1: \begin{abstract}

2: For many applications involving a sequence of linear systems with slowly changing system matrices, subspace recycling,

3: which exploits relationships among systems and reuses search space information, 

4: can achieve huge gains in iterations across the total number of linear system solves in the sequence. 

5: However, for general (i.e., non-identity) shifted systems with the shift value varying over a wide range, the properties of the linear systems vary widely as well, which makes recycling less effective. If such a sequence of systems is embedded in a nonlinear iteration, the problem is compounded, and special approaches are needed to use recycling effectively.

6: 

7: In this paper, we develop new, more efficient, Krylov subspace recycling approaches for large-scale image reconstruction and restoration techniques that employ a nonlinear iteration to compute 

8: a suitable

9: regularization matrix. For each new regularization matrix, we need to solve regularized linear systems, $\MA + \c_\ell \ME_k$, for a sequence of regularization parameters, $\c_\ell$, to find the optimally regularized solution that, in turn, will be used to update the regularization matrix.   

10: 

11: In this paper, we analyze system and solution characteristics to choose appropriate techniques to solve each system rapidly.

12: Specifically, we use an inner-outer recycling approach

13: with a larger, principal recycle space for

14: each nonlinear step 

15: and smaller recycle spaces

16: for each shift.

17: We propose an efficient way to obtain good initial guesses from the principle recycle space and smaller shift-specific recycle spaces that lead to fast convergence.

18: Our method is substantially reduces the total number of matrix-vector products that would arise in a naive approach.   

19: Our approach is more generally applicable to sequences of shifted systems where the matrices in the sum are 

20: positive semi-definite.  

21: 

22: \end{abstract}

23: