Error with inequality constraints with patternsearch
1 Ansicht (letzte 30 Tage)
Ältere Kommentare anzeigen
Tan
am 12 Feb. 2011
Beantwortet: Xinlei Wang
am 7 Jul. 2017
Hi,
I encountered the following error while using patternsearch in Matlab :???
Error using ==> lcondirections at 39
Constraints are dependent at current iterate
Try increasing OPTIONS.TolBind (<eps).
The default value of TolBind is already large enough and anyway I tried to increase the TolBind value but to no avail. Does anyone here encountered a similar problem or know how to solve it? Thanks.
0 Kommentare
Akzeptierte Antwort
Alan Weiss
am 21 Mai 2014
TolBind relates to linear constraints. So I suggest that you use the GSS polling algorithm instead of the default GPS. Set the PollMethod option to 'GSSPositiveBasis2N' or 'GSSPositiveBasisNp1' with psoptimset. See Compare the efficiency of poll options in the documentation.
Alan Weiss
MATLAB mathematical toolbox documentation
2 Kommentare
Mehul Agrawal
am 7 Jun. 2016
Bearbeitet: Walter Roberson
am 7 Jun. 2016
Hi Alan,
I tried both of these things in the poll method. But still got the same error.
Optimization running.
Error running optimization.
Constraints are dependent at current iterate.
Try increasing OPTIONS.TolBind (< eps).
Alan Weiss
am 8 Jun. 2016
You might also try different initial points. And try to determine whether your problem is feasible by using linprog to find an initial point that satisfies your bounds and linear constraints.
Alan Weiss
MATLAB mathematical toolbox documentation
Weitere Antworten (1)
Xinlei Wang
am 7 Jul. 2017
Hi Tan, I happened to encounter this error and, in my problem, solved it. My linear constraints were redundant. I deleted one of the constraints and then the solver works well. I think you could check whether your constraints are too many, if this is also your case.
0 Kommentare
Siehe auch
Kategorien
Mehr zu Direct Search finden Sie in Help Center und File Exchange
Produkte
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!