1: \begin{abstract}

2: Anderson Acceleration (AA) is a popular algorithm designed to  enhance 

3: the convergence of fixed-point iterations.

4: In this paper, we introduce a variant of AA based on a Truncated

5: Gram-Schmidt process (AATGS) which has a few advantages over the classical AA. 

6: In particular, an attractive

7: feature of AATGS is that its iterates obey a three-term recurrence in the situation

8: when it is applied to solving  symmetric linear problems and this can  lead to a

9: considerable reduction of memory and computational costs.

10: We analyze the convergence of AATGS in

11: both full-depth and limited-depth scenarios and establish its

12: equivalence to the classical AA in the linear case.

13: We also report on the effectiveness of AATGS  through a set of numerical

14: experiments, ranging from solving nonlinear partial differential

15: equations to tackling nonlinear optimization problems. In particular, the performance of the

16: method is compared with that of the classical AA algorithms.

17: \end{abstract}

18: