Information Technology Reference
In-Depth Information
Table 5.3
Performance and settings used in the polynomial function problem
without numerical constants (
GEP
) and with a fixed set of NCs (
GEP-NC
).
GEP
GEP-NC
Number of runs
100
100
Number of generations
1000
1000
Population size
50
50
Number of fitness cases
10 (Table 5.2)
10 (Table 5.2)
Function set
+ - * /
+ - * /
Terminal set
x
x a b c d e
Head length
6
6
Gene length
13
13
Number of genes
2
2
Linking function
+
+
Chromosome length
26
26
Mutation rate
0.044
0.044
Inversion rate
0.1
0.1
IS transposition rate
0.1
0.1
RIS transposition rate
0.1
0.1
One-point recombination rate
0.3
0.3
Two-point recombination rate
0.3
0.3
Gene recombination rate
0.3
0.3
Gene transposition rate
0.1
0.1
Fitness function
Equation (3.3b)
Equation (3.3b)
Selection range
100%
100%
Precision
0.0%
0.0%
Average best-of-run fitness
934.534
987.384
Average best-of-run R-square
0.9999190824
0.9999802294
and a fitness of 983.813. Its chromosome is shown below (the sub-ETs are
linked by addition):
01234567890120123456789012
+x**x+xabbcae**+x+*xabbcae
(5.3a)
And as you can see in Figure 5.2, it codes for the following function:
2
y
2
762
x
3
19258
x
(5.3b)
which is a pretty good approximation to the target function (5.1).
It is worth pointing out, however, that in real-world problems one never
knows either the type or the range of the numerical constants that are needed
