−520.17	−637.08	−403.25	5.00E − 07
	PSO	NSGA-III	455.00	338.09	571.91	2.10E − 06
Fitness values	NSGA-III	Best-fit	−21.20	−25.83	−16.58	3.60E − 07
	NSGA-III	First-fit	−18.14	−22.77	−13.52	1.90E − 06
	NSGA-III	MOEA/D	−5.49	−10.12	−0.87	0.02377
	NSGA-III	Next-fit	−22.34	−26.97	−17.72	2.10E − 07
	PSO	NSGA-III	16.53	11.90	21.15	5.00E − 06