Constrained optimization and image space analysis. Vol. 1: Separation of sets and optimality conditions (Q2487865)

As the author claims, ``the book is devoted to the basic theory of smooth and nonsmooth constrained extremum problems and variational inequalities. This approach is based on the analysis in the image space, namely the space where the images of the involved functions run. The book is divided into five chapters: Ch.~1: Introduction, Ch.~2: Elements of convex analysis and separation, Ch.~3: Introduction to image space analysis, Ch.~4: Alternative and separation, Ch.~5: Optimality conditions, Preliminary results; it ends with a glossary of notations and a subject index. Every chapter ends with a quite long section of interesting comments and a list of references. The most part of the notions and results are given in finite dimensional spaces, sometimes being given hints for the case of infinite dimensional Banach spaces. In the first chapter a general constrained extremum problem is presented and several examples of problems which enter in this general framework. In the second chapter the usual notions and results concerning convex sets and functions are given. Among such results we mention the Carathéodory theorem concerning the convex hull of a set, separation theorems for convex sets, existence and uniqueness of the best approximation by elements of a closed convex subset of \({\mathbb R}^{n},\) properties of the polar sets, continuity and sub\-differentiability properties of convex functions. Some extensions of convexity are discussed, too. In the first section of Chapter 3 the notion of \(G\)-differentiability is introduced, \(G\) being a subset of the class of positive homogeneous functions defined on a certain cone. Some calculus is developed, as well as relationships with other directional derivatives are studied. Many examples are considered. In fact the presence of many examples is an important feature of this book. In the second section the image problem is formulated, while in the third one several results on stationary points are given. In the fourth chapter a general setting for alternative theorems is presented; this is based by the consideration of weak separation functions and strong separation functions, respectively. Using such a setting several more or less classic alternative results are obtained. The last chapter is dedicated to optimality conditions. It is shown that weak separation functions are intimately related to sufficient optimality conditions, while strong separation functions are related to necessary optimality conditions. The mathematical style of this book is not one I am used to; this is probably the reason for the difficulties I met in reading it. The fact that the book is addressed, besides mathematicians, to mechanical engineers could be an explanation. Unfortunately, there are some misprints and omissions, which ask readers' care; probably the author will mention them as errata.