Information Technology Reference
In-Depth Information
Ta b l e 5 . 9 .
Comparison for Optimal Instances of DE and RKGA Implementations
DE
RKGA
Instance
F
avg
Δ
avg
F
avg
Δ
avg
11EIL51
0
0.08
0
0.2
14ST70
0
0.1
0
0.2
16EIL76
0
0.12
0
0.2
16PR76
0
0.14
0
0.4
20KROA100
0
0.21
0
0.4
20KROB100
0
0.22
0
0.3
20KROC100
0
0.2
0
0.4
20KROD100
0
0.21
0
0.6
20KROE100
0
0.2
0
0.5
20RAT99
0
0.2
0
0.5
20RD100
0
0.2
0
0.4
21EIL101
0
0.19
0
0.5
21LIN105
0
0.21
0
0.4
22PR107
0
0.23
0
0.8
25PR124
0
0.28
0
0.4
26BIER127
0
0.33
0
0.5
28PR136
0
1.27
0
1
29PR144
0
0.37
0
0.7
30KROA150
0
0.48
0
0.9
30KROB150
0
0.46
0
1.2
31PR152
0.01
1.49
0
0.8
32U159
0
0.55
0
1
39RAT195
0.07
4.6
0
1.6
40D198
0.04
3.54
0
1.8
40KROA200
0
1.81
0
1.9
40KROB200
0.04
2.03
0
2.1
45TS225
0.25
2.98
0.02
1.5
46PR226
0
0.76
0
1.9
53GIL262
1.24
5.65
0.75
2.1
53PR264
0.01
4.38
0
3.2
60PR299
0.71
10.4
0.11
3.5
64LIN318
0.77
8.89
0.62
5.9
80RD400
1.64
18.89
1.19
5.3
84FL417
0.09
25.26
0.05
9.5
88PR439
1.13
22.94
0.27
9
89PCB442
1.78
12.12
1.7
1.72
Avg
0.22
3.67
0.13
0.2
where
H
i
,
OPT
and R are the objective function values generated by the DE in each
run, the optimal objective function value, and the number of runs, respectively. For the
computational effort consideration,
t
avg
denotes average CPU time in seconds to reach
the best solution found so far during the run, i.e., the point of time that the best so
Search WWH ::
Custom Search