1: \begin{abstract}

2: This paper studies the Craig variant of the Golub-Kahan bidiagonalization algorithm

3: as an iterative solver for linear systems with saddle point structure. Such symmetric indefinite systems in 2x2 block form

4: arise in many applications, but standard iterative solvers are often found

5: to perform poorly on them and robust preconditioners may not be available. Specifically, such systems arise in structural mechanics, when a

6: semidefinite finite element stiffness matrix is augmented with linear multi-point constraints via Lagrange multipliers. Engineers often use such multi-point constraints to introduce boundary or coupling conditions into complex finite element models. The article will present a systematic convergence study of the Golub-Kahan algorithm

7: for a sequence of test problems of increasing complexity, including concrete structures enforced with pretension cables and the coupled finite element model of a reactor containment building.

8: When the systems are suitably transformed using

9: augmented Lagrangians on the semidefinite block and 

10: when the constraint equations are properly scaled, the

11: Golub-Kahan algorithm is found to exhibit excellent convergence that depends only

12: weakly on the size of the model.

13: The new algorithm is found to be robust in practical cases that are otherwise

14: considered to be difficult for iterative solvers. 

15: \end{abstract}

16: