Information Technology Reference
In-Depth Information
deterministic selection is more problematic as for more complex problems the
plots for both schemes become more intertwined (Figure 12.16).
However, there are several reasons why one should choose the roulette-
wheel selection. (1) Real-world problems are more complex than the prob-
lems analyzed here. (2) The ideal exclusion factor depends on population
size and the complexity of the problem. (3) Deterministic selection requires
more CPU time as individuals must be sorted by rank and, for large
populations, this is a factor to take seriously into consideration. (4) Deter-
ministic selection is not appropriate in systems undergoing recombination
alone as it reduces dramatically the genetic diversity of the population. And
(5), roulette-wheel selection is easy to implement and mimics nature more
faithfully and therefore is much more appealing.
This topic finishes here and I hope to have shown you that gene expres-
sion programming is not only easy to implement but also easy to understand.
Although natural evolution seems sometimes something beyond our grasp,
the simple artificial system of gene expression programming can be minutely
dissected in order to reveal all its secrets.
And, most of all, I hope to have convinced you that gene expression pro-
gramming can help you find very good solutions to difficult real-world prob-
lems, not easily or satisfactorily solved by conventional mathematical or
statistical methods.
