Newsgroups: comp.ai.alife
Path: cantaloupe.srv.cs.cmu.edu!rochester!udel!news.mathworks.com!news.kei.com!simtel!harbinger.cc.monash.edu.au!yoyo.aarnet.edu.au!spasun.tpa.com.au!dmsaffron.adl.dms.CSIRO.AU!dmssyd.syd.dms.CSIRO.AU!its.csiro.au!news
From: Shaun Green <shaun@prospect.anprod.csiro.au>
Subject: Re: Cellular Universes
Content-Type: text/plain; charset=us-ascii
Message-ID: <D9HGML.B7A@its.csiro.au>
Sender: news@its.csiro.au (News Manager)
Nntp-Posting-Host: green.prospect.anprod.csiro.au
Content-Transfer-Encoding: 7bit
Organization: DSL Systems Centre, CSIRO
References: <3qf086$r87@status.gen.nz> <Pine.ULT.3.91.950531013159.25022B-100000@rac10.wam.umd.edu>
Mime-Version: 1.0
Date: Thu, 1 Jun 1995 07:25:32 GMT
X-Mailer: Mozilla 1.1N (Windows; I; 16bit)
Lines: 36

Keith Wiley <keithw@wam.umd.edu> wrote:
>> I have no formal training in A.I, but have been developing my own little 
>> programs.  Initially I copied programs that itterate on a frame by frame 
>> basis, meaning each cell of the 'universe' is processed each frame.  A 
>> new array is generated from the old, then new becomes old, and the 
>> process repeated.  Two problems with this approach occur.
>> 
>> First you have an inbuilt bias towards or against the cells processed 
>> first.  Second, infinite static loops develop within the cell structures. 
>> The outcome of a specific situation is defined exactly by the initial 
>> settings.
>
>The only way to do this accurately (that I know of) is to read data from 
>your universe and save the results in a buffer array which is then 
>transfered to the universe once all the cells have been calculated.  That 
>way, it makes absolutely no difference what order the cells are processed 
>in.

If you are not worried about strict determinism, then how about:

    repeat
        pick a location at random
        process it
    until satisfied with the result

If your run goes for long enough and your cell-processing doesn't rely on
all cells being exactly the same age, then everything more or less evens out.

Still having a cow,

+--------------------------------------------------------------------------+
| Shaun Green               | "To err is human, to moo bovine"             |
| DSL Systems Centre, CSIRO | Don't bust me: I have to rent my disclaimer! |
+--------------------------------------------------------------------------+


