Greedy expansions in convex optimization (Q483395): Difference between revisions

The paper is a continuation of the author's previous paper [``Greedy approximation in convex optimization'', Preprint (2012), \url{arXiv;1206.0393}] devoted to the problem of adjusting greedy-type algorithms for finding sparse solutions of convex optimization problems. The author continues to study sparse approximate solutions to convex optimization problems and concentrates the attention on greedy algorithms that provide so-called expansions. The expansions can be characterized as follows: the approximating expression at an iteration step \(m\) is equal to the sum of the approximating expression at the preceding iteration step \(m-1\) and one element from the dictionary with an appropriate coefficient. Convergence of the proposed method as well as the corresponding rate of convergence are proved.