Analysis of some Krylov subspace approximations to the matrix exponential operator. (English) Zbl 0749.65030
An algorithm is described which computes the exponential function of a matrix multiplying a vector by applying the Arnoldi algorithm to the matrix and computing the exponential function of the resulting small Hessenberg matrix. Error estimates based on the theory for Lagrange interpolation are given and a numerical test is reported.
Reviewer: A.Ruhe (Göteborg)
MSC:
65F30 | Other matrix algorithms (MSC2010) |
65F10 | Iterative numerical methods for linear systems |