Newsgroups: comp.ai.genetic
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!yale!gumby!newsxfer.itd.umich.edu!zip.eecs.umich.edu!panix!news.mathworks.com!news.alpha.net!uwm.edu!spool.mu.edu!howland.reston.ans.net!pipex!sunic!news.chalmers.se!news.gu.se!gd-news!d6241.shv.hb.se!sa209
From: sa209@utb.shv.hb.se (Claes Andersson)
Subject: Re: Optima... when?
Message-ID: <sa209.60@utb.shv.hb.se>
Sender: usenet@gdunix.gd.chalmers.se (USENET News System)
Nntp-Posting-Host: d6241.shv.hb.se
Organization: Department of Scocial Science
References: <sa209.54@utb.shv.hb.se> <39ltr6$det@lawelawe-f0.mrtc.maui.com>
Date: Wed, 9 Nov 1994 15:18:56 GMT
Lines: 22

In article <39ltr6$det@lawelawe-f0.mrtc.maui.com> Paul Billings <pab@maui.com> writes:


>In article <sa209.54@utb.shv.hb.se>, <sa209@utb.shv.hb.se> writes:
>>  I have a question.. When the genomes converge depends much on the size of 
>> the population but can one always say that one have reached a local optima 
>> when the genomes have converged?

>No.  (Consider the case when population size=1.)

 That's of course correct but with a case where one optimizes strategies for 
robots trying to get from point A to point B in a random course, is there 
really any way to know for sure when a local optima is reached? Isn't the 
only way to do it to assume (if the population originally was reasonably 
large) that there is some sort of local optima when they have converged, or 
at least that it's not efficient to go further.. When optimizing an 
algorithm one can test it in some way, but an intricate stragtegy that is 
generally optimized for guiding a vehicle through a course with random 
obstacles ought to be impossible to test.. Am I wrong? I would be glad if I 
were.

Claes Andersson. University of Bors. Sweden
