A class of diagonal quasi-Newton methods for large-scale convex minimization (Q503542): Difference between revisions

This paper considers the diagonal quasi-Newton method with the Zhu-Nazareth-Wolkowicz updating formula in minimizing a smooth function. The author shows that the scaled method with the Wolfe's line search is globally and linearly convergent on uniformly convex objective functions. Some numerical results on solving a set of 32 test problems are presented.