1: \begin{abstract}

2: Facial reduction, FR, is a regularization technique for convex

3: programs where the strict feasibility constraint qualification, CQ, fails.

4: Though this CQ holds generically, failure is pervasive in applications

5: such as semidefinite relaxations of hard discrete optimization problems.

6: In this paper we relate FR

7: to the analysis of the convergence behaviour of a semi-smooth

8: Newton root finding method for the projection onto a spectrahedron, 

9: i.e., onto the intersection of a linear manifold and the semidefinite cone.

10: In the process, we derive and use an elegant formula for the 

11: projection onto a face of the semidefinite cone.

12: We show further that

13: the ill-conditioning of the Jacobian of the Newton method near optimality 

14: characterizes the degeneracy of the nearest point in the spectrahedron.

15: We apply the results, both theoretically and empirically,

16:  to the problem of finding nearest points to the

17: sets of: (i) correlation matrices or the elliptope; and 

18: (ii) semidefinite relaxations of permutation matrices or the vontope,

19: i.e., the feasible sets for the semidefinite relaxations of the

20: max-cut and quadratic assignment

21: problems, respectively. 

22: %This further allows for characterizations of degeneracy

23: %based on singularity of the Jacobian for the projection problem.

24: 

25: 

26: \end{abstract}

27: