1: \begin{abstract}

2: We introduce a new iterative regularization method for solving inverse problems that  can be

3: written as systems of linear or non-linear  equations in Hilbert spaces.

4: The proposed averaged Kaczmarz (AVEK)  method  can be seen as a hybrid method  between the Landweber

5: and the Kaczmarz method. As the Kaczmarz method, the  proposed method only requires evaluation

6: of one direct and one adjoint sub-problem per iterative update.

7: On the other, similar to the Landweber iteration,  it uses an average over previous auxiliary  iterates   which

8: increases stability. We present a convergence analysis of the AVEK iteration. Further, detailed numerical

9: studies are presented for a tomographic image reconstruction problem, namely the limited data problem in

10: photoacoustic tomography.  Thereby, the AVEK  is compared with other iterative regularization methods

11: including standard Landweber and Kaczmarz iterations, as well as recently proposed  accelerated versions 

12: based on error minimizing relaxation strategies.

13: 

14: \bigskip\noindent\textbf{Keywords:}

15: Inverse problems, system of ill-posed equations, regularization method, Kaczmarz iteration,  ill-posed equation,  convergence analysis,  tomography, circular Radon transform.

16: 

17: \bigskip\noindent\textbf{AMS Subject Classification:}

18: 	65J20; 65J22; 45F05.

19: \end{abstract}

20: