- GP -> evolution of programs. An easy way to evolve

programs is to evolve a tree, were the nodes are

operations like +, -, sin, div. To evolve a program

you must choose a computing model : trees,

stack-based, etc... an then you evolve sentences from

this computing model.

Good article this place

http://en.wikipedia.org/wiki/Genetic_programming

--- Will Sturgis <cassiniovals@...> a écrit :

---------------------------------

I am new to gp and need some hints on how to evolve

expressions. I have a fair amount of experience with

genetic algorithms, but I am having trouble making the

leap from finding a value and finding an equation.

I downloaded PerlGP, found PI, and some other

constants, so I know it works correctly.

My internet searches have not yielded very much, but

maybe I am not looking in the right place.

Any help would be greatly appreciated.

-Will

__________________________________

Yahoo! Mail - PC Magazine Editors' Choice 2005

http://mail.yahoo.com

---------------------------------

YAHOO! GROUPS LINKS

Visit your group "genetic_programming" on the web.

To unsubscribe from this group, send an email to:

genetic_programming-unsubscribe@yahoogroups.com

Your use of Yahoo! Groups is subject to the Yahoo!

Terms of Service.

---------------------------------

_______________________________

Marmakoide aka Alexandre Devert

_______________________________

___________________________________________________________________________

Appel audio GRATUIT partout dans le monde avec le nouveau Yahoo! Messenger

Téléchargez cette version sur http://fr.messenger.yahoo.com - Hi Will,

Sorry for the delay - I started a new job this week. I can certainly help

you with PerlGP, since I wrote it! But we can do that off-list. First, take

a look at the "How it all works" section of the manual. It goes through an

example doing symbolic regression of a sine function (an expression/equation

of 'x' is produced).

http://perlgp.org/docs/manual/manual/node11.html

Note that in the sin(x) example, the data files (train.dat and test.dat) are

generated on-the-fly because I did resampling and other tricks. You will

probably want to read in static data, and that is no problem - just remove

the 'refresh' method.

Anyway, as I said, I'm happy to help!

cheers,

Bob

> ---------------------------------

[Non-text portions of this message have been removed]

> I am new to gp and need some hints on how to evolve

> expressions. I have a fair amount of experience with

> genetic algorithms, but I am having trouble making the

> leap from finding a value and finding an equation.

>

> I downloaded PerlGP, found PI, and some other

> constants, so I know it works correctly.

>

> My internet searches have not yielded very much, but

> maybe I am not looking in the right place.

>

> Any help would be greatly appreciated.

>

> -Will

>

>

>

>

> GP -> evolution of programs. An easy way to evolve

And additionnaly, if you need reasonable speed, stack-based is

> programs is to evolve a tree, were the nodes are

> operations like +, -, sin, div. To evolve a program

> you must choose a computing model : trees,

> stack-based, etc... an then you evolve sentences from

> this computing model.

generally faster. There is a C implementation called XCLE of a generic

stack-based interpreter, where you can choose the primitives (i.e.

nodes) you want to use, or implement yours... In google, first result

should be the FSF entry for XCLE.

But of course, perlPG is a lot more flexible, if possibly a bit

slower. You can't beat Perl at that.