Dual coordinate ascent methods for non-strictly convex minimization (Q2368079): Difference between revisions

A special problem of convex programming possessing a certain separable structure is considered. The suggested method for solving this problem appears to be a dual version of a block coordinate ascent method. The convergence for this dual method is proved. It is shown that Han's method for the minimization of a strictly convex quadratic function on the intersection of a collection of closed convex sets and the method of multipliers for the problem of convex programming with linear constraints are special cases of the suggested method.