Research Article

A High-Performance Genetic Algorithm: Using Traveling Salesman Problem as a Case

Table 3

Simulation results of GA, HeSEA, LEM, and SA.

Data set PREGA PREHeSEAPRELEMPRESA

a280 ( )−0.71 (1.68%) −91.840.22 (1.47%)−94.791.17 (1.90%)−95.92−0.07 (1.48%)−57.62
2,679.442,681.442,671.602,743.77
u574 ( )−5.51 (0.84%) −94.76−5.88 (0.65%)−93.141.25 (0.83%)−95.920.02 (1.17%)−52.54
38,837.7039,211.4039,156.4039,448.30
u724 ( )−5.12 (2.39%) −94.61−3.95 (0.77%)−93.751.04 (0.89%)−96.14−0.27 (0.95%)−52.68
44,399.1044,347.7044,388.7045,044.90
u1060 ( )−3.69 (4.31%) −94.55−5.52 (2.28%)−93.970.85 (0.77%)−96.25−0.29 (1.01%)−49.33
241,758.00238,428.00237,896.00238,928.00
u1432 ( )−2.76 (4.20%) −93.76−0.45 (3.80%)−93.781.27 (0.80%)−95.970.13 (0.75%)−53.26
164,093.00165,570.00162,950.00162,973.00
pr2392 ( )−3.32 (3.34%) −95.06−5.58 (2.43%)−95.890.14 (0.46%)−96.61−0.25 (0.58%)−53.20
405,470.00405,612.00407,597.00419,796.00
pcb3038 ( )−0.55 (4.64%) −94.72−0.97 (4.22%)−95.340.40 (0.66%)−95.73−0.05 (0.73%)−59.18
148,567.00148,293.00148,353.00152,449.00
fnl4461 ( )−4.09 (4.52%) −95.32−1.02 (3.06%)−95.580.00 (0.51%)−95.190.12 (0.36%)−61.21
195,074.00195,938.00195,907.00202,525.00
usa13509 ( )−0.77 (5.23%) −92.186.09 (0.42%)−95.47−0.17 (0.28%)−94.89−1.41 (0.46%)−85.03
21,500,000.0023,700,000.0021,900,000.0022,600,000.00

Average −2.95−94.09−1.90−94.630.66−95.85−0.23−58.23

: time in seconds; : best solution in 30 runs; : coefficient of variation as defined in Table 2.