[Ipopt] Start in positive area with optimal solution in negativearea

["Michael Gißler"]

Pardon, I tried to keep my request as short as possible this time..

Structural topology optimization is an optimization process based on finite elements models where a fe-solver solves functions values and gradients.

E.g. one could say that a certain weight of the structure is the constraint by minimizing the displacement of a special node of the fe-model by maniplulating the element densitiy which is linked to other mechanical properties such as stiffnesses of the finite elements.

For more information the wikipedia entry is helpfull:
http://en.wikipedia.org/wiki/Topology_optimization

With "area" I just mean the value of the objective function:

In my special case I try to optimize an fe-model of a mechanism.

The obj. function for this special opt. is the movement of a node.

At the beginning the node moves along the posivite x-axes, the obj. function is to minimize this movement. The clue is, that it is possible to get a result where the the position of this node ends on a negative x-value.


At the moment IPOPT's opt. ends up with a result where the x-axes displacement is nearly zero.

Opt. algorithms like the MMA from Svanberg could handle this.

So my question is: Is it possible in principile? Do I have to set some special values?


Best regards,

Michael


-------- Original-Nachricht --------
> Datum: Fri, 20 Jan 2012 09:12:00 -0500
> Von: "Ayotte-Sauvé, Étienne" <Etienne.Ayotte-Sauve at RNCan-NRCan.gc.ca>
> An: "Michael Gißler" <michael.gissler at gmx.net>, ipopt at list.coin-or.org
> Betreff: RE: [Ipopt] Start in positive area with optimal solution in negativearea

> 
> Hi Michael,
> 
> Is it possible to define the context, namely what you mean by "structural
> topology optimizations" and "area"?
> 
> Apologies in advance if the context is obvious or has already been defined
> in a previous message.
> 
> Regards,
> 
> Etienne Ayotte-Sauve  
> 
> -----Original Message-----
> From: ipopt-bounces at list.coin-or.org
> [mailto:ipopt-bounces at list.coin-or.org] On Behalf Of "Michael Gißler"
> Sent: January 20, 2012 8:50 AM
> To: ipopt at list.coin-or.org
> Subject: [Ipopt] Start in positive area with optimal solution in
> negativearea
> 
> Is it possible for IPOPT to optimize a NLP where the objective starts in a
> positive area but has it's optimum in negative one?
> 
> I try to run several structural topology optimizations which work quite
> well except that one special case.
> --
> Empfehlen Sie GMX DSL Ihren Freunden und Bekannten und wir belohnen Sie
> mit bis zu 50,- Euro! https://freundschaftswerbung.gmx.de
> _______________________________________________
> Ipopt mailing list
> Ipopt at list.coin-or.org
> http://list.coin-or.org/mailman/listinfo/ipopt

-- 
Empfehlen Sie GMX DSL Ihren Freunden und Bekannten und wir
belohnen Sie mit bis zu 50,- Euro! https://freundschaftswerbung.gmx.de