Daniele, you've got to have something misconfigured.  We've been using NSGA-II very heavily this year with no issues at all.

- Are you saying that the produced Pareto Front doesn't move to the corner of the alternative objective?

- Do a    -p print-params=true    and tell us what parameters were printed out during your run.

Sean

On Jun 6, 2012, at 1:08 PM, Danielle de Man wrote:

> Hi,
> I have noticed that when I use multi-objective fitness the selection process acts 
> a bit strange. It seems to select only for the first objective. When I switch the 
> two objectives the result is completely different as the first objective is always 
> optimized. I'm using the standard NSGA-2 algorithm so it should work fine. How is 
> this possible?
> Cheers,
> Danielle