Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!think.com!mintaka!bloom-beacon!eru!hagbard!sunic!dkuug!diku!skipper
From: skipper@diku.dk (Jakob Skipper)
Newsgroups: comp.ai.neural-nets
Subject: Re: Core Wars
Keywords: core wars, genetic algorithms, ga, mule dna
Message-ID: <1991May30.170107.26863@odin.diku.dk>
Date: 30 May 91 17:01:07 GMT
References: <1991May29.001505.12073@leland.Stanford.EDU> <1991May30.115136.24223@odin.diku.dk>
Sender: skipper@freja.diku.dk
Organization: Department of Computer Science, U of Copenhagen
Lines: 32

skipper@diku.dk (Jakob Skipper) writes:

>reaper@leland.Stanford.EDU (Dan Michaeloff) writes:


>>   I am using a GA to evolve a Core Wars fighter program, using tournaments
>>against past contest winners to measure goodness.  However, one of the '88
>>finalists, "Mule DNA" by Douglas McDaniels of the US, looks like it may have
>>been created by a GA, as it contains many seemingly arbitrary constants.
>>However, it seems to consist of repeated segments.

>>   Does anyone know more on this?  Has anyone applied GAs to Core Wars?

>A friend of mine once did it, with very little success. There are, however,
  .. stuff deleted ..

OOPS, when i wrote that answer, I completely forgot the reference that
is perhaps closest to what you are looking for. Four people from DTH (The
technical University of Denmark) Steen Rasmussen, Carsten Knudsen,
Rasmus Feldberg and Morten Feldberg has written the report:

  The Coreworld: Emergence and evolution of cooperative structures in a
  computational chemistry

You find it in Physica D 42, 1990, pp. 111-134. It is a description of
a system called VENUS that tries to make programs by evolution. The system
uses the Redcode language from the Core War game.

Kind regards and good luck,

Jakob Skipper
Institute of Datalogy, University of Copenhagen