Tangency and differentiation: Marginal functions

\(\Gamma\)-functionals \(\Gamma\)-limits generalized derivatives hypocompactoidness infinite dimensional spaces marginal functions subgradient inclusions tangency differentiation

Mathematics Subject Classification ID

Nonsmooth analysis (49J52) Optimality conditions for problems in abstract spaces (49K27) Differentiation theory (Gateaux, Fréchet, etc.) on manifolds (58C20) Sensitivity, stability, parametric optimization (90C31) Programming in abstract spaces (90C48)

Review text

The paper deals with upper and lower estimates (contingent and equi- tangent) of derivatives of the marginal functional \(\inf_ X f(y):=\inf_{x\in X} f(x,y)\) in terms of the corresponding derivatives of \(f\). Introducing simplified notions the calculus of \(\Gamma\)- functional and \(\Gamma\)-limits of De Giorgi is used. Results of Penot, Clarke, Gauvin, Gollan, Rockafellar and Hiriart-Urruty are extended to a broad class of topologies for problems in infinite dimensional spaces. Formerly used compactness properties are weakened applying the theory of compactoid and compact filters. The paper starts with the definition of generalized derivatives (chapter 1) and approximating cones (chapter 2) by applying \(\Gamma\)-functionals to the indicator functions of sets. Using the epigraph of a function connections between both the notions are obtained. In chapter 3 a short introduction to hypocompactoid filters is done. By using the above theories the main results are obtained in chapter 4 ``Bounds for contingent epi-derivatives of marginal functions e.g. \[ \begin{alignedat}{2} D_{(\alpha;-)}\inf_ X f(y)k &\leq \inf_{h\in X} D_{(\alpha;-)} f(x,y)(h,k)\quad && \text{ with } x\in\text{Min}_ X f(y)\text{ and } \alpha=-\text{ or } +,\\ D_{(-;-)}\inf_ X f(y)k &\geq \inf_{h\in X} D_{(-;-)} f(x,y)(h,k) \quad && \text{ under some hypocompactoidness} \end{alignedat} \] and in chapter 5 ``Upper bounds for equi-tangent epi-derivatives of marginal functions e.g. \[ D_{(+,\theta^ +,\sigma^ -)} \inf_ X f(y)k \leq \sup_{x\in\text{Min}_ X f(y)} \text{cl}_ \sigma \inf_{h\in X} D_{(+,\xi\times\theta^ +;\tau\times\sigma^ -)} f(x,y)(h,k), \] for \(f\) being l.s.c. and \(\text{Min}_ Xf\) being hypocompactoid at \(y\) from \(\theta_{epi(\inf_ X f)}\) to \(\xi\). In chapter 6 subgradient inclusions are considered. The reader should be familiar with the subjects introduced in the chapters 1-3. 43 references are given.

Recommendations

Cites work

Cited in

(6)

This page was built for publication: Tangency and differentiation: Marginal functions

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1174197)