Tiny GP and random numbers - Field Guide to Genetic Programming

Message from discussion Tiny GP and random numbers
MIME-Version: 1.0
Message-ID: <81ecfb17-655f-44f4-b0d0-64580b7a7015@i76g2000hsf.googlegroups.com>
Date: Thu, 8 May 2008 05:45:30 -0700 (PDT)
In-Reply-To: <4764a9bd-215d-4f83-9b27-0a246d40e6dc@q1g2000prf.googlegroups.com>
References: <4764a9bd-215d-4f83-9b27-0a246d40e6dc@q1g2000prf.googlegroups.com>
User-Agent: G2/1.0
Subject: Re: Tiny GP and random numbers
From: rp...@essex.ac.uk
To: Field Guide to Genetic Programming <field-guide-to-genetic-programming@googlegroups.com>
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

Hi,

the TinyGP program was designed to do continuous symbolic regression.
If one replaced the arithmetic operations in the eval procedure with
Boolean functions one could also easily do Boolean induction problems.
Anything that deviates significantly from symbolic regression,
however, requires bigger changes. In particular, I don't think it is
possible to get what you are asking for just by changing some numbers
in the input file for TinyGP. Of course, nothing would prevent the
addition of a rand() function (with no arguments), but the fitness
function would definitely be changed.

I hope this helps.

Riccardo

On May 6, 12:12=A0pm, sanny <rs.sa...@gmail.com> wrote:
> hallo,
>
> i'm using java program, Tiny GP (written by Riccardo Poli), to learn
> about
> Genetic Programming.
>
> is it possible to this program using random numbers as the input
> (terminal) that will result an output with random numbers as a
> variable?
> and if it's possible, how should i write the input file?
>
> is anybody have studied this program? i really need a help.
> thanks
>
> --
> =A0best regards,