On second-order subdifferentials and their applications (Q2784406)
From MaRDI portal
| This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use this page instead for the normal view: On second-order subdifferentials and their applications |
scientific article; zbMATH DE number 1732297
| Language | Label | Description | Also known as |
|---|---|---|---|
| English | On second-order subdifferentials and their applications |
scientific article; zbMATH DE number 1732297 |
Statements
23 April 2002
0 references
variational analysis
0 references
Lipschitzian stability in optimization
0 references
second-order subdifferentials
0 references
calculus rules
0 references
piecewise smooth functions
0 references
variational and hemivariational inequalities
0 references
mechanical equilibrium
0 references
contact problems
0 references
shape design problems
0 references
nonmonotone friction
0 references
On second-order subdifferentials and their applications (English)
0 references
The authors study second-order subdifferentials of extended-real-valued functions. The starting point is the second-order subdifferential introduced by the first author as the coderivative of the first-order subdifferentiable mapping. Beside this, a semiconvex version is introduced as the coderivative of the convexified first-order subdifferential. Several sum and chain rules are established for both subdifferentials. For a class of separable piecewise \(C^2\) functions, the subdifferentials are efficiently computed.NEWLINENEWLINENEWLINEThe theory is then applied to the stability analysis of parametric variational systems of the form NEWLINE\[NEWLINE0\in f(x, y)+ Q(x, y),NEWLINE\]NEWLINE where \(f\) is a continuously differentiable vector function and \(Q\) is a multifunction. In this connection, \(y\in \mathbb{R}^m\) is to be interpreted as the decision variable and \(x\in \mathbb{R}^n\) is a perturbation vector. Conditions are established ensuring that the solution map \(S\) defined by NEWLINE\[NEWLINES(x):= \{y\in \mathbb{R}^m\mid 0\in f(x, y)+ Q(x,y)\}NEWLINE\]NEWLINE is pseudo-Lipschitzian. Finally, the results are applied to contact problems as well as shape design problems with nonmonotone friction.
0 references