An algorithm for solving nonlinear least-squares problems with a new curvilinear search (Q911233)

This paper brings a modification of the algorithm by the author [Computing 39, 307-325 (1987; Zbl 0618.65050)] for solving the nonlinear least squares problem \(\min \{\| F(x)\|^ 2;\quad x\in R^ n\}\) which uses a cheaper curvilinear search instead of the former trust- region strategy. It is proved that the modified algorithm has global convergence properties.