Information Technology Reference
In-Depth Information
a.
01234560123456
A bbabcAOc abc-[m] = 6
A bbabcAOc abc-[d] = 8
N
a
a
b
b.
Sub-ET
1
Sub-ET
2
A
A
c
N
b
O
b
a
a
c.
Sub-ET
1
Sub-ET
2
A
A
a
c
b
O
a
b
Figure 3.13.
Illustration of point mutation.
a)
The mother and daughter chromo-
somes with the mutation points in bold.
b)
The sub-ETs encoded in the mother
chromosome (before mutation).
c)
The sub-ETs codified by the daughter chromo-
some (after mutation). The mutated nodes are shown in gray. Note that the first
mutation changed significantly the sub-ET
1
, shortening the original sub-ET in one
node. Note also that with these two point mutations, a perfect solution to the
Majority(
a
,
b
,
c
) function was discovered (the sub-ETs are linked by OR).
On the other hand, we can see that, in gene expression programming, mu-
tations in the coding sequence of a gene have most of the times a very pro-
found effect, reshaping drastically the sub-ETs . However, this capability to
reshape profoundly expression trees is fundamental for an efficient evolu-
tion (see chapter 12 for a discussion of the Genetic Operators and Their
Power). Indeed, the results presented throughout this topic clearly show that
our too human wish to keep intact the small functional building blocks and
recombine them carefully without disrupting them (as is done in genetic pro-
gramming through tree crossover) is conservative and works poorly in evo-
lutionary terms. In fact, genotype/phenotype systems can find much more
efficient ways of creating their own building blocks and manipulating them
