1: \begin{abstract}

2: The convergence to the self-consistency in the dynamical-mean-field-theory

3: (DMFT) calculations for models of correlated electron systems can

4: be significantly accelerated by using an appropriate mixing of hybridization

5: functions which are used as the input to the impurity solver. It is shown

6: that the techniques and the past experience with the mixing of input charge

7: densities in the density-functional-theory (DFT) calculations are also

8: effective in DMFT. As an example, the increase of the computational

9: requirements near the Mott metal-insulator transition in the Hubbard model

10: due to critical slowing down can be strongly reduced by using the modified

11: Broyden's method to numerically solve the non-linear self-consistency

12: equation. Speed-up factors as high as 3 were observed in practical

13: calculations even for this relatively well behaved problem. Furthermore, the

14: convergence can be achieved in difficult cases where simple linear mixing is

15: either not effective or even leads to divergence. Unstable and metastable

16: solutions can also be obtained.

17: %

18: %

19: We also determine the linear response of the system with respect to the

20: variations of the hybridization function, which is related to the

21: propagation of the information between the different energy scales during

22: the iteration.

23: \end{abstract}

24: