Newsgroups: comp.ai.genetic
Path: cantaloupe.srv.cs.cmu.edu!bb3.andrew.cmu.edu!newsfeed.pitt.edu!gatech!swrinde!newsfeed.internetmci.com!portal.gmu.edu!hearst.acc.Virginia.EDU!murdoch!design.mech.Virginia.EDU!sec3a
From: sec3a@design.mech.Virginia.EDU (Susan E. Carlson)
Subject: mutation
X-Nntp-Posting-Host: design.mech.virginia.edu
Message-ID: <DMD7FG.Ly3@murdoch.acc.Virginia.EDU>
Followup-To: Interesting idea... 
Sender: usenet@murdoch.acc.Virginia.EDU
Organization: University of Virginia
Date: Tue, 6 Feb 1996 17:33:15 GMT
Lines: 10

Omard,

I use the idea of changing mutation and crossover rate based on the 
the variability of the population.  If the population is very diverse
then I use a low mutation rate and a high crossover rate.  If the
population has niched and has many members that are the same genetically,
then I raise the mutation rate and decrease crossover (crossover won't
do anything for you anyway if the parents are the same genetically.)

susan
