Optim.jl icon indicating copy to clipboard operation
Optim.jl copied to clipboard

IPNewton() and Constrained Nonlinear Optimization

Open alh99 opened this issue 5 years ago • 2 comments

Hi,

I was wondering if IPNewton() could be used in conjunction with equality constraints. The tutorial was quite nice, but noted that this had not yet been implemented.

Something else I noticed in using IPNewton() for a multivariable optimization problem (with second order Hessian information) was that the returned solution didn't satisfy the inequality constraints that I imposed on the problem (although the initial point did). It seems like the algorithm shouldn't return "success" when the solution doesn't satisfy the original inequality constraints.

alh99 avatar May 01 '20 07:05 alh99

Thanks for the comments. Certainly, that part of the code can still use a lot of improvements. The last part is news to me, and I would love if you could provide a small example if at all possible.

pkofod avatar May 04 '20 11:05 pkofod

if IPNewton() could be used in conjunction with equality constraints

Does it work if you set the lower and upper bounds to the same value?

timholy avatar May 05 '20 12:05 timholy