Abstract
Consider the problem of minimizing a real functionalf. A Newton-like method requires first an approximationD(d) off(x + d)−f(x) at the current iteratex, valid for smalld to an order higher than 1, and consists in minimizingD. In this paper, we will introduce a new concept of such an approximation for convex functions without differentiability assumptions. The connections with the classical concept based on the Taylor development of a differentiablef are exhibited. The material is used to study a conceptual (nonimplementable) algorithm. Some hints are given which could lead to an implementable version.
Similar content being viewed by others
References
Ben-Tal A, Zowe J (1982) Necessary and sufficient optimality conditions for a class of nonsmooth minimization problems. Math Programming 24:70–91
Hiriart-Urruty JB (to appear) Limiting behaviour of the approximate first order and second order directional derivatives for a convex function. In: Nonlinear analysis: Theory, methods, and applications
Holmes RB (1975) Geometric functional analysis and its applications. Springer-Verlag, New York-Heidelberg-Berlin
Laurent P-J (1972) Approximation et optimisation. Hermann, Paris
Lemarechal C (1978) Bundle methods in nonsmooth optimisation. In: Lemaréchal C, Mifflin (eds) Nonsmooth optimization. Pergamon Press
Moreau JJ (1966–67) Fonctionelles convexes (Séminaires sur les équations aux dérivées partielles II). Collège de France, Paris
Pchenichnyi B (1970) Necessary conditions for an extremum. Marcel Dekker, New York
Rockafellar, RT (1970) Convex analysis. Princeton University Press, Princeton NJ
Author information
Authors and Affiliations
Additional information
Communicated by J. Stoer
Rights and permissions
About this article
Cite this article
Lemaréchal, C., Zowe, J. Some remarks on the construction of higher order algorithms in convex optimization. Appl Math Optim 10, 51–68 (1983). https://doi.org/10.1007/BF01448379
Accepted:
Issue Date:
DOI: https://doi.org/10.1007/BF01448379