Preconditioning the Restarted and Shifted Block FOM Algorithm for Matrix Exponential Computation
The approximation of $e^{tA}B$ where $A$ is a large sparse matrix and $B$ a rectangular matrix is the key ingredient in many scientific and engineering computations. A powerful tool to manage the matrix exponential function is to resort to a suitable rational approximation such as the Carath$\acute{\rm e}$odory-Fej$\acute{\rm e}$r approximation, whose core reduces to solve shifted linear systems with multiple right-hand sides. The restarted and shifted block FOM algorithm is a commonly used technique for this problem. However, determining good preconditioners for shifted systems that preserve the original structure is a difficult task. In this paper, we propose a new preconditioner for the restarted and shifted block FOM algorithm. The key is that the absolute values of the poles of the Carath$\acute{\rm e}$odory-Fej$\acute{\rm e}$r approximation are medium sized and can be much smaller than the norm of the matrix in question. The advantages of the proposed strategy are that we can precondition all the shifted linear systems simultaneously, and preserve the original structure of the shifted linear systems after restarting. Theoretical results are provided to show the rationality of our preconditioning strategy. Applications of the new approach to Toeplitz matrix exponential problem are also discussed. Numerical experiments illustrate the superiority of the new algorithm over many state-of-the-art algorithms for matrix exponential.