Hi Robert, we're not actively avoiding you!  :-)  I'm in Kyoto on  

very limited bandwidth, and Liviu's working on his dissertation  

(which he's defending in a few days).  Liviu's the man you want to  

have look into this anyway -- he wrote the coevolution code and knows  

it better than I do.  Give him a little while though!  :-)



Sean



On Aug 23, 2006, at 2:27 AM, Robert Baruch wrote:



> OK, I see the problem. It was based on a misunderstanding of the  

> GroupedProblemForm.evaluate

> parameters. I thought that the ind[] array would have ind[0] the  

> individual under evaluation, and ind

> [1] the opponent. That's how it was when I only had one  

> subpopulation and was running the

> CompetitiveEvaluator.

>

> I switched to MultiPopCoevoluationaryEvaluator, figuring it would  

> do the same thing, but it doesn't.

> Apparently the ind[] array is the size of the number of subpops,  

> and the update[] element that's true

> is the individual under evaluation.

>

> Now the only problem is that it seems that ind[0] is always an  

> individual from subpop 0, ind[1] is an

> individual from subpop 1, and ind[2] is one from subpop 2. What I  

> really wanted was to choose a

> random individual from a random subpop (not in my own subpop).

>

> I guess that means that in my GPProblem's evaluate method, I will  

> have to randomly choose one of

> the non-updated ind's.

>

> It would be nice to be able to choose the number of opponents in  

> the parameters for

> MultiPopCoevolutionaryEvaluator...

>

> --Rob