A Successive Linear Programming Method and Its Convergence on Nonlinear Problems.
TEXAS UNIV AT AUSTIN CENTER FOR CYBERNETIC STUDIES
Pagination or Media Count:
Since Griffith and Stewart firstly proposed as successive linear programming method for solving general nonlinear programming problems, such methods have been widely used in practice because of their ease of implementation and their ability to deal with large scale problems. However, neither the original version, nor a more recent one contain convergence proofs possibly because of non-robustness of their algorithms. Using exact penalty functions and Levenberg Marquardtlike steps, an improved algorithm has been recently devised. In this paper, we give this modified SLP method a theoretical analysis and covergence proof, and thereby provide a sound basis for it. Author
- Operations Research