A numerical study of multiplier methods for constrained parameter optimization

Abstract
Several algorithms based on the multiplier or augmented penalty function method for minimizing a function subject to equality constraints are surveyed and compared numerically. These methods employ an approximation of the Lagrange multiplier to increase efficiency. A revised method for choosing the multiplier is presented and compared numerically with the other algorithms. One scheme for automatically choosing the penalty constant associated with the algorithms is also presented.

This publication has 6 references indexed in Scilit: