Iterative processes, functional equations and the Newton method (Q1121643)

In this carefully written expository paper the author illustrates the idea of his method of nondiscrete induction [Linear Algebra Appl. 13, 223-238 (1976; Zbl 0323.46005)] for measuring convergence of iterative processes by applying it to Newton's method. The method of nondiscrete induction replaces the order of convergence (which is a number) with a new definition of the rate of convergence. It is now defined as a function which is the solution of a system of functional inequalities. These inequalities depend on the process and on the initial data. This rate of convergence is then used to obtain sharp error estimates and optimal convergence criteria. As described in the book by \textit{F.-A. Potra} and the author [Nondiscrete induction and iterative processes (1984; Zbl 0549.41001)], the method has been successfully applied to other iterative processes in numerical analysis. The rate of convergence of Newton's method was first established by the author [Numer. Math. 25, 279-285 (1976; Zbl 0304.65037)].