qpOASES
qpOASES copied to clipboard
coin-or
Possible bug when using equality constraints
Issue created by migration from Trac.
Original creator: gabrielenava
Original creation time: 2018-08-02 17:15:32
Assignee: ferreau
Version: 3.2.0
Hi all, I think I may have found a bug inside the qpOASES algorithm when performing a constrained optimization with equality constraints in MATLAB. In particular, I considered the following optimization problem:
min |u|^2
s.t. Au = 0
where u is a [29 * 1] vector and A is a [6 * 29] matrix with rank 6. In theory, this problem is well posed and the solution is u = 0. However, for some valid constraint matrices A qpOASES cannot find a solution, and returns the following error:
ERROR: Division by zero
->ERROR: Abnormal termination due to TQ factorisation
->ERROR: Determination of step direction failed
->ERROR: Abnormal termination due to TQ factorisation
->ERROR: Initialisation failed! QP could not be solved!
Attached to this script there is a matlab code with an example. The constraint matrix is stored in a .mat file (attached to the issue, too). Also, relaxing the problem with inequality constraints of the form:
min |u|^2
s.t. -eps < Au < eps,
eps > 0
did not solve the issue. As an additional information, I noted that in case the problem is really unfeasible the error from qpOASES is different.
May you help me in understanding this issue? Thanks in advance.
This issue is quite old by now, but in case someone else runs into this with a deadline too close to try switching to another solver (like me!), I ran into the same error with a very similar problem setup (but using the C++ interface) and have seemingly been able to resolve it. My problem included a constraint of the form Ax = b, where A is 6x9, and the problem is known to be feasible.
To fix it, I set the solver options to "reliable". In C++, this is:
qpOASES::Options options;
options.setToReliable();
// qp is my QProblem object
qp.setOptions(options);
(see page 30 of the user manual).
I have the same issue. setting
options.setToReliable();
didn't really solve the problem for me. I can solve a Problem with 148 constraints but as soon as the number of constraints is increasing, say to 180, I again get the error.
