On the convergence of descent methods
For many iterative processes of descent methods it is not possible to obtain an estimate of the convergence to solution point \(x_*\) at a linear rate \[ \rho{(x^k,x_*)}\leq{q^k\rho{(x^0,x_*)}} \forall{k}\in\mathbb{N}. \] At the same time, one is often able to obtain the required estimates in terms of a certain function that is more or less naturally defined. This approach, which may be classified as the second Lyapunov method, is used in this paper. Applications to the numerical solution of nonlinear operator equations with approximate initial data are discussed.
- Convergence properties of descent methods for unconstrained minization
- Convergence of descent method without line search
- A study on the rate of convergence of descent methods
- scientific article; zbMATH DE number 4106165
- Some notes about the iterative gradient and quasi-gradient methods of approximate solution for the nonlinear operator equations
This page was built for publication: On the convergence of descent methods
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1571169)
