# On Cone Approximations and Generalized Directional Derivatives

• K.-H. Elster
• J. Thierfelder
Chapter
Part of the Ettore Majorana International Science Series book series (EMISS, volume 43)

## Abstract

By the classical notion of differentiability necessary optimality conditions for smooth problems can be formulated as stationarity conditions. But often optimization problems connected with real life problems have a “nondifferentiable” structure. Hence it is useful to develop appropriate notions of generalized differentiability in order to obtain analogous optimality conditions in nondifferentiable case, too. The concepts of the directional derivative and the subdifferential of a convex function were used with advantage for treating convex optimization problems. Since more than ten years much effort was made to establish similar concepts in the nonconvex case. Several trends can be distinguished, for example modifications of the directional derivative given by Clarke and many other authors. In accordance with such investigations, in Refs.[4, 5, 16] an axiomatic approach was given for constructing generalized directional derivatives of arbitrary extended real-valued functionals. The basic idea is the fact that the epigraphs of the different directional derivatives of a function $$f:X \to \overline R$$ (X a linear topological space) can be considered as cone approximations of the epigraph epi f of f. Conversely, so called K-directional derivatives can be introduced in such a way that epi f is locally approximated by a cone K(epi f, (x, f (x))) at the point (x, f (x)), where a positively homogeneous functional $${f^K}(x, \cdot ):X \to \overline R$$ according
$${f^K}(x,y):\inf \left\{ {\xi \in R|(y,\xi ) \in K(epif,(x)))} \right\}$$
is determined uniquely by that cone.

## Keywords

Convex Cone Directional Derivative Generalize Derivative Tangent Cone Nonsmooth Optimization
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

## References

1. [1]
F.H. Clarke. “Generalized gradients and applications”. Trans. A.M.S., 205 (1975), pp. 247–262.
2. [2]
V.F. Demyanov, A.M. Rubinov. “On quasidifferentiable functionals”. Soviet. Math. Dokl., 21 (1980) 1, pp. 14–17.Google Scholar
3. [3]
V.F. Demyanov, D. Pallaschke (eds). “Nondifferentiable optimization: Motivations and applications”. Proc., Sopron, Hungary, 1984. Lecture Notes in Economics and Mathematical Systems, Vol. 255, Springer-Verlag, Berlin, Heidelberg, New York, Tokyo (1985).Google Scholar
4. [4]
K.-H. Elster, J. Thierfelder. “Abstract cone approximations and generalized differentiability in nonsmooth optimization”. Optimization, Akademie-Verlag, 19 (1988) 3, pp. 315–341.
5. [5]
K.-H. Elster, J. Thierfelder. “A general concept of cone approximations in non-differentiable optimization”. In: V.F. Demyanov, D. Pallaschke (eds). “Non-differentiable Optimization: motivations and applications”. Lecture Notes in Economics and Mathematical Systems, Vol. 255, Springer-Verlag (1985), pp. 170–189.
6. [6]
K.-H. Elster, J. Thierfelder. “On cone approximations of sets”. In: P. Kenderov, (ed.). Mathematical Methods in Operations Research. Lectures presented at the Summer School on Operations Research, Primorsko (1984), pp. 33–59.Google Scholar
7. [7]
K.-H. Elster, J. Thierfelder. “Generalized notions of directional derivatives”. Research Paper No. 155, Dept. of Mathematics (Applied Section, Optimization ), Univ. of Pisa (1988), pp. 1–54.Google Scholar
8. [8]
J.B. Hirriart-Urruty. “Tangent cones, generalized gradients and mathematical programming in Banach-spaces”. Math. Oper. Res. 4 (1979) 1, pp. 79–97.
9. [9]
A.D. Ioffe. “On the theory of subdifferential”. In: J.B. Hirriat-Urruty (ed.). “ Fermat-Days 85: mathematics for optimization”, North-Holland Mathematics Studies, Vol, 129, North-Holland, (1986), pp. 183–200.
10. [10]
P. Michel, J.P. Penot. “Calcul sous-différential pour des fonctions lipschitziennes et non lipschitziennes”. C.R. Acad. Sc. Paris, t. 298, 1 (1984) 12, pp. 269–272.Google Scholar
11. [11]
J.P. Penot. “Variations on the theme of nonsmooth analysis: another subdifferential”. In: V.F. Demyanov, D. Pallaschke (eds). “Nondifferentiable optimization: motivations and applications”. Lecture Notes in Economics and Mathematical Systems, Vol. 255, Springer-Verlag (1985), pp. 41–55.
12. [12]
R.T. Rockafellar. “Convex analysis”. Princeton Univ. Press., Princeton (1978).Google Scholar
13. [13]
R.T. Rockafellar. “Directionally lipschitzian functions and subdifferential calculus”. Proc. of the London Math. Soc., 39 (1979), pp. 331–355.
14. [14]
R.T. Rockafellar. “Generalized directional derivatives and subgradients of non-convex functions”. Canadian J. Math., 32 (1980) 2, pp. 257–280.
15. [15]
R.T. Rockafellar. “The theory of subgradients and its applications to problems of optimization”. Heldermann-Verlag, Berlin (1981).
16. [16]
J. Thierfelder. “Beiträge zur Theorie der nichtglatten Optimierung”. Diss. (A ), Technical University of Ilmenau (1984).Google Scholar
17. [17]
C. Ursescu. “Tangent set’s calculus and necessary conditions for extremality”. SIAM J. Control Optim., 20 (1982) 4, pp. 563–574.
18. [18]
M. Vlach. “Approximation operators in optimization theory”. Zeitschrift fzir Operations Research. Theory 25 (1981) 1, pp. 15–24.
19. [19]
M. Vlach. “ Closures and neighbourhoods properties included by tangential approximations”. Research Paper No. 129, University of Bielefeld (1983).Google Scholar
20. [20]
D.E. Ward. “Isotone tangent cones and nonsmooth optimization”. Optimization, Akademie-Verlag, 18 (1987) 6, pp. 769–783.Google Scholar