On a globally convergent trust region algorithm with infeasibility control for equality constrained optimization (Q2634338)

A previously known algorithm in solving the nonlinear equality constrained optimization problem uses a technique in controlling infeasibility via a set of constraint violation of some previous iterations. The global convergence of that algorithm was established by requiring a strong assumption on a double trust regions strategy used for the step computations. The author replaces the double trust regions strategy by a single trust region strategy, and the global convergence to first order critical points is established under mild assumptions. Some numerical results are shown to demonstrate the proposed trust region infeasibility control algorithm.