I had a similar problem before and then discovered that the parameter  
needed to be the following:

breed.elite.0 = 10

-Dave

On Nov 19, 2007, at 5:37 AM, Anna Esparcia wrote:

> I'm running a very simple problem which uses simple.params, plus some
> modifications on tutorial2.params. When looking at the out.stats  
> file I
> noticed that the best fitness went up and down from one generation  
> to the
> next. I thought this could be sorted using elitism, so i added the  
> following
> line to the params file:
>
> breed.elites.0 = 10
>
> I thought this would keep the best 10 individuals from one  
> generation to the
> next and hence, the fitness of the best individual would only  
> increase.
> However, this is not the case and I still see the fitness  
> oscillate. Is this
> normal behaviour? Were am I going wrong?
>
> Cheers,
> Anna