Abstract
In this paper we prove that, given a \(n \times n\) symmetric matrix A, a matrix V with r orthonormal columns and an integer \(m \ge 1\), \(mr \le n\), it is possible to devise a matrix algebra \(\mathcal {L}\) such that, denoting by \(\mathcal {L}_A\) the matrix closest to A from \(\mathcal {L}\) in the Frobenius norm, one has \(\mathcal {L}^j_AV=A^jV\) for \(j=0,\dots ,m-1\). The algebra \(\mathcal {L}\) is the space of all matrices that are diagonalized by a given orthogonal matrix L. We show, moreover, that L can be constructed as the product of mr Householder matrices, so that \(\mathcal {L}\), for \(mr \ll n\), is a low complexity matrix algebra. The new theoretical results here introduced allow to investigate new possible preconditioners \({\mathcal {L}}_A\) for the Conjugate Gradient method and new quasi-Newton algorithms suitable to solve large scale optimization problems.
Similar content being viewed by others
References
Bell, N., Garland, M.: Cusp: Generic Parallel Algorithms for Sparse Matrix and Graph Computations. Version 0.5.1. http://cusplibrary.github.io/ (2015)
Benzi, M.: Preconditioning techniques for large linear systems: a survey. J. Comput. Phys. 182(2), 418–477 (2002)
Bergamaschi, L., Martínez, Á.: Two-stage spectral preconditioners for iterative eigensolvers. Numer. Linear Algebra Appl. 24(3), e2084 (2017)
Bertaccini, D., Durastante, F.: Iterative Methods and Preconditioning for Large and Sparse Linear Systems with Applications. Chapman and Hall/CRC, Boca Raton (2018)
Bortoletti, A., Di Fiore, C., Fanelli, S., Zellini, P.: A new class of quasi-Newtonian methods for optimal learning in MLP-networks. IEEE Trans. Neural Netw. Learn. Syst. 14(2), 263–273 (2003)
Bridson, R., Tang, W.P.: Multiresolution approximate inverse preconditioners. SIAM J. Sci. Comput. 23(2), 463–479 (2001)
Byrd, R.H., Nocedal, J., Yuan, Y.X.: Global convergence of a cass of quasi-Newton methods on convex problems. SIAM J. Numer. Anal. 24(5), 1171–1190 (1987)
Capizzano, S.S.: Superlinear PCG methods for symmetric Toeplitz systems. Math. Comput. 68, 793–803 (1999)
Chen, J., Saad, Y.: Lanczos vectors versus singular vectors for effective dimension reduction. IEEE Trans. Knowl. Data Eng. 21(8), 1091–1103 (2009)
Cipolla, S., Di Fiore, C., Zellini, P.: Updating Broyden Class-type descent directions by Householder adaptive Transforms (2018)
Cipolla, S., Di Fiore, C., Tudisco, F., Zellini, P.: Adaptive matrix algebras in unconstrained minimization. Linear Algebra Appl. 471, 544–568 (2015)
Cipolla, S., Di Fiore, C., Tudisco, F.: Euler–Richardson method preconditioned by weakly stochastic matrix algebras: a potential contribution to Pagerank computation. Electron. J. Linear Algebra 32, 254–272 (2017)
Cipolla, S., Di Fiore, C., Durastante, F., Zellini, P.: Regularizing properties of a class of matrices including the optimal and the superoptimal preconditioners. Numer. Linear Algebra Appl. 26, e2225 (2019)
Davis, T.A., Hu, Y.: The university of Florida sparse matrix collection. ACM Trans. Math. Softw. (TOMS) 38(1), 1–25 (2011)
Di Benedetto, F., Capizzano, S.S.: A unifying approach to abstract matrix algebra preconditioning. Numer. Math. 82, 57–90 (1999)
Di Benedetto, F., Estatico, C., Serra Capizzano, S.: Superoptimal preconditioned conjugate gradient iteration for image deblurring. SIAM J. Sci. Comput. 26(3), 1012–1035 (2005)
Di Fiore, C.: Structured matrices in unconstrained minimization methods. In: Olshevsky, V., (ed.) Fast Algorithms for Structured Matrices: Theory and Applications, Contemporary Mathematics, vol. 323. AMS SIAM (2003)
Di Fiore, C., Zellini, P.: Matrix algebras in optimal preconditioning. Linear Algebra Appl. 335, 1–54 (2001)
