On the Linear Convergence of Descent Methods for Convex Essentially Smooth Minimization (Q3988952)
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 33610
| Language | Label | Description | Also known as |
|---|---|---|---|
| default for all languages | No label defined |
||
| English | On the Linear Convergence of Descent Methods for Convex Essentially Smooth Minimization |
scientific article; zbMATH DE number 33610 |
Statements
On the Linear Convergence of Descent Methods for Convex Essentially Smooth Minimization (English)
0 references
28 June 1992
0 references
strictly convex essentially smooth function
0 references
gradient projection
0 references
matrix splitting
0 references
0.80980575
0 references
0.8073929
0 references
0.8058528
0 references
0.79899997
0 references
0.7979803
0 references
0.7940488
0 references
0.7931467
0 references
The authors consider the following minimization problem: (P) \(\min f(x)\), s.t. \(x\in X\), where \(X\) is a polyhedral set in \(R^n\) and \(f(x)=g(Ex)+\langle q,x\rangle\) with \(g: R^n\to (-\infty,\infty]\), \(E\) is a \(m\times n\) matrix and \(q\) is a vector in \(R^n\). The following assumptions are made for the function \(g\): (a) the effective domain of \(g\), denoted by \(C_g\) is non-empty and open; (b) \(g\) is strictly convex twice continuously differentiable on \(C_g\); (c) \(g(t)=\infty\) as \(t\) approaches any boundary point of \(C_g\). In the terminology of \textit{R. T. Rockafeller} [see ``Convex analysis.'' Princeton, N. J.: Princeton University Press (1970; Zbl 0193.18401)] \(g\) is said to be a strictly convex essentially smooth function. Under additional assumptions on \(f\) and \(X\), general conditions under which a sequence of feasible points converges at least linearly to an optimal solution of (P) are introduced. Further, it is shown that these conditions are satisfied by the gradient projection algorithm and by the matrix splitting algorithm using regular splitting. Possible extensions of the convergence conditions are discussed.
0 references