[Ipopt] Failure handling in ipopt

Stefan Vigerske stefan at math.hu-berlin.de
Sat May 29 05:09:24 EDT 2010


Hi,

you can return false in the evaluation functions.
If eval_f or eval_g returns false during linesearch, Ipopt should reduce
its steplength.

Stefan

Am 29.05.2010 07:48, schrieb ipopt at praveen.e4ward.com:
> Hi
> 
> I would like to ask if there is any way to handle failure of the simulation
> code ?
> 
> I am doing shape optimization using CFD. If the shape changes in a crazy
> way, the CFD will fail. Can we tell ipopt that there was a failure so that
> it should reduce the step size and then try again ?
> 
> My constraints involve flow and geometry. If ipopt tries to make a large
> shape change, then the geometry constraint will be violated. In that case
> ipopt should not ask for any flow related constraint but must change its
> search step, etc. and try again. I would like to know if it is possible to
> get this type of behaviour. If the geometry constraints are violated then
> the simulation will likely fail.
> 
> What would be desirable is to specify some constraints as absolutely
> essential and they should be checked first by ipopt. If they are not
> strictly satisfied, then the search step/direction has to be changed to get
> feasibilty of those constraints.
> 
> Thanks
> praveen
> 
> 
> 
> 
> _______________________________________________
> Ipopt mailing list
> Ipopt at list.coin-or.org
> http://list.coin-or.org/mailman/listinfo/ipopt



More information about the Ipopt mailing list