Newton-like methods in generalized Banach spaces (Q1343482)

Let \(F\) be a nonlinear operator defined on a convex subset of a generalized Banach space \(E\) and with values in \(E\). The approximations of zeros of \(F\) by a Newton-like method of the form \(x_{n+1}= x_ n+ y_ n\), \(A_ n y_ n+ F(x_ n)= 0\), \(n\geq 0\), \(x_ 0\in D\), where \(A_ n\) is a conscious approximation to the Fréchet derivative \(F'(x_ n)\) of \(F\) evaluated at \(x_ n\), are considered. If some conditions for the space \(E\) and the operator \(F\) are fulfilled, the convergence for this method is proved and error estimates are provided. The idea of a generalized norm is used which is defined to be a map from a linear space into a partially ordered Banach space. The obtained convergence results and error estimates are improved and compared with those from the usual norm theory.