On Fri, 01 Apr 2005 14:52:03 +0100, Rob Myers wrote
> (with-delurk
>
> On Friday, April 01, 2005, at 01:40PM, Dave Griffiths
> <dave_at_pawfal.org> wrote:
>
> >Coming back to what I first blurted out, I have to admit that there is
> >something about the pure GP approach with fitness functions and so on, that
> >doesn't sit well for me.
>
> Supervised evolution, the best-known examples probably being the art
> of Karl Sims or William Latham, might be the sweet spot for live
> genetics when combined with the editor for tweaking the code.
>
> You could have the variations played in order and mark the best ones
> as they play to be the seed for the variations for the next sequence.
This is what I call manual selection, as opposed to automatic selection by
fitness function.
fwiw, Mr Latham used to be my boss. A game with a genetic type interface:
http://pc.ign.com/articles/161/161756p1.html
dave
Received on Fri Apr 01 2005 - 15:05:19 BST