[Ipopt] ipopt in structural optimization

Wed Jun 10 20:16:57 EDT 2009

Hi Peter,

>>>     2) We “warmstart” the optimization each iteration and
>>>     leave the optimization to calculate the new primary
>>>     solution. Thus we have changed the MMA so that we call it
>>>     with objective function, constraints and sensitivities of
>>>     the actual iteration plus some saved information from
>>>     former iterations, e.g. the last asymptotes and other
>>>     values. Could this be easily done with Ipopt ?

> Has anyone a comment on the “warmstart” issue?

I don't quite understand what you mean by this.  If you use Ipopt, it 
should usually be in control of choosing the primal variables, and the 
application code computes the values and derivatives of the functions in 
the optimization problem formulation.

There is an option to warmstart Ipopt (see the documentation), where you 
can not only give a primal starting point, but also values for all 
multipliers.  However, the formulation that Ipopt uses internally is 
somewhat different from what is exposed to the user, and some translation 
happens when the starting point is taken, or the final iterate is 
returned.  Also, there are a number of options that play a crucial role 
when trying to use Ipopt in a warmstart setting, such as

mu_init
bound_push
bound_frac

and possibly others.  And this does not restore exactly the state that 
Ipopt was when it most recently returned.  Finally, as an interior point 
method, Ipopt is not as good as making use of a very good starting point 
as an active set method typically is.

By the way, Ipopt's runtime performance per iteration will critically 
depend on the linear solver you pick, and whether you use an efficient 
BLAS library.  I would suggest to try Pardiso for such large problems, 
together with some good BLAS.

Sorry for the late reply, and maybe this helps a little bit...

Andreas

>
> Thanks in advance.
>
> /Peter
>
>> Ruhollah Tavakoli wrote:
>>> Peter
>>>
>>> Look at SICON 06 paper of Martin Burger
>>> (http://dx.doi.org/10.1137/05062723X), He used IPOPT to solve
>>> discretized form of structural topology optimization.
>>>
>>> In my experience, IPOPT is *very* robust for such problems, but is
>>> very expensive in contrast to e.g. OC or MMA.
>>>
>>> Cheers
>>>
>>> RT
>>>
>>> On Wed, May 27, 2009 at 9:06 AM, Clausen Peter M.
>>> <peter.clausen at fe-design.de <mailto:peter.clausen at fe-design.de>> wrote:
>>>
>>>     Hello
>>>
>>>     I am working at FE-DESIGN GmbH (www.fe-design.com
>>>     <http://www.fe-design.com/>) and our
>>>     program TOSCA Structure is a structural optimization
>>>     program. We use the Method of Moving Asymptotes (MMA of
>>>     Krister Svanberg) as optimizer which in most cases does a
>>>     really good job. In some cases though it has its drawbacks
>>>     (linear object function, constraints which become active
>>>     later in optimization) and we would like to test an
>>>     alternative.
>>>
>>>     Our problems may exceed 1 mio. design variables and we
>>>     usually only need around 30-50 iterations to converge. The
>>>     primal solution is a FEM system solved by an external
>>>     program, the sensitivity analysis (1. order) is done by
>>>     ourselves. We have no 2. order information.
>>>
>>>     My questions:
>>>     1) How would you expect Ipopt to perform in a benchmark
>>>     against MMA ? Has anybody got experiences?
>>>
>>>     We would prefer to use the fortran version, but
>>>     interfacing to C++ is nothing new to us. It would simply
>>>     mean that Ipopt’s objects would be created and destroyed
>>>     each iteration (would this be a performance problem?).
>>>
>>>     Best regards
>>>
>>>     Peter M. Clausen, developer at FE-DESIGN GmbH
>>>
>>>     _______________________________________________
>>>
>>>      Peter M. Clausen
>>>
>>>      Email: peter.clausen at fe-design.de
>>> <mailto:peter.clausen at fe-design.de>
>>>      Tel. : +49 721 96467 - 236
>>>     _______________________________________________
>>>
>>>      FE-DESIGN GmbH, Haid-und-Neu-Str. 7, 76131 Karlsruhe,
>>>     Germany
>>>      Registration Court: D-Mannheim HRB 107382
>>>      Director: Dr. Juergen Sauter
>>>      Fax: +49 (0)721 96467-290
>>>      Web: www.fe-design.com <http://www.fe-design.com/>
>>>     _______________________________________________
>>>
>>>     --
>>>     This email was Anti Virus checked by Astaro Security Gateway.
>>>     http://www.astaro.com <http://www.astaro.com/>
>>>
>>>     _______________________________________________
>>>     Ipopt mailing list
>>>     Ipopt at list.coin-or.org <mailto:Ipopt at list.coin-or.org>
>>>     http://list.coin-or.org/mailman/listinfo/ipopt
>>>
>>>
>>
>
> -- 
> Mit freundlichen Grüßen / Best Regards
>
> Peter M. Clausen
>
> _______________________________________
>
>  Peter M. Clausen
>
>  Email: peter.clausen at fe-design.de
>  Tel. : +49 721 96467 - 236
> _______________________________________
> _______________________________________________________________
>
>  FE-DESIGN GmbH, Haid-und-Neu-Str. 7, 76131 Karlsruhe, Germany
>  Registration Court: D-Mannheim HRB 107382
>  Director: Dr. Juergen Sauter
>  Fax: +49 (0)721 96467-290
>  Web: www.fe-design.com
> _______________________________________________________________
>
> Veranstaltungsvorschau:
> +++ 10.-13.März 2009 in Lyon: "Industrie Lyon 2009" +++
> +++ 23.-24.März 2009 in London: "STAR European Conference+++
> +++ 30. März – 01. April 2009 in Karlsruhe: „TOSCA Structure Einführungsseminar (Deutsch)“ +++
> +++ 20.04.-24.04.2009 in Hannover: "HMI"+++
> +++ 27.04.-28.04.2009 in Graz: "2nd Graz Symposium Virtual Vehicle"+++
> +++ 02. April 2009 in Karlsruhe: „TOSCA Sructure Advanced Seminar (Deutsch)“ +++
> +++ 20.04.-24.04.2009 in Hannover: "HMI"+++
> +++ 27.04.-28.04.2009 in Graz: "2nd Graz Symposium Virtual Vehicle"+++
> +++ 18.-21. Mai 2009 in London: "SIMULIA Customer Conference" +++
>
> Event Preview:
> +++ March 10-13, 2009 in Lyon/France: "Industrie Lyon 2009" +++
> +++ March 23-24, 2009 in London/GB: "STAR European Conference"+++
> +++ March 30-April 01, 2009 in Karlsruhe/Germany: “TOSCA Structure Basic Seminar (German)”
> +++
> +++ April 02, 2009 in Karlsruhe/Germany: “TOSCA Structure Advanced Seminar (German)+++
> +++ April 20-24, 2009 in Hannover/Germany: "HMI"+++
> +++ April 27-28, 2009 in Graz/Austria: "2nd Graz Symposium Virtual Vehicle"+++
> +++ May 18-21, 2009 in London/GB: "SIMULIA Customer Conference"+++
>
> _______________________________________________
> Ipopt mailing list
> Ipopt at list.coin-or.org
> http://list.coin-or.org/mailman/listinfo/ipopt
>
>