Information Technology Reference
In-Depth Information
OHOWbabababbaaba-[11] = 9
OOOTOabbababbaab-[12] = 8
OTHTWabbbabbbabb-[13] = 4
OHbWTabaababbaab-[14] = 8
TOWbWababbaabbaa-[15] = 7
ObOTTabbababbaab-[16] = 10
OTWOOabbaabbbabb-[17] = 7
OTHbTababbaabaab-[18] = 7
HOWbbaaaabbbbaba-[19] = 10
Finally, in the next generation an individual with maximum fitness (chro-
mosome 2), solving all the 14 instances of Table 9.1, was created. Its expres-
sion is shown in Figure 9.7. The complete population is shown below:
Generation N: 5
0123456789012345
HOWbaaaababbbaba-[ 0] = 12
HOabaaaabbbababa-[ 1] = 12
OHOWbababbabbaba-[ 2] = 14
OaWOTababbaabaaa-[ 3] = 8
HOWbbaaaaabbaaba-[ 4] = 10
TOabbabbbbabbbaa-[ 5] = 9
TOabbabbbbabbaaa-[ 6] = 9
OOOTTabbababbaab-[ 7] = 7
OOOTOabbababaaaa-[ 8] = 8
HOaTbabbababbaab-[ 9] = 10
OOWbaaaabaabbaba-[10] = 6
OTOTOabbababbaab-[11] = 7
TOabWabbaaabbaab-[12] = 8
TbaaOabaabbbbabb-[13] = 9
TabbOabbbaaabbaa-[14] = 5
OaOOTabbababbaab-[15] = 8
OOWbaaaabbabbaba-[16] = 6
TOaOaabaabaabbaa-[17] = 9
TaOOTabaabaabbaa-[18] = 8
WbWaWaaabbabbaba-[19] = 9
Note that the perfect solution created in this generation is not a direct
descendant of the best of the previous generation; indeed, it was most prob-
ably created by an event of two-point recombination between chromosomes
3 and 11, both of which low profile individuals (chromosome 3 has fitness 6
and chromosome 11 is slightly better with fitness 9).
You probably noticed that all the chromosomes presented here encode
decision trees with an attribute node in the root. Indeed, creating decision
trees with a terminal at the root would be a total waste of resources and,
