[Octave-bug-tracker] [bug #50067] function: qp.m, incorrect vectors dimension

[nicodemus]

URL:
  <http://savannah.gnu.org/bugs/?50067>

                 Summary: function: qp.m, incorrect vectors dimension 
                 Project: GNU Octave
            Submitted by: nicmus
            Submitted on: Tue 17 Jan 2017 03:07:14 PM GMT
                Category: Octave Function
                Severity: 3 - Normal
                Priority: 5 - Normal
              Item Group: Crash
                  Status: None
             Assigned to: None
         Originator Name: 
        Originator Email: 
             Open/Closed: Open
         Discussion Lock: Any
                 Release: 4.2.0
        Operating System: Microsoft Windows

    _______________________________________________________

Details:

With the new release 4.2.0 the call to the function qp.m results in a crash
for an infeasible initial guess x0. 

The function that aborts the execution is glpk called at qp.m:380.

The problem seems to be due to bad dimension of vectors ctmp and lb.

Octave 4.2.0 on microsoft windows
------------------------------------
376: ctmp = [zeros(n, 1); ones(n_in, 1)];
377: lb = [-Inf(n,1); zeros(n_in,1)];

Octave 4.0.3 on linux
------------------------------------
376: ctmp = [zeros(n-n_eq, 1); ones(n_in, 1)];
377: lb = [-Inf(n-n_eq,1); zeros(n_in,1)];

The problem is solved by using the old version of qp.m




    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?50067>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/