1: \begin{abstract}

2: A novel method which is called the Chebyshev inertial iteration for accelerating the convergence speed of 

3: fixed-point iterations is presented. 

4: The Chebyshev inertial iteration can be regarded as a valiant of

5: the successive over relaxation or Krasnosel'ski\v{\i}-Mann iteration

6: utilizing the inverse of roots of a Chebyshev polynomial as 

7: iteration dependent inertial factors.

8: One of the most notable features of the proposed method is that it can be applied to

9: nonlinear fixed-point iterations in addition to linear fixed-point iterations.

10: Linearization around the fixed point is the key 

11: for the analysis on the local convergence rate of the proposed method.

12: The proposed method appears effective in particular for accelerating the proximal gradient methods 

13: such as ISTA. It is also proved that the proposed method can successfully accelerate almost any fixed-point iterations if all the eigenvalues of 

14: the Jacobian at the fixed point are real. 

15: \end{abstract}

16: