[Ipopt-tickets] [Ipopt] #112: Incorrect number of nonzeros in Lagrangian Hessian causes MA27 to fail

[Ipopt]

#112: Incorrect number of nonzeros in Lagrangian Hessian causes MA27 to fail
-------------------------------------------+--------------------------------
Reporter:  arrieta                         |       Owner:  andreasw
    Type:  clarification                   |      Status:  new     
Priority:  lowest                          |   Component:  Ipopt   
 Version:  3.7 (C++ Version)               |    Severity:  trivial 
Keywords:  hessian lagrangian derivatives  |  
-------------------------------------------+--------------------------------
Changes (by andreasw):

  * owner:  ipopt-team => andreasw


Comment:

 Hi Juan,

 I would be happy to include code that checks the correctness of the
 Hessian, even though the check should probably be enabled only on demand
 to reduce runtime.

 However, except for making sure that the indices are within bounds (and
 MA27 is already flagging this if it is violated) I don't know what the
 check would be.  The user can provide the lower OR the upper triangular
 part of the Hessian, or even mixed, and also provide entries multiple
 times, in which case they are added.

 So, let me know if you want to contribute code for Hessian correctness
 checks.

 Thanks

 Andreas

-- 
Ticket URL: <https://projects.coin-or.org/Ipopt/ticket/112#comment:1>
Ipopt <http://projects.coin-or.org/Ipopt>
Interior-point optimizer for nonlinear programs.