Hello everyone,
It is often useful for diagnostic or tracking purposes to print out
information about the progress of an optimization at each iteration.
Some libraries (e.g. pylab.optimize) allow for a callback function to be
defined for that should handle any per-iteration processing like this.
For some simpler algorithms, putting this code in the objective function
makes sense, but for others that might make many "exploratory" calls to
the objective function without taking an actual step (e.g. gradient
algorithms), this will produce a large amount of undesired noise that is
often indistinguishable from the desired output.
Are there any work-arounds that I'm unaware of, and if not, any plans
for this kind of functionality in the future?
Thanks,
Greg Nicholas
_______________________________________________
NLopt-discuss mailing list
[email protected]
http://ab-initio.mit.edu/cgi-bin/mailman/listinfo/nlopt-discuss
