You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
In doing a relatively involved optimization with NLopt, I find that it seems to ignore stopval!() and ftol_rel!(), though maxtime!() is working. I'm using the optimizer LN_COBYLA.
Based on the links below, I have some sense it may be about the optimizer or the constraints (though the inequality constraints don't seem to be violated at the value it returns at maxtime, which has a function value well smaller than the stopval), but wanted to check if it was a bug/known issue, or if there is some clear way to get the optimization to adhere to the stopval.
Closing as stale, and because there is no reproducible example. If anyone encounters this in future, please post on the community forum, https://discourse.julialang.org/c/domain/opt/13, and tag me @odow.
In doing a relatively involved optimization with NLopt, I find that it seems to ignore stopval!() and ftol_rel!(), though maxtime!() is working. I'm using the optimizer LN_COBYLA.
Based on the links below, I have some sense it may be about the optimizer or the constraints (though the inequality constraints don't seem to be violated at the value it returns at maxtime, which has a function value well smaller than the stopval), but wanted to check if it was a bug/known issue, or if there is some clear way to get the optimization to adhere to the stopval.
https://stackoverflow.com/questions/34755612/unexpected-behaviour-of-ftol-abs-and-ftol-rel-in-nlopt
https://discourse.julialang.org/t/nlopt-stopval-and-termination-criteria/7333/2
The text was updated successfully, but these errors were encountered: