1: \begin{abstract}

2: Generalized alternating projections is an algorithm that alternates

3: relaxed projections onto a finite number of sets to find a point in

4: their intersection. We consider the special case of two linear

5: subspaces, for which the algorithm reduces to a matrix iteration. For

6: convergent matrix iterations, the asymptotic rate is linear and decided

7: by the magnitude of the subdominant eigenvalue. In this paper, we show

8: how to select the three algorithm parameters to optimize this magnitude,

9: and hence the asymptotic convergence rate. The obtained rate depends on

10: the Friedrichs angle between the subspaces and is considerably better

11: than known rates for other methods such as alternating projections and

12: Douglas-Rachford splitting. We also present an adaptive scheme that,

13: online, estimates the Friedrichs angle and updates the algorithm

14: parameters based on this estimate. A numerical example is provided that

15: supports our theoretical claims and shows very good performance for the

16: adaptive method.

17: \end{abstract}

18: