Second-order characterizations of Lipschitz stability in nonlinear programming (Q1417347): Difference between revisions
From MaRDI portal
Set profile property. |
Set OpenAlex properties. |
||
Property / full work available at URL | |||
Property / full work available at URL: https://doi.org/10.1023/a:1023695203500 / rank | |||
Normal rank | |||
Property / OpenAlex ID | |||
Property / OpenAlex ID: W197072406 / rank | |||
Normal rank |
Latest revision as of 12:06, 30 July 2024
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Second-order characterizations of Lipschitz stability in nonlinear programming |
scientific article |
Statements
Second-order characterizations of Lipschitz stability in nonlinear programming (English)
0 references
4 January 2004
0 references
We study critical and stationary points of perturbed \(C^{1,1}\) optimization problems and characterize three types of Lipschitz stability in terms of contingent and Thibault derivatives for the mapping in question: strong regularity, upper Lipschitz behavior, and pseudo-Lipschitz behavior. It turns out that there is a full similarity with respect to stability conditions, quadratic approximation, and geometrical interpretation when comparing strong regularity and upper Lipschitz behavior of critical points. The only difference arises from the derivatives one has to apply, and these derivatives are determined in terms of the original data. In particular, the explicit formula for the contingent derivative of the stationary point map \(X\) under the Mangasarian-Fromowitz constraint qualification leads to criteria for both the pseudo- and upper Lipschitz behavior of \(X\). Our approach is essentially based on the analysis of the associated Kojima function, a product of a matrix function \(M(x)\) and a vector function containing terms of the form \((y^+,y^-)\), only. Applications to Lipschitz stability of optimal solutions are also discussed.
0 references