Re: Fuseki and Joseki Database with Neural Network

[Juergen Kahnert <Kahnert@xxxxxxxxxxxxxxxx>]

Gary Boos wrote:
>
> I think GA will be better.
>

I'm sorry, but i can't follow you.
To use genetic algorithm you need a quick and easy evaluation function.
But that's the problem. We are not able to do a simple test and get a
result which one is better...
For travelling salesman it's easy to calculate the way from each
generation.
I think GA is good to optimize things if you are able to quickly
evaluate each generation because you need hundreds or thousands much
more generations and to get a result before the sun explodes it has to
bee quick.

	Juergen