[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

sqp function

From: Siddharth Jain
Subject: sqp function
Date: Tue, 8 Oct 2013 09:05:45 +0530

I am facing a problem with optimizing a problem with sqp. Sqp throws an "error: sqp: operator *: nonconformant arguments" in 10th or 11th iteation. To be specifc on the following lines of sqp.m:

if (! isempty (A))
t = ((A_new - A)'*lambda);
y -= t;

After looking up this around, I found sqp failed, because qp was not able to optimize the problem in that iteration. This led me to bug #32717. Gabriele, the person who filed bug #32717, mentions "I briefly skimmed the relevant chapter of Nocedal and Wright optimization book, which maybe was used as a reference for writing the code, and the solution does not seem trivial and may involve using slack variables and large linear penalties to ensure that the qp sub-problems are always feasible at each iteration. "  Does anyone think this is possible?

Also while I was going through sqp.m,  I found a comment " ## Check QP solution and attempt to recover if it has failed.", has this been implemented now?

Also bug#38875 mentions some problem with sqp in octave 3.6.4, do you think falling back to older versions might help?

This specific problem we are dealing with is feasible, as we do get results in matlab. Is there any version of octave in which sqp will be able to recover from the failed qp and continue optimization? or is there any way how we can fix it?

Any help will be highly appreciated.

Best regards,
Siddharth Jain

reply via email to

[Prev in Thread] Current Thread [Next in Thread]