Nondiscrete mathematical induction and iterative existence proofs

The author proves a simple general theorem about complete metric spaces which forms an abstract basis of existence theorem in functional analysis and numerical analysis. He shows that this theorem, the so called induction theorem, contains the classical fixed point theorem for contractive mappings as well as the closed graph theorem.He then explains the principles of application of the induction theorem, the method of nondiscrete mathematical induction which consists in reducing the given problem to a system of functional inequalities, to be satisfied by a certain function, called the rate of convergence. The fact that the rate of convergence is defined as a function and not a number makes it possible to obtain sharp estimates valid for the whole iterative process, not only asymptotically. The method of nondiscrete mathematical induction is then illustrated by means of the example of eigenvalues of almost decomposable matrices.