Subgradients and Variational Analysis
نویسنده
چکیده
The study of problems of maximization or minimization subject to constraints has been a fertile field for the development of mathematical analysis from classical times. In recent decades, convexity has come forward as an important tool, and the geometry of convexity has been translated into notions of directional derivatives and subgradients of functions that may not be differentiable in the ordinary sense. Now there has emerged a form of analysis able to deal robustly, even in the absence of convexity, with the phenomena of nonsmoothness that arise in variational problems. The Need for Variational Analysis Differential calculus has been so successful in treating a variety of physical phenomena that mathematics has long relied on differentiable functions as the main tools of analysis. The domain of such a function, in a finite-dimensional setting, is typically an open subset of IR or of a differentiable manifold such as might be defined by a system of equations in IR and coordinatized locally by IR for some d < n. Many of the systems and phenomena that have come under mathematical scrutiny in recent decades, however, are of interest especially at their frontiers of feasibility. They involve functions and mappings whose domains may be closed sets with very complicated boundaries, expressed often by numerous inequalities as well as equations. Behavior around boundary points of these domains is seen as crucial, but it cannot well be investigated without a development of ideas beyond the customary framework. A major source of this trend lies in the fact that mathematical models are being used more and more for prescriptive as well as merely descriptive purposes. Nowadays one seeks not only to describe what happens in the world but to influence or improve the way it happens. New subjects have been created like optimization theory, control theory, and viability theory, which are heavily involved with finding the extremes of what may be possible under given circumstances. This has stemmed from an increasing preoccupation ∗ The written text of this lecture has been adapted from the introduction of the forthcoming book Variational Analysis by R. T. Rockafellar and R. J-B Wets.
منابع مشابه
Variational Analysis of the Spectral Abscissa at a Matrix with a Nongeneric Multiple Eigenvalue
The spectral abscissa is a fundamental map from the set of complex matrices to the real numbers. Denoted α and defined as the maximum of the real parts of the eigenvalues of a matrix X, it has many applications in stability analysis of dynamical systems. The function α is nonconvex and is non-Lipschitz near matrices with multiple eigenvalues. Variational analysis of this function was presented ...
متن کاملNecessary Conditions in Nonsmooth Minimization Via Lower and Upper Subgradients
The paper concerns first-order necessary optimality conditions for problems of minimizing nonsmooth functions under various constraints in infinite-dimensional spaces. Based on advanced tools of variational analysis and generalized differential calculus, we derive general results of two independent types called lower subdifferential and upper subdifferential optimality conditions. The former on...
متن کاملVariational Analysis of the Abscissa Mapping for Polynomials∗
The abscissa mapping on the affine variety Mn of monic polynomials of degree n is the mapping that takes a monic polynomial to the maximum of the real parts of its roots. This mapping plays a central role in the stability theory of matrices and dynamical systems. It is well known that the abscissa mapping is continuous on Mn, but not Lipschitz continuous. Furthermore, its natural extension to t...
متن کاملSubgradients of Distance Functions at Out-of-Set Points
This paper deals with the classical distance function to closed sets and its extension to the case of set-valued mappings. It has been well recognized that the distance functions play a crucial role in many aspects of variational analysis, optimization, and their applications. One of the most remarkable properties of even the classical distance function is its intrinsic nonsmoothness, which req...
متن کاملAn Approximate Bundle Method for Solving Variational Inequalities
We present an approximate algorithm for solving generalized variational inequality problem (P) by combining the bundle technique with the auxiliary problem method. For subproblems we construct a new lower approximation to the involved function φ, as in the bundle method for nonsmooth optimization, which only requires the approximate function values and approximate subgradients instead of the ex...
متن کامل