Unfortunately, there's not currently a way to register new search algorithms 
with NLopt, except by hacking the NLopt source code directly.

On Nov 4, 2013, at 5:51 AM, federico vaggi wrote:

> Hi everyone,
> 
> I'm currently using the Python wrappers of NLopt.  I'm working on a 
> non-convex least squares problem, where the local search portion of the 
> optimization is best handled by a LS algorithm like Levemberg Marquardt (the 
> SciPy implementation, or any other one, really).
> 
> Originally, I was thinking of simply doing something like Latin Hypercube 
> Sampling to explore the full parameter landscape, but I was wondering if I 
> could combine the local least-squares optimization step with one of the 
> global algorithms in NLopt like MLSL to search the whole parameter landscape 
> more efficiently.
> 
> Federico
> _______________________________________________
> NLopt-discuss mailing list
> [email protected]
> http://ab-initio.mit.edu/cgi-bin/mailman/listinfo/nlopt-discuss


_______________________________________________
NLopt-discuss mailing list
[email protected]
http://ab-initio.mit.edu/cgi-bin/mailman/listinfo/nlopt-discuss

Reply via email to