Theory of duality in mathematical programming

From MaRDI portal





The book is intended as an introduction to the quantitative theory of mathematical programming in vector spaces. It consists of three chapters. The first chapter presents fundamentals of convex analysis, including separation theorems and conjugacy correspondence for convex functions (which are essentially the geometric and analytic equivalent forms of duality principles). Differentiability properties of convex functions are also treated but only in the Fréchet sense. Chapter 2 deals with the general duality theory of mathematical programming in its three main aspects: duality theorems, minimax theorems and saddle point theorems. Special attention is paid to the Lagrange form \(\psi (x,y)=\phi (x,y)- \alpha (x)-\beta (y)\), where \(\phi\) (x,y) is defined on \(K\times L\), \(\alpha\) (x) on K and \(\beta\) (y) on L. The last chapter is devoted to dual problems of mathematical programming, including linear programming, quadratic programming, convex programming and nonlinear programming. Duality theorems for these extremal problems are derived by specializing the general principles in Chapters 2 to suitable Lagrange forms. Also linear programming in Banach spaces and duality in transportation problems on networks are studied. In the whole the book is a lucid presentation of duality theory and can serve as a useful textbook for graduate or advanced undergraduate students. The specialist might wish to find in it more material on recent developments connected with nondifferentiable and nonconvex programming.











This page was built for publication: Theory of duality in mathematical programming

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3993055)