Convex and quadratic programming
This paper deals primarily with the theory of convex and quadratic programming problems and discusses algorithmic procedures used in solving such problems.
A general overview of mathematical programming is given in Chapter I. It includes formulations of linear, convex, and quadratic problems and a brief history of the development of mathematical programming.
Chapter II deals with the general theory of convex programming. An equivalent problem, the saddle point problem is examined and then used to develop the Kuhn-Tucker conditions. This chapter includes a formulation of the dual convex problem.
Quadratic programming theory is the topic of discussion in Chapter III. After a brief introduction dealing with quadratic forms, the Kuhn-Tucker conditions for quadratic programming are developed, and the dual quadratic problem is formulated.
Chapter IV is a discussion of a convex programming algorithm developed by G. Zoutendijk. It includes computational details of the procedure and a discussion of a criterion for identifying the optimal solution.
Several procedures for solving quadratic problems are discussed briefly in Chapter V, which is mainly a presentation of Wolfe's algorithm, an indirect method of solving the quadratic problem by looking at an equivalent linear problem.
Thesis75.L397.pdf
2.01 MB
Adobe PDF
79b5328d0d66716fa67b95ef4fafe7b2