[jgap-users] GABreeder / BulkFitnessFunction question

[Sjs S. <arc...@ho...>]

Hello JGAP development team,=20
I had a question/bug on the GABreeder. I don't know if this is a bug or not=
.=20
I'm using JGAP 3.2.2 distribution.=20
In
the evolve(pop) method of GABreeder, the bulk fitness function is
evaluated AFTER the natural selectors and genetic operators are
applied. This causes NullPointerExceptions inside the
BestChromosomesSelector natural selector, as well as in my custom
natural selector. Should the bulk fitness function be evaluated before
the natural selectors and genetic operators are applied? By comparison,
if there is no bulk fitness function, the chromosomes are each
evaluated BEFORE the natual selectors and genetic operators are applied
(lines 69 - 74).=20

If this behavior is intentional, can you
please tell me how I can avoid the NullPointerExceptions I get when
BestChromosomesSelector tries to determine which chromosomes are best,
even though all their multi objectves are null?=20

Thanks for the help,=20
        -Simon-=20
_________________________________________________________________
Windows Live Hotmail and Microsoft Office Outlook =96 together at last. =A0=
Get it now.
http://office.microsoft.com/en-us/outlook/HA102225181033.aspx?pid=3DCL10062=
6971033=