[Ipopt-tickets] [Ipopt] #263: Quasi-Newton approximation and the size of the hessian matrix

[Ipopt]

#263: Quasi-Newton approximation and the size of the hessian matrix
---------------------------------------+------------------------
Reporter:  leung                       |      Owner:  ipopt-team
    Type:  clarification               |     Status:  new
Priority:  normal                      |  Component:  Ipopt
 Version:  3.12                        |   Severity:  normal
Keywords:  Quasi-Newton approximation  |
---------------------------------------+------------------------
 I am trying to set up an optimization problem with a non-sparse Hessian
 matrix. My optimization problem, which has 100000 variables, will result
 in a really large Hessian matrix (with ~(100000*100000)/2 elements) if the
 derivatives are coded in the eval_h sub-routine. I was going through the
 IPOPT code, and it seemed like such a large matrix may not be necessary if
 I use the limited memory quasi-Newton approximation to approximate the
 hessian. Could anyone clarify whether my understanding is correct please?

--
Ticket URL: <https://projects.coin-or.org/Ipopt/ticket/263>
Ipopt <http://projects.coin-or.org/Ipopt>
Interior-point optimizer for nonlinear programs.