Information Technology Reference
In-Depth Information
Table 4.17
Parameters for the
n
-exactly-one-on problem with UDFs.
3-1On
4-1On
5-1On
6-1On
Number of runs
100
100
100
100
Number of generations
1,000
10,000
10,000
10,000
Population size
30
30
30
30
Number of fitness cases
8
16
32
64
Function set
A O N
A O N
A O N
A O N
User defined functions
2-1On
3-1On
4-1On
5-1On
Terminal set
a b c
a b c d
a b c d e
a b c d e f
Head length
10
10
10
10
Gene length
21
21
21
21
Number of genes
2
2
2
2
Linking function
A
A
A
A
Chromosome length
42
42
42
42
Mutation rate
0.044
0.044
0.044
0.044
Inversion rate
0.1
0.1
0.1
0.1
IS transposition rate
0.1
0.1
0.1
0.1
RIS transposition rate
0.1
0.1
0.1
0.1
One-point recombination rate
0.3
0.3
0.3
0.3
Two-point recombination rate
0.3
0.3
0.3
0.3
Gene recombination rate
0.3
0.3
0.3
0.3
Gene transposition rate
0.1
0.1
0.1
0.1
Fitness function
Eq. (3.14)
Eq. (3.14)
Eq. (3.14)
Eq. (3.14)
Success rate
98%
83%
85%
53%
means that there is indeed a pattern in the structure of these functions. Let's
then see if we can find it.
Compared to the odd-parity functions, the structural pattern governing the
design of the exactly-one-on functions is much harder to find, and being able
to find the most parsimonious representations is therefore crucial in order to
discern this pattern. For this task, we are going to use the already familiar
strategy of letting the system evolve freely without constraints concerning
the size of the evolving solutions, but as soon as a perfect solution is found,
parsimony pressure is applied in order to create perfect solutions that are as
compact as possible. Thus, before applying the parsimony pressure the fit-
ness is evaluated by equation (3.14), whereas during the simplification proc-
ess the fitness is evaluated by equation (4.21), where
rf
i
corresponds obvi-
ously to the sensitivity/specificity raw fitness.
