Convergence Conditions for Nonlinear Programming Algorithms

Conditions which are necessary and sufficient for convergence of a nonlinear programming algorithm are stated. It is also shown that the convergence conditions can be easily applied to most programming algorithms. As examples, algorithms by Arrow, Hurwicz and Uzawa; Cauchy; Frank and Wolfe; and Newton-Raphson are proven to converge by direct application of the convergence conditions. Also the Topkis-Veinott convergence conditions for feasible direction algorithms are shown to be a special case of the conditions stated in this paper.

• 저널Management Science · 16(1) · 1–13
• 토픽Advanced Optimization Algorithms Research · Numerical Analysis
• DOI10.1287/mnsc.16.1.1
• 저자Willard I. Zangwill