Convex analysis and nonlinear optimization. Theory and examples. (Q817931): Difference between revisions

This book is divided into 11 chapters and provides a comprehensive presentation of the main features of convex analysis and nonlinear optimisation. Each result is sustained by a set of theorems, propositions and corollaries and includes rigorous proofs and clarifying discussions. They are complemented by a series of theoretical exercises. This is the second edition of [Convex analysis and nonlinear optimization. Theory and examples. CMS Books in Mathematics/Ouvrages de Mathématiques de la SMC. 3. New York, NY: Springer (2000; Zbl 0953.90001)]. In Chapter 1 the needed theoretical background on Euclidean spaces and symmetric matrices, for dealing with convexity, is given. The second chapter develops a similar study of optimality conditions, the so called theorems of the alternatives and the max-functions. They may be considered as an introductory part, not needed in some courses. The third chapter deals with functions, which do not have to be differentiable. An alternative approach to the usual Karush-Kuhn-Tucker (KKT) conditions is established. Different properties of convex functions are established within the Fenchel duality theoretical frame. The fourth chapter deals with convex analysis where the key relation for studying duality is studied. The convex function-Fenchel conjugate is considered. The fifth chapter presents some consequences of the advanced concepts discussed previously. Henceforth, polyhedral convex functions, eigenvalue functions, duality for linear semi-definite programming, as well as convex duality are discussed using these concepts. Chapter 6 is devoted to the study of nonsmooth optimisation where nonsmooth max-formulae, calculus and nonsmooth necessary conditions theorems are discussed. A similar study is made for regularity and regularity of convex functions, strict differentiability, unique Clarke's subgradient exact penalization, and other related problems. KKT theory is the theme of chapter 7, which introduces the necessary metric regularity concepts and results, allowing to revisit the KKT theorem establishing KKT and second order necessary and sufficient conditions. Chapter 8 deals with fixed points where Brouwers's theorem as well as Kakutani-Fan results are proved. Variational inequalities are studied extensively. Chapter 9 revisits nonsmooth problems establishing how to deal with them if more complicated structures are present. Then Rademacher's theorem is proved, when some changes are introduced into the original hypothesis, a proximal normal formula is introduced and Clarke's results are reworked conveniently and partly smooth sets are considered and the previously derived results are visited considering their role in sensitivity analysis of optimisation problems. Chapter eleven presents a list of the results (theorems, propositions, exercises of theoretical importance, etc.) clustered by sections and chapters of the book and the used notation is given. This book is warmly recommended for an advanced course in analysis for mathematicians or as a first graduate course for students involved with optimisation theory.