Generalised Adaptive Harmony Search: A Comparative Analysis of Modern Harmony Search
Table 2
Average scores and standard deviations over 50 runs for 30-dimensional problems. All optimisers were allowed to run for 100,000 function evaluations, except when the Goldstein-Price polynomial was being minimised. In that case, 10,000 evaluations were allowed. The best results from each benchmark function are indicated in bold.
Ackley
Griewank
Rosenbrock
Rastrigin
Goldstein-Price
HS
Best
0.0066
0.0322
30.7287
0.0142
6.8464
Standard deviation
0.0006
0.0334
22.6459
0.0027
9.3353
Hits
50
48
1
2
42
DLHS
Best
0.8464
0.0170
32.7372
0.2003
6.2444
Standard deviation
0.5173
0.0180
22.8117
0.4200
8.7724
Hits
2
50
0
31
44
EHS
Best
0.0008
0.0006
28.9183
15.4119
4.0809
Standard deviation
0.0024
0.0022
10.7493
10.8267
5.2910
Hits
50
50
0
5
48
AHS
Best
0.0003
0.0042
26.5041
1.4819
3
Standard deviation
0.0020
0.0111
0.5624
0.9055
0
Hits
50
50
0
1
50
Goldstein-Price was implemented in two dimensions as it is only defined in two.