Information Technology Reference
In-Depth Information
chosen to make the comparisons must follow different evolutionary dynamics
so that the results discussed here could be useful not only theoretically but
also for understanding the evolutionary strategies chosen by different artificial
evolutionary systems.
For this analysis, we are going to use the already familiar test sequence
(5.14) of section 5.6.1. In all the experiments, the first 10 positive integers
n
and their corresponding term were used as fitness cases (see Table 5.5); the
fitness function was evaluated by equation (3.3b) and a selection range of
20% and maximum precision (0% error) were chosen, giving maximum fit-
ness
f
max
= 200; population sizes
P
of 50 individuals and evolutionary times
G
= 100 generations were used; the success rate of each experiment was
evaluated over 100 independent runs; F = {+, -, *, /} and the terminal set
T
consisted only of the independent variable; and six-genic chromosomes of
length 78 (head length
h
= 6) linked by addition were used. The parameters
of all the five experiments are summarized in Table 12.2.
We have already seen that point mutation is by far the single most important
genetic operator and that populations undergoing mutation display
Table 12.2
Success rates and parameters for a non-homogenizing system undergoing muta-
tion (
Mut
) and homogenizing systems undergoing two-point recombination
(
Rec2P
), one-point recombination (
Rec1P
), gene recombination (
RecG
), and three
different kinds of recombination (
RecMix
).
Mut
Rec2P
Rec1P
RecG
RecMix
Number of runs
100
100
100
100
100
Number of generations
100
100
100
100
100
Population size
50
50
50
50
50
Number of fitness cases
10
10
10
10
10
Head length
6
6
6
6
6
Number of genes
6
6
6
6
6
Chromosome length
78
78
78
78
78
Mutation rate
0.05
--
--
--
--
Two-point recombination rate
--
1.0
--
--
0.8
One-point recombination rate
--
--
1.0
--
0.8
Gene recombination rate
--
--
--
1.0
0.8
Selection range
20%
20%
20%
20%
20%
Precision
0%0%0%0%0%
Success rate
96%
0.04%
0.03%
0.0%
0.13%
