1: \begin{abstract}

2: The conjugate gradient method (CG) is typically used with a preconditioner which improves efficiency and robustness of the method.  

3: Many preconditioners include parameters and a proper choice of a preconditioner and its parameters is often not a trivial task.  

4: Although many convergence estimates exist which can be used for

5: optimizing preconditioners, these estimates typically

6: hold for all initial guess vectors, in other words,

7: they reflect the worst convergence rate.  

8: To account for the mean convergence rate instead,

9: in this paper, we follow a  stochastic approach. 

10: It is based on trial runs with random initial guess vectors

11: and leads to a functional which can be used to monitor

12: convergence and to optimize preconditioner parameters in CG.

13: Presented numerical experiments show that optimization of this 

14: new functional with respect to preconditioner parameters usually 

15: yields a better parameter value than optimization of the functional based on the spectral condition number. 

16: \end{abstract}

17: