No label defined (Q4255463)

This book is an elementary introduction to optimization theory and algorithms. It assumes only a limited knowledge in calculus in finite-dimensional spaces and ordinary differential equations. Chapter 1 starts with simple examples in one and two dimensions, and then introduces the notion of convexity as well as several examples (some of them being linked to the theory of dynamical systems rather than to optimization itself) such as Chebyshev polynomials. The properties of the latter, however, are given without proof. Chapter 2 is devoted to the side conditions, the Karush-Kuhn-Tucker optimality system, whose necessity is obtained under the condition of linear independence of gradients of active constraints; the proof is based on the Farkas lemma. Chapter 3 deals with linear and quadratic optimization, and discusses duality for nonlinear, possibly convex, programs. Finally, Chapter 4 presents the Fenchel-Rockafellar conjugate duality scheme. Applications to the Fermat-Weber problem and some generalizations of it are discussed. The material is classical and is not related to the improvements of the field during the last decades. This book may be useful for a first course in nonlinear programming with students having little mathematical background.