[Ipopt] How to make Ipopt faster when all variables have 1e-4 precision?

[Greg Horn]

try setting the option "tol" to a larger value
http://www.coin-or.org/Ipopt/documentation/node41.html#SECTION000112010000000000000

I'm not sure if multiplying by 1e4 has any effect, perhaps someone can
comment on if ipopt is using scaled primal/dual tolerances or unscaled

What do you mean "say that they are integers"? I'm not aware of this
functionality

On Fri, Apr 3, 2015 at 9:53 PM Илья Палачев <palachev.ilya at yandex.ru> wrote:

> Hi,
>
> I'm optimizing some quadratic programming problem using Ipopt. It works
> fast :)
> It's known that all data is obtained from physical measurements that have
> precision 1e-4.
> Is it possible to set some parameter in Ipopt to make it working faster
> according to this information?
>
> Maybe I should multiply all values by 10000 and say that they are
> integers? Will this help to make it faster?
>
> --
> Best regards,
> Ilya Palachev
> _______________________________________________
> Ipopt mailing list
> Ipopt at list.coin-or.org
> http://list.coin-or.org/mailman/listinfo/ipopt
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://list.coin-or.org/pipermail/ipopt/attachments/20150403/db283660/attachment.html>