Convexity and well-posed problems (Q817460): Difference between revisions

This book studies convex functions in Banach spaces and the stable behavior under perturbations of the optimization problems associated to them. The adopted point of view is to link, in a fruitful and coherent way, convexity properties with several forms of well-posedness in optimization, in order to gain insight in both theories (and applications). An interesting feature of the book is the inclusion of some topics, like elements of game theory, hypertopologies and genericity of well-posedness, not usually found in textbooks devoted to convexity and optimization. The first three chapters describe the basic properties of convex sets and functions, lower semicontinuity, derivatives and the subdifferential. Chapter 4 is devoted to optimization issues: existence of minimizers, the Ekeland principle, properties of level sets. Chapter 5 deals with conjugacy, and chapter 6 with duality. Chapter 7 contains some elements of linear programming and zero sum game theory: some aspects of non cooperative games are presented in an appendix. Duality in linear programming is derived via game theory. In chapter 8 the basic elements of topologies and convergences on hyperspaces of closed subsets are considered and compared. Epiconvergence is introduced and corresponding stability results for optimization problems are proved. Chapter 9 deals with continuity properties of operations between functions. In chapter 10 the basic notions of Tikhonov and Levitin-Polyak well-posedness for optimization problems are presented, and stability results are described related to convergence properties of minimizers and the optimal values under perturbations. The particular case of the projection of a point on a closed convex set is considered in detail. The book ends with a chapter devoted to genericity and porosity properties of the set of well-posed optimization problems. The exposition is clear, the blend between standard basic notions and more advanced results is well balanced, leading the reader to understand recent developments without requiring a very specialized background. The reader finds several useful examples, comments and remarks scattered throughout, and over 120 exercices of varying level of difficulty. This book is suitable for graduate courses on convex optimization from a mathematical standpoint.