Abstract :
A method is presented which allows computation of eATby a truncated series method. The method given has low incremental computational cost if the series approximation to eATis of much higher degree than the degree of the characteristic polynomial of A.
