[Coin-ipopt] Re: ipopt

Juan Jose Arrieta-Camacho jarrieta at andrew.cmu.edu
Fri Jun 17 19:30:04 EDT 2005


> I am trying to use your ipopt for my project, which, for a moderate
> problem, can easily have 20 million variables. Do you think IPOPT can
> handle so large system, and do you have an estimation of the cpu and
> memory it requires on a linux of 1.5G CPU?
 What type of application are you talking about? I assume ---given the problem size--- we are talking about a PDE constrained optimization or probably the optimal control of a large scale nonlinear system.
Also, how would you solve the problem in a 1.5Gb computer? Maybe the memory allocation would not suffice.

 Finally, I don't think IPOPT is constrained _theoretically_ to solve such a problem (nor MA27); the issues arise in the implementation. Do you have a code describing the problem? Can this code return derivatives? Do you plan to implement it from scratch? Maybe an iterative linear solver would do the trick if MA27 fails, but the implementation would be nontrivial.

 Hopefully, you can describe in greater detail the type of application you are dealing with, so we can get some input from the community.

 Have a nice day,

 Juan Arrieta

| J. J. Arrieta-Camacho                       5000 Forbes Avenue |
| Aerospace Systems Engineering                Doherty Hall 4200 |
| Carnegie Mellon University                 Pittsburgh PA 15213 |
| Mobile: (+1) 412.901.9184            Office: (+1) 412.268.2543 |
|         http://andrew.cmu.edu/user/jarrieta                    |
                        !  !  !            

More information about the Coin-ipopt mailing list