Newsgroups: comp.ai.genetic
Path: cantaloupe.srv.cs.cmu.edu!rochester!cornellcs!newsstand.cit.cornell.edu!news.acsu.buffalo.edu!news.uoregon.edu!ddsw1!news.mcs.net!newspump.sol.net!howland.erols.net!news-peer.gsl.net!news.gsl.net!EU.net!usenet2.news.uk.psi.net!uknet!usenet1.news.uk.psi.net!uknet!uknet!lyra.csx.cam.ac.uk!hgmp.mrc.ac.uk!ebi.ac.uk!news
From: Cedric Notredame <Cedric.Notredame@ebi.ac.uk>
Subject: parrallel GA
Sender: news@ebi.ac.uk (usenet news)
Message-ID: <32B150EB.41C6@ebi.ac.uk>
Date: Fri, 13 Dec 1996 12:49:47 GMT
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=us-ascii
Mime-Version: 1.0
X-Mailer: Mozilla 2.02 (X11; I; OSF1 V3.0 alpha)
Organization: EBI - European Bioinformatics Institute
Lines: 31

Hi,

I am currently using a parrallel GA that I designed myself. It involves
several GA running independently on several machines. From time to time
( e.g. every 5 generation) some individuals are exchanged, according to
some simple rules ( best individuals are moving to the next GA).
In practice all the  GAs are aranged on a k-branched tree.  Individuals
travel in a single direction ( from the leaves to the root).
Any GA that is not a leaf receives individuals from the k GAs that are
'above' it. 

      O    O   O
       |   |   |
        |  |  |
         | | | 	
           O
           |
          ...
	
In practice, this does very well. And after a while, the best solutions
are found in the root. I would be surprised if this model had not been
described somewhere. Does anybody know if the properties of such models
have been investigated, and is there some bib on line available? 
Thanks in advance, 
cedric 
-- 
 Cedric NOTREDAME                   E-mail:
cedric.notredame@ebi.ac.uk    
 EMBL Outstation - The EBI          URL:       http://www.ebi.ac.uk  
 Hinxton Hall, Hinxton,             Telephone: +44 (0)1223 494 416
 Cambridge CB10 1RQ U.K.            Fax:       +44 (0)1223 494 468
