Hi, First of all thank you very much for creating NLOPT and making it free/open-source - it is a very powerful package which has been easy to use.
I am using NLOPT to do non-linear least squares fitting of a model (w/o analytic Jacobian) to experimental data. BOBYQA seems to work well for this case. I was wondering if there is any existing way to obtain the covariance matrix and/or Jacobian and/or uncertainty/error estimates for the optimized fit parameters? I believe these are typically calculated using the Jacobian. See for example: https://www.mathworks.com/help/stats/nlinfit.html CovB = inv(J'*J)*MSE Unfortunately I did not see anything regarding this in the NLOPT documentation. Secondly, are there any plans to include an implementation of the Levenberg-Marquadt algorithm, which is quite popular for non-linear least squares fitting, in the NLOPT package? See for example: https://www.physics.wisc.edu/~craigm/idl/cmpfit.html Thanks, Shaun
_______________________________________________ NLopt-discuss mailing list NLopt-discuss@ab-initio.mit.edu http://ab-initio.mit.edu/cgi-bin/mailman/listinfo/nlopt-discuss
