[Coin-ipopt] First order KKT

[Rohallah Tavakoli]

Dear All

- Has IPOPT routine that is based on first order KKT optimality condition
that essentially doesn't need Hessian or its approximation with
quasi-newton?

- In engineering application, experience shows that first order methods are
reasonble and sometimes second order methods do'nt give considrable
improvement in expence of their cost (e.g. topology optimization), so my
question is that for such problem is it logical that when we have only first
derivatives, we use quasi-newton method instead of using pure frist order
methods?

Regards, ... rt.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://list.coin-or.org/pipermail/ipopt/attachments/20070224/0a65d539/attachment.html