Approximate iterations for structured matrices (Q929359)

The paper deals with the evaluation of a matrix-valued function \(f(A)\), where \(A\) is a large matrix with a special structure. The authors present a transformation of an iterative fixed-point like process, under certain general assumptions, into another process which preserves the convergence rate and benefits from the underlying structure. Applications are also given for \(f(A) = A^{-1}\) and \(f(A) = \sqrt{A}\).