Evolutionary computation
In the year 1975 John Holland proposed an algorithm that should imitate the process of
evolution. This algorithm called nowadays simple genetic algorithm (SGA)
looks as follows [1]:
Simple genetic algorithm
First, the primary population of individuals is initialised, then using the fitness function
(function that evaluates the population of individuals) the next population is generated.
After that, the individuals are reproduced, using cross-over algorithm, then mutation of
genotypes may occur. In the end, the population of individuals is ready to be submitted to
the process of selection again.
Main processes occuring in this algorithm are:
- Selection - is the process of choosing appropriate individuals (from the aspect of given task)
from the population of ancestors.
- Cross-over - is generation of new individuals basing on genotypes of parents in the process of
reproduction.
- Mutation - is the process of applying slight changes to the genotypes in the population.
Bibliography
[1] J. Arabas, Wyklady z algorytmow ewolucyjnych, Wydawnictwa Naukowo-Techniczne, Warszawa 2001.
|
