Newsgroups: comp.ai.genetic
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!fas-news.harvard.edu!newspump.wustl.edu!news.ecn.bgu.edu!psuvax1!uwm.edu!cs.utexas.edu!swrinde!ihnp4.ucsd.edu!newshub.nosc.mil!news!news
From: "R. Scott Starsman" <r_starsman@nise-p.nosc.mil>
Subject: Q: Coding a GA for a TSP problem
Message-ID: <1995Mar10.193743.10655@nosc.mil>
Sender: news@nosc.mil
Organization: NCCOSC RDT&E Division, San Diego, CA
Date: Fri, 10 Mar 1995 19:37:43 GMT
Lines: 19

I'm writing some GA code to solve the TSP problem.  

I'm running into the problem of multiple visits to nodes
during evolution.  How do you solve this problem?

I have tried both heavy fitness penalties for multiple visits
and I have also devised a data structure that allows only a
single visit per given number of nodes.  Unfortunately, neither
of these methods seem to acheive a solution nearly as good as
one I can guess.  

I'm using 30 nodes as a test case (perhaps this is too many).

Can anyone provide some insight?

Thanks,

Scott
r_starsman@nise-p.nosc.mil
