Possibly Low-Storage BFGS or the Shifted limited-memory variable metric. I am unfamiliar with the latter.
-Grey On Sep 5, 2014, at 10:07 AM, Miguel Angel Salazar de Troya <[email protected]> wrote: > I'm using the MMA method, I believe it requires a hessian approximation? What > are the alternatives? I thought this method was popular for topology > optimization, where usually many variables are employed. > > Take care > Miguel > > > On Fri, Sep 5, 2014 at 8:21 AM, Grey Gordon <[email protected]> wrote: > Sounds like you may just be running out of memory causing a segfault. If you > are using some derivative method that requires a hessian approximation, 1000 > x 1000 matrix is quite large. > > Best, > Grey Gordon > > > On Sep 4, 2014, at 11:51 PM, Miguel Angel Salazar de Troya > <[email protected]> wrote: > >> Hi all >> >> I'm trying to run nlopt inside of a loop. For each iteration, I increase the >> number of variables to optimize, using the last result as initial estimation >> (it's a field in a mesh that has been refined, hence the increase in the >> number of variables) and inialize again nlopt, but I'm running into some >> segfault problems that are hard to track (it only happens with a large >> number of variables, +1000 thousand, and valgrind runs really slow for those >> cases). >> >> I would just like to know if this is possible (running nlopt inside a for >> loop) and what is the best way to do it. >> >> Thanks >> Miguel >> -- >> Miguel Angel Salazar de Troya >> Graduate Research Assistant >> Department of Mechanical Science and Engineering >> University of Illinois at Urbana-Champaign >> (217) 550-2360 >> [email protected] >> >> _______________________________________________ >> NLopt-discuss mailing list >> [email protected] >> http://ab-initio.mit.edu/cgi-bin/mailman/listinfo/nlopt-discuss > > > > > -- > Miguel Angel Salazar de Troya > Graduate Research Assistant > Department of Mechanical Science and Engineering > University of Illinois at Urbana-Champaign > (217) 550-2360 > [email protected] >
_______________________________________________ NLopt-discuss mailing list [email protected] http://ab-initio.mit.edu/cgi-bin/mailman/listinfo/nlopt-discuss
