Newsgroups: comp.ai.genetic
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!news.mathworks.com!uunet!in1.uu.net!pipex!warwick!bsmail!ian
From: ian@bris.ac.uk (Ian Craddock)
Subject: Re: Question: On changing genome size
Message-ID: <D63FG9.2rv@info.bris.ac.uk>
Sender: usenet@info.bris.ac.uk (Usenet news owner)
Nntp-Posting-Host: kryten.ccr.bris.ac.uk
Organization: University of Bristol, England
X-Newsreader: TIN [version 1.2 PL2]
References: <mjg12.1.001849AE@po.cwru.edu>
Date: Mon, 27 Mar 1995 09:52:55 GMT
Lines: 32

Michael Golebiewski (mjg12@po.cwru.edu) wrote:
>      I am doing a project and I would like to apply genetic algorithms, but 
> one part of the problem is that the size of the solution may expand or 
> contract from that of the initial population. I need some background on 
> changing the genome soze from individual to individual

> Mike Golebiewski
> Case Western Reserve University

I've been having similar thoughts and am quite interested in this idea;
so far what I've considered is that (along with many other features of
GA's) the change in genome size should be modelled on what happens in
nature. As far as I understand it the number of genes may increase
by a section of the DNA being duplicated (ie you get two copies of 
one or more genes - obviously this is a good place for new genes
to start as the section being duplicated is presumably already 'useful'
so small changes to the duplicated section via mutation etc may
well yield similar 'useful' genes).

I'm not aware of anyone having tried this and this isnt my main research
interest so I havent had time to give it a go. One problem that I can
see is that in many cases the genome _may_ just keep on growing! This
increase in length might be limited by the increasing chance of mutation
adversely affecting the genome but if so the mutation rate might become
a critical parameter - which is undesirable.

Any other thoughts?

Ian.
ian.craddock@bristol.ac.uk


