Newsgroups: comp.ai.neural-nets
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!howland.reston.ans.net!usc!crash!churritz
From: churritz@crash.cts.com ()Skito()
Subject: Re: Help! ftp site for self-growing traveling salesman (Flexmap)
Organization: establishment met nine
Date: Tue, 13 Jun 1995 22:21:14 GMT
Message-ID: <DA4u3F.Dvu@crash.cts.com>
References: <3rk49r$fmh@news.ox.ac.uk>
Sender: news@crash.cts.com (news subsystem)
Nntp-Posting-Host: crash.cts.com
Lines: 27

In article <3rk49r$fmh@news.ox.ac.uk>,
Michael Zhu Ning <c94nz@PROBLEM_WITH_INEWS_DOMAIN_FILE> wrote:
>Hi!
>
>Last year, some people mentioned a kind of so-called "self-growing cell"
>structure neural-nets on this group for solving "traveling salesman"
>problem. Could anybody tell where I can find the paper?
>
>many thanks!

You may be refering to the Linear order Flexmap Algorithm, who's co-author is

fritzke@immd2.informatik.uni-erlangen.de
Universit:at Erlangen-N:urnberg
Bernd Fritzke

In my opinion the algorithm doesn't do as well as the paper suggests, in
speed and in goodness of tour.  The algorithm is linear, but the linear
constants must be pushed to the 100's to get good solutions.

-skito

--
                                                               here i am
                                                               here i am
                                                               waiting to hold
churritz@cts.com                           This Mortal Coil -- you
