Lagrangian transformation and interior ellipsoid methods in convex optimization (Q2342139): Difference between revisions

The author first considers smooth convex optimization problems with Slater condition and describes the so-called Lagrangian transformation, which serves for creation of various augmented Lagrangians. Then he describes a class of multiplier methods, shows its equivalence to the interior ellipsoid method for the dual problem, proves its convergence and some rates of convergence. It is also shown that application of this method to the linear programming problem gives the Dikin affine scaling method with respect to the dual problem.