Information Technology Reference
In-Depth Information
For this problem, two genes per chromosome were used, each with a head
length of 12 and, therefore, encoding sub-ETs with a maximum of 25 nodes
each. The sub-ETs were posttranslationally linked by addition. Both the per-
formance and the parameters used per run are summarized in Table 4.3.
As you can see in Table 4.3, small populations of just 50 individuals were
used, but they were left to evolve for 5000 generations, enough time to dis-
cover good solutions, as the average best-of-run R-square of 0.99043015
indicates.
The best-of-experiment solution has a fitness of 999.992 and an R-square
of 0.99998525, and was discovered in generation 4946 of run 31:
01234567890123456789012340123456789012345678901234
*S*aCCQbSce/bccbabbadceeeT-deb/-EceTebeccbbebedcad
(4.7a)
which, mathematically, corresponds to the following expression:
y
sin(a)
cos(
b
)
cos
sin
(
c
)
tan(
d
e
)
(4.7b)
Table 4.3
Settings for the five-parameter function problem.
Number of runs
100
Number of generations
5000
Population size
50
Number of fitness cases
100
Function set
+ - * / Q E S C T
Terminal set
a b c d e
Head length
12
Gene length
25
Number of genes
2
Linking function
+
Chromosome length
50
Mutation rate
0.044
Inversion rate
0.1
IS transposition rate
0.1
Gene transposition rate
0.1
One-point recombination rate
0.3
Two-point recombination rate
0.3
Gene recombination rate
0.3
RIS transposition rate
0.1
Fitness function
Equation (3.4a)
Average best-of-run fitness
997.661
Average best-of-run R-square
0.99043015
