Inverse \(q\)-columns updating methods for solving nonlinear systems of equations (Q1410869)

The paper is concerned with a special class of quasi-Newton methods for solving large-scale systems of nonlinear equations, i.e., the inverse column updating methods (ICUM). The inverse \(q\)-columns updating methods for solving nonlinear equations of the form \(F(x)=0\) where \(F:\mathbb{R}^{n}\rightarrow \mathbb{R}\) is continuously differentiable, consist in constructing a sequence \(\{x^{n}\},\) defined by \[ x^{k+1}=x^{k}-H_{k}(x^{k}), \] where the matrix \(H_{k}\) is updated in such a way that \(H_{k+1}\) differs from \(H_{k}\) in exact \(q\) columns and that the last \(q\) secant equations are satisfied, that is, \[ H_{k+1}y^{j}=s_{j},j=k,k-1,\ldots,k-q+1, \] where \[ y^{j}=F(x^{j})-F(x^{j-1}),\text{ and }s_{j}=x^{j}-x^{j-1},\text{ } j=k,k+1,\ldots \] The value \(q=2\) is shown to be optimal for the ICUM. The numerical tests performed use both large-scale and short-scale problems.