The Backpropagation Artificial Neural Network Based on Elite Particle Swam Optimization Algorithm for Stochastic Linear Bilevel Programming Problem

Table 4

The comparison results of Example 2 from 11 runs.

Algorithm	means constrain	optimal solutions	means		variance		CPU time
			UL	LL	UL	LL	(s)
BPANN-PSO	Y	(0.0529, 0.2509, 0,0.0072, 0.9501; 0.5987,1.2321,1.1399)	9.9769	-2.2822	5.9689	77.7253	4.3898
Method in [51]	Y	(0.0530,0.2510, 0, 0.0070, 0.9500; 0.5990,1.2320, 1.1400)	9.9770	-2.2820	5.9700	77.7250	21.2172
BPANN-PSO	N	(0.4101,0.6320,0.0671,0.2702,0.8012; 0.2757,0.9731, 0.7488)	9.5979	-3.3727	20.6536	48.3552	0.9654
Method in [51]	N	(0.4100,0.6320,0.0670,0.2700, 0.8010,0.2760,0.9730,0.7490)	9.5980	-3.3720,	20.6540	48.3550	7.1836