RESUMEN

Standard iterative methods for the solution of nonlinear constrained minimization problems must solve two or three linear systems of equations at each iteration. The theory developed by Dembo, Eisenstat, and Steihaug [SIAM J. Numer. Anal., 19 (1982), pp. 400–408] on inexact Newton methods is used to solve one of the linear systems inexactly. In particular, the case where the number of constraints is much smaller than the number of variables is studied. Conditions are given for a tradeoff between the speed of convergence (2-step q-superlinear) and the amount of computations per iteration.