Do the fitness scores also "jump around" when you run the first float-based regression problem that comes with the cgp package?
On May 21, 2012, at 10:31 AM, Andreas Meier wrote:
> With the help of the CGP framework, I implemented a small problem which is
> roughly the same as the first float-based regression problem that is part of
> the cgp-package. Maybe I'm doing something wrong but I'm wondering that the
> achieved fitness scores are "jumping around". Whereas in my classic ECJ/GP
> implementation the scores are monotonically decreasing with every generation
> in ECJ/CGP they often alternate and do not show such a trend. Do you have an
> idea what the reason for this behaviour could be?
> Thanks in advance!