No label defined (Q4033228)
From MaRDI portal
!
This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:
scientific article; zbMATH DE number 165426
| Language | Label | Description | Also known as |
|---|---|---|---|
| default for all languages | No label defined |
||
| English | No label defined |
scientific article; zbMATH DE number 165426 |
Statements
16 May 1993
0 references
gradient methods
0 references
method of the heavy ball
0 references
unconstrained minimization
0 references
convex, differentiable function
0 references
0.824151337146759
0 references
0.8094369769096375
0 references
0.7942042350769043
0 references
0.7919168472290039
0 references
In order to solve the unconstrained minimization problem associated to a convex, differentiable function \(f\) with \(f'\) being \(L\)-Lipschitz- continuous on \(R^ m\), the author considers the following iterative scheme (the method of the ``heavy ball'' with constant coefficients): \((*)\) \(x_{n+1}=x_ n-\alpha f'(x_ n)+\beta(x_ n-x_{n-1})\); \(x_{-1}=x_ 0\). The main result of the paper shows that if \(0\leq\beta<1\) and \(0<\alpha<2(1-\beta)^ 2/((1+\beta)L)\) and if \(\text{Argmin}\{f(x): x\in R^ m\}\neq\emptyset\), then the convergence of the sequence generated by \((*)\) to an optimal solution is guaranteed. Numerical experiments are presented in the last part of the paper.
0 references