Information Technology Reference
In-Depth Information
a.
01234567890120123456789012
/x-/-+xxxxxxx*++x/+xxxxxxx
b.
Sub-ET
1
Sub-ET
2
x
x
x
x
x
x
x
x
x
x
x
x
x
§
·
y
(
x
2
3
x
)
3
x
2
3
x
c.
©
¹
2
Figure 5.1.
Best solution created in the experiment summarized in the first column
of Table 5.3. This program has an R-square of 0.999994826 and was found in
generation 395 of run 2.
a)
The chromosome of the individual.
b)
The sub-ETs
codified by each gene.
c)
The corresponding mathematical expression after linking
with addition (the contribution of each sub-ET is shown in brackets). Note that this
model is a rather simplistic approximation to the target function (5.1).
to solve a problem, and choosing the right ones for a problem might be really
complicated. So, a more flexible approach is required to handle a wider di-
versity of random numerical constants, and an elegant and efficient way of
doing so is presented below.
5.2 Genes with Multiple Domains to Encode RNCs
A facility for handling random numerical constants can be easily implemented
in gene expression programming. We have already met two different do-
mains in GEP genes: the head and the tail. And now another one - the Dc
domain - will be introduced. This domain was especially designed to handle
