Questions about the convergence of infinite sequences or series of matrices can be treated with vector norms just as one treats infinite se quences or series of vector.
Ex. Let { A } < M be a given infinite sequence of matrices. Show that the series M = 0 converges to some matrix in M if there is a vector norm || on M such that the numerical series M= o is convergent (or even if its partial sums are bounded ). Hint: Show that the partial sums form a Cauchy sequence.
One special case for matrices that does not arise in the study of infinite series of vectors is the case of power series of matrices. But because of the submultiplicative property of matrix norms, it is easy to give a simple sufficient condition for convergence of matrix power series.
5.6.15 Theorem. If A < M , then the series M… converges if there is a matrix norm || on M such that the numerical series M = || converges, or even if the partial sums of this series are bounded.
Ex. Prove (5.6.15)
Ex. Show by example that it is possible that the series M =0 converges and the series M = || diverges. This is analogous to conditional convergence (convergence but not absolute convergence) for numerical series.
Ex. Let the function f(z) be defined by the power series f(z) = M , which has radius of convergence R > 0 , and let || be a matrix norm on M . Show that f(A) =_ M is well defined for all A < M such that || . More generally, show that f(A) is well defined for all A