cyipopt.minimize_ipopt now supports the definition of constraints with Bounds, LinearConstraint, and NonlinearConstraint objects like scipy.optimize.minimize. Unlike scipy.optimize.minimize with method='trust-constr', however, the keep_feasible option:
of these constraints is ignored.
Does Ipopt have any way to ensure that (inequality) constraints are never violated (as long as the guess is feasible)?
If not, this issue can be closed, IMO. If so, however, this is an enhancement request for minimize_ipopt to respect keep_feasible.
cyipopt.minimize_ipopt
now supports the definition of constraints withBounds
,LinearConstraint
, andNonlinearConstraint
objects likescipy.optimize.minimize
. Unlikescipy.optimize.minimize
withmethod='trust-constr'
, however, thekeep_feasible
option:of these constraints is ignored. Does Ipopt have any way to ensure that (inequality) constraints are never violated (as long as the guess is feasible)? If not, this issue can be closed, IMO. If so, however, this is an enhancement request for
minimize_ipopt
to respectkeep_feasible
.