Selection strategies in projection methods for convex minimization problems
Andrzej Cegielski, Robert Dylewski · Discussiones Mathematicae Differential Inclusions Control and Optimization · 2002
We propose new projection method for nonsmooth convex mini-mization problems. We present some method of subgradient selection, which is based on the so called residual selection model and is a gen-eralization of the so called obtuse cone model. We also present nu-merical results for some test problems and compare these results with some other convex nonsmooth minimization methods. The numerical results show that the presented selection strategies ensure long steps and lead to an essential acceleration of the convergence of projection methods.