Quasi-iteration methods of Chebyshev type for the approximate solution of operator equations (Q1904093)

The authors describe a general method for the approximate solution of the operator equation \(Ax= f\), where \(A\) is a bounded linear invertible operator in a Banach space \(X\). A Chebyshev quasi-iteration method is defined and basic convergence results as well as some error estimates are given. The method builds on the construction of a sequence of polynomials of \(A\) which approximates the operator \(A^{- 1}\) in the norm of \({\mathcal L}(X)\). The authors also discuss an effective algorithm for constructing a sequence of polynomials of \(A\) which tends to \(A^{- 1}\) with sharp Lyapunov exponent \(\exp[- g(0, \infty)]\). Some examples are given.