On Convergence Properties of a Subgradient Method

From MaRDI portal





It is shown that the use of normalized support vectors to level sets in subgradient methods can be used to accelerate the convergence properties for the minimization of quasi-convex functions. As example, Hölder continuity of the function can be used to obtain convergence of the method, and the existence of both majorant and minorant functions leads to a linear convergence rate. To compute the step size in the algorithm (approximate) knowledge of the optimal function value is necessary.




Cited in
(38)








This page was built for publication: On Convergence Properties of a Subgradient Method

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4810572)