Greedy strategies for convex optimization (Q2363679): Difference between revisions

The paper considers two greedy procedures for solving the minimization of a convex function on a given Hilbert space. The first one is a convex minimization analogue of the orthogonal matching pursuit algorithm. The second is the weak Chebyshev greedy algorithm. The authors establish a priori convergence rates for both algorithms under certain conditions of uniform smoothness and uniform convexity.