1: \begin{abstract}

2: Exponential integrators that use Krylov approximations of matrix functions 

3: have turned out to be efficient for the time-integration of certain ordinary 

4: differential equations (ODEs). This holds in particular

5: for linear homogeneous ODEs, where the exponential integrator

6: is equivalent to approximating the product of the matrix exponential and a vector.

7: In this paper, we consider  linear inhomogeneous ODEs, $y'(t)=Ay(t)+g(t)$,

8: where the function $g(t)$ is assumed to satisfy certain regularity

9: conditions. We derive an algorithm for this problem which is equivalent to

10: approximating the product of the matrix exponential and a vector using

11: Arnoldi's method. The construction is based on 

12: expressing the function $g(t)$ as a linear combination

13: of given basis functions $[\phi_i]_{i=0}^\infty$ 

14: with particular properties. The properties

15: are such that the inhomogeneous ODE can be restated

16: as an infinite-dimensional

17: linear homogeneous ODE. Moreover, the linear homogeneous

18: infinite-dimensional ODE has properties that directly allow us to extend

19: a Krylov method for finite-dimensional linear ODEs. Although the

20: construction is based on an infinite-dimensional

21: operator, the algorithm can be carried out with operations

22: involving matrices and vectors of finite size.

23: This type of

24: construction resembles

25: in many ways the infinite Arnoldi method for nonlinear eigenvalue

26: problems \cite{Jarlebring:2012:INFARNOLDI}. We prove

27: convergence of the algorithm under certain natural conditions, and 

28: illustrate properties of the algorithm with  examples stemming from the

29: discretization of partial differential equations.

30: \end{abstract}

31: