Hi all,
when running a parameter optimization, I understand that Zorro selects the best parameter values according to the most stable/robust area and not the peak of the fitness function (as defined by objective() within default.c)

I would like to better understand how the parameter selection is made, specially under brute force optimization (i.e. all possible combinations).

When there is only one parameter, the selection is straight forward and I can check it directly with a 2D plot of the fitness function result vs parameter values.
When there are 2 parameters to be optimized, I undersant Zorro select the best combination by finding the most stable area in a 3D plot. Is this correct?
What happens when there are 3 or more parameters? Is Zorro selecting the most stable n-dimensional area of the fitness function, or it is looking at single or pairs of parameters to decide?

And last question: is it possible to customize the parameter selection algorithm after an optimization cycle, to be able to define my own?

Thanks for your help!

The parameter dimension is not relevant for the peak filtering, since you can apply a lowpass filter in any dimension. But the training method is. It is only implemented for Ascent and brute force, not for Genetic. For implementing your own algorithm, look in the manual for the parameters() function.

Thanks for the info. When looking into the parameters() function, I understand that it allows me to customize the parameter list which is evaluated at each optimization iteration, however what I am interested is in applying my own algorithm to the full list of results calculated through brute force. Is there any built in function which I can modify to do that, or would I need to manually do it after the optimize loop is completed?

Thanks