Dynamic Reduction-Expansion Operator to Improve Performance of Genetic Algorithms for the Traveling Salesman Problem

Table 4

Results on set of 14 TSP instances: comparison with KMC [2] (same average execution time).

			REDEXP							KMC
Instance	Size	Optimal Value	RedExp	Error	_RedExp	Error	_RedExp	Error	Best Error	Average Time (s)	Best Error
berlin52	52	7542	7903.8	4.7967	7985.8	5.8841	7841.4	3.9696	3.9696	16.3208	0.0000
kroA100	100	21282	22445.1	5.4649	22445.1	5.4649	22381.9	5.1683	5.1683	20.3318	0.1035
pr144	144	58537	61399.2	4.8896	61399.2	4.8896	61243.8	4.6241	4.6241	26.5335	2.0176
ch150	150	6528	6707.4	2.7477	6784.3	3.9264	6798.6	4.1450	2.7477	25.4257	5.1038
kroB150	150	26130	27439.0	5.0095	26787.4	2.5158	27084.9	3.6543	2.5158	25.5628	3.3544
pr152	152	73682	77133.3	4.6841	77082.1	4.6145	75852.8	2.9461	2.9461	26.4204	2.5998
rat195	195	2323	2419.8	4.1675	2411.1	3.7936	2411.8	3.8231	3.7936	30.0717	8.7133
d198	198	15780	16253.2	2.9985	16270.4	3.1075	16504.8	4.5930	2.9985	32.5780	2.8149
kroA200	200	29368	30026.6	2.2427	30290.5	3.1413	30303.1	3.1840	2.2427	31.8691	5.3725
ts225	225	126643	138526.9	9.3837	139043.0	9.7913	130054.6	2.6938	2.6938	34.1776	10.3008
pr226	226	80369	82685.6	2.8825	82113.5	2.1706	83026.1	3.3061	2.1706	36.3008	4.9741
pr299	299	48191	51740.7	7.3658	53164.5	10.3205	51731.9	7.3476	7.3476	44.6964	14.6584
lin318	318	42029	46135.6	9.7710	46507.3	10.6552	45821.8	9.0243	9.0243	47.2462	15.0669
pcb442	442	50778	57006.4	12.2660	57917.6	14.0604	57032.5	12.3173	12.2660	62.8187	22.5887
	204		Average =	5.6193		6.0240		5.0569	4.6078	32.8824	6.9763