This data set may be cited by referencing:
This is a collection of symmetric and quasi-definite linear systems in MatrixMarket format.
The systems arise from some of the CUTE quadratic optimization problems, and are output during the iterations of an interior-point method. As the iteration number grows, the system becomes more ill conditioned.
The 2x2 and 3x3 formulations of each system are given with accompanying right-hand side.
D. Orban. Limited-Memory LDLT Factorization of Symmetric
Quasi-Definite Matrices with Application to Constrained Optimization. Cahier
du GERAD G-2013-87. GERAD, Montreal, Canada.
Technical Report.
Published version, Numerical Algorithms, November 2014.