An efficient algorithm for nonlinear optimization based on sequential quadratic programming
by
M. Chuedoung
Curtin University of Technology
Coauthors: Y.H. Wu

Sequential quadratic programming generally has superlinear convergence and is considered to be one of the best methods for solving nonlinear constrained programming problems. However, in the traditional sequential quadratic programming method, a large matrix, the Hessian matrix, needs to be calculated and stored in each iteration, which limits its application to large-scale optimization problems.

In this paper, we present an efficient sequential quadratic programming algorithm in which, the quadratic programming subproblem can be implemented in such a way that the Hessian matrix is not needed explicitly. The algorithm is illustrated through an example.

Date received: August 12, 1999