[Ipopt] solution of infeasile problem

Andreas Waechter andreasw at watson.ibm.com
Wed Apr 22 19:14:35 EDT 2009


Hi,

> Is IPOPT suitable to solve a nonlinear program (constraints are also
> nonlinear) which is infeasible,

By definition, a problem that is infeasible does not have a solution ;-)

> i.e., feasible set is empty and goal is
> to minimize the objective function simultaneously having the minimum
> infeasibility?

No, if you give an infeasible problem to Ipopt, it will eventually enter 
the restoration phase, and this will converge to a point that minimizes 
the 1-norm of the constraint violation (possibly scaled), without 
considering the objective function.

Andreas



More information about the Ipopt mailing list