0

In the context of linear programming:

A Kuhn Tucker Point is a point that satisfies certain conditions, making it a candidate for being a minimizer of a function (global or local).

Now, given a local minimum, can we necessarily say it is a Kuhn Tucker Point?

  • a local minimum of what ? and you should look instead to quadratic programming it is easier to understand the KKT conditions for it – reuns Mar 06 '16 at 21:38
  • if you assume one of many standard forms for LP, then yes. The feasible set is already described linearly. You won't lose anything if you linearize the tangent cone. – user251257 Mar 07 '16 at 15:50
  • @user1952009: KKT conditions are more about constraints than the objective, aren't they? – user251257 Mar 07 '16 at 15:51
  • @user251257 : the KKT conditions are about the lagrangian. the LP is harder to understand than the QP because there are no local extremum for the objective function nor for the lagrangian, there are only saddle points for the lagrangian, – reuns Mar 07 '16 at 19:33

0 Answers0