When I asked how about computing in parallel, I wasn't meaning computing the constraints in parallel, that I know I can do it using the multiple constraint method, but computing the constraints in parallel with the energy function.
For the energy function I'm using CUDA and it takes a while, so it will be great to use that CPU idle time to compute the constraints. 2016-11-10 23:40 GMT+00:00 Juan Jose Casafranca <[email protected]>: > Then, coming back to the first method to express my constraints, as I have > a lot of them, is it possible to compute them simultaneously in different > cores? > > Thanks for your answer > > 2016-11-10 17:00 GMT+00:00 <[email protected]>: > >> Send NLopt-discuss mailing list submissions to >> [email protected] >> >> To subscribe or unsubscribe via the World Wide Web, visit >> http://ab-initio.mit.edu/cgi-bin/mailman/listinfo/nlopt-discuss >> or, via email, send a message with subject or body 'help' to >> [email protected] >> >> You can reach the person managing the list at >> [email protected] >> >> When replying, please edit your Subject line so it is more specific >> than "Re: Contents of NLopt-discuss digest..." >> >> >> Today's Topics: >> >> 1. Fwd: Action after step is done (Juan Jose Casafranca) >> 2. Re: Fwd: Action after step is done (Steven G. Johnson) >> >> >> ---------------------------------------------------------------------- >> >> Message: 1 >> Date: Wed, 9 Nov 2016 19:15:07 +0000 >> From: Juan Jose Casafranca <[email protected]> >> To: [email protected] >> Subject: [NLopt-discuss] Fwd: Action after step is done >> Message-ID: >> <CAMe8Lxsv_CHHqtP9+v-jMkx+yBE7GMpSV5=oo_r4FAJbJ9sn6Q@mail. >> gmail.com> >> Content-Type: text/plain; charset="utf-8" >> >> Hi, >> >> I'm using NLOPT for a project where I need to optimize a non linear >> function subject to non linear constraints. I have found that I can >> express >> my constraints in several ways. In one way I have to deal with a very high >> number of constraints while in the other way I can reduced it due to some >> equalities. The problem is that this second way is not easy to be >> expressed >> as a mathematical expression. I have several polylines and each one >> defines >> a zone in space where it has control. I'm trying to optimize the position >> of this polylines and the constraints says that the new points must be in >> the zone of space that it's under the polyline control. The space is >> sampled as a regular grid. >> >> As you can see, its not easy to define the mathematical expression for the >> constraint so I was wondering if it is possible to call a user function >> after each step the optimization algorithm performs. This would allow me >> to >> project the points to the feasible space after the step is done. >> >> Thanks! >> -------------- next part -------------- >> An HTML attachment was scrubbed... >> URL: <http://ab-initio.mit.edu/pipermail/nlopt-discuss/attachment >> s/20161109/0db1df05/attachment-0001.html> >> >> ------------------------------ >> >> Message: 2 >> Date: Wed, 9 Nov 2016 15:46:30 -0500 >> From: "Steven G. Johnson" <[email protected]> >> To: nlopt-discuss <[email protected]> >> Subject: Re: [NLopt-discuss] Fwd: Action after step is done >> Message-ID: <[email protected]> >> Content-Type: text/plain; charset=us-ascii >> >> >> > On Nov 9, 2016, at 2:15 PM, Juan Jose Casafranca <[email protected]> >> wrote: >> > >> > Hi, >> > >> > I'm using NLOPT for a project where I need to optimize a non linear >> function subject to non linear constraints. I have found that I can express >> my constraints in several ways. In one way I have to deal with a very high >> number of constraints while in the other way I can reduced it due to some >> equalities. The problem is that this second way is not easy to be expressed >> as a mathematical expression. I have several polylines and each one defines >> a zone in space where it has control. I'm trying to optimize the position >> of this polylines and the constraints says that the new points must be in >> the zone of space that it's under the polyline control. The space is >> sampled as a regular grid. >> > >> > As you can see, its not easy to define the mathematical expression for >> the constraint so I was wondering if it is possible to call a user function >> after each step the optimization algorithm performs. This would allow me to >> project the points to the feasible space after the step is done. >> >> No, optimization algorithms will typically not work very well if you >> discontinuously project points back into the feasible region after each >> step. If I were you I would try to parameterize the polylines in such a >> way that they are automatically in the desired spatial zone. >> >> >> ------------------------------ >> >> Subject: Digest Footer >> >> _______________________________________________ >> NLopt-discuss mailing list >> [email protected] >> http://ab-initio.mit.edu/cgi-bin/mailman/listinfo/nlopt-discuss >> >> >> ------------------------------ >> >> End of NLopt-discuss Digest, Vol 81, Issue 1 >> ******************************************** >> > >
_______________________________________________ NLopt-discuss mailing list [email protected] http://ab-initio.mit.edu/cgi-bin/mailman/listinfo/nlopt-discuss
