Jump to content

Talk:Mutation (genetic algorithm)

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

meat

[edit]

now that this has some meat to it, should we consider de-stubifying it. There is much more that we can do with it, but I would question whether it still exists as a stub or not. Rgarvage 16:09, Aug 21, 2004 (UTC)


I don't think this is a stub either. Unless someone has plans on expanding it in the near future, it should be removed from stubhood. --Fastfinge 22:50, Sep 19, 2004 (UTC)


An information is not given in the article

[edit]

It is not mentioned in the article where the mutation operation is used in the GA. There are N chromosomes in the last step. ( I say them old N chromosomes) There are three choises, I think.

1) N new chromosomes are generated by three operations: a) Some are directly copied from initial population b) Some are generated by crossover c) Some are generated by mutation (This is the algoritm used in MATLAB's implementation)

2) N old chromosomes enter crossover and after mating and crossover N new chromosomes are generated. There are 2 choises for this case.

  2.a) N new chromosomes enter mutation.
  2.b) Both N old and N new chromosomes enter mutation.

Which one above (1, 2.a and 2.b) is true? OR Are all of them true? Kavas (talk) 08:17, 17 September 2009 (UTC)[reply]

local minima

[edit]

local maxima or minima are denote in a mathematics function. The function presented the continuous relation of x-y direction. Mutation in GA fitnesses are developed random or variance interval and not in sequence steps. So local maxima or minima of mathematic analysis is not suitable term to describe fitness evolution (will stop). — Preceding unsigned comment added by 202.170.57.247 (talk) 04:50, 27 April 2019 (UTC)[reply]

Thorough revision of the old article

[edit]

The old article was greatly expanded, since mutations occur not only in genetic algorithms but in almost all other evolutionary algorithms (EA). In this respect, the title is also rather too restrictive. The design of mutation operators depends not only on the EA type but also very much on the representation of the decision variables of the chromosome (genetic representation). Some examples were given and, above all, references were made to the literature. Studi90 (talk) 17:00, 4 January 2023 (UTC)[reply]