Information Technology Reference
In-Depth Information
to choose the appropriate settings for a problem, including the fitness func-
tion, the number of genes, the head length, the function set and the linking
function. This kind of analysis is useful for developing an intuitive under-
standing of the fundamental parameters of the algorithm. The parameters
chosen for this problem are summarized in Table 4.1; how and why they
were chosen is discussed below.
Consider we are given a sampling of numerical values from the test func-
tion (4.1) over 10 random points chosen from the interval [-10, 10] (Table
4.2), and we wanted to find a function fitting those values within 0.01 of the
correct value. We could, therefore, evaluate the fitness by equation (3.3a) to
make sure that all the solutions with maximum fitness match indeed the tar-
get function.
Table 4.1
Settings for the polynomial function problem.
Number of runs
100
Number of generations
50
Population size
30
Number of fitness cases
10 (Table 4.2)
Function set
+ - * /
Terminal set
a
Head length
6
Gene length
13
Number of genes
4
Linking function
+
Chromosome length
52
Mutation rate
0.0385
Inversion rate
0.1
One-point recombination rate
0.3
Two-point recombination rate
0.3
Gene recombination rate
0.3
IS transposition rate
0.1
RIS transposition rate
0.1
Gene transposition rate
0.1
Fitness function
Equation (3.3a)
Selection range
100
Precision
0.01
Success rate
100%