Di Fiore, C., Fanelli, S., Lepore, F., Zellini, P.: Matrix algebras in quasi-Newton methods for unconstrained minimization. Numer. Math. 94, 479–500 (2003)
Di Fiore, C., Lepore, F., Zellini, P.: Hartley-type algebras in displacement and optimization strategies. Linear Algebra Appl. 366, 215–232 (2003)
Di Fiore, C., Fanelli, S., Zellini, P.: Low-complexity minimization algorithms. Numer. Linear Algebra Appl. 12(8), 755–768 (2005)
Estatico, C.: A class of filtering superoptimal preconditioners for highly ill-conditioned linear systems. BIT 42(4), 753–778 (2002)
Fezzani, R., Grigori, L., Nataf, F., Wang, K.: Block filtering decomposition. Numer. Linear Algebra Appl. 21(6), 703–721 (2014)
Freitag, M.A., Spence, A.: A tuned preconditioner for inexact inverse iteration applied to Hermitian eigenvalue problems. IMA J. Numer. Anal. 28(3), 522–551 (2007)
Golub, G.H., Van Loan, C.F.: Matrix Computations, vol. 3. JHU Press, Baltimore (2012)
Goreinov, S.A., Tyrtyshnikov, E.E., Zamarashkin, N.L.: A theory of pseudoskeleton approximations. Linear Algebra Appl. 261, 1–21 (1997a)
Goreinov, S.A., Zamarashkin, N.L., Tyrtyshnikov, E.E.: Pseudo-skeleton approximations by matrices of maximal volume. Math. Notes 62, 515–519 (1997b)
Gutknecht, M.H.: Block Krylov space methods for linear systems with multiple right-hand sides: an introduction. In: Siddiqi, A.H., Duff, I.S., Christensen, O. (eds.) Modern Mathematical Models, Methods and Algorithms for Real World Systems, pp. 420–447. Anamaya Publishers, New Delhi (2007)
Hanke, M., Nagy, J., Plemmons, R.: Preconditioned iterative regularization for lll-posed problems. Retrieved from the University of Minnesota Digital Conservancy (1992). http://hdl.handle.net/11299/2169
Horn, R.A., Johnson, C.R.: Matrix Analysis, 2nd edn. Cambridge University Press, Cambridge (2013)
Jin, X.Q., Wei, Y.M.: A survey and some extensions of T. Chan’s preconditioner. Linear Algebra Appl. 428(2), 403–412 (2008)
Kailath, T., Sayed, A.H.: Fast Reliable Algorithms for Matrices with Structure. SIAM, Philadelphia (1999)
Lopez, L., Simoncini, V.: Preserving geometric properties of the exponential matrix by block Krylov subspace methods. BIT 46(4), 813–830 (2006)
Martínez, Á.: Tuned preconditioners for the eigensolution of large spd matrices arising in engineering problems. Numer. Linear Algebra Appl. 23(3), 427–443 (2016)
Ming, G., Li, X.S., Vassilevski, P.S.: Direction-preserving and Schur-monotonic semiseparable approximations of symmetric positive definite matrices. SIAM J. Matrix Anal. Appl. 31, 2650–2664 (2010)
Nocedal, J., Wright, S.J.: Numerical Optimization, 2nd edn. Springer, Berlin (2006)
Osinsky, A., Zamarashkin, N.: Pseudo-skeleton approximations with better accuracy estimates. Linear Algebra Appl. 537, 221–249 (2018)
Powell, M.J.D.: Some global convergence properties of a variable metric algorithm for minimization without exact line searches. Nonlinear Program. 9, 53–72 (1976)
Saad, Y.: Analysis of some Krylov subspace approximations to the matrix exponential operator. SIAM J. Numer. Anal. 29(1), 209–228 (1992)
Saad, Y.: Iterative Methods for Sparse Linear Systems. SIAM, Philadelphia (2003)
Saad, Y.: Numerical Methods for Large Eigenvalue Problems. SIAM, Philadelphia (2011)
Sadkane, M.: Block-Arnoldi and Davidson methods for unsymmetric large eigenvalue problems. Numer. Math. 64(1), 195–211 (1993)
Tyrtyshnikov, E.E.: Optimal and superoptimal circulant preconditioners. SIAM J. Matrix Anal. Appl. 13(2), 459–473 (1992)
Tyrtyshnikov, E.E.: A unifying approach to some old and new theorems on distribution and clustering. Linear Algebra Appl. 232, 1–43 (1996)
Wagner, C.: Tangential frequency filtering decompositions for unsymmetric matrices. Numer. Math. 78(1), 143–163 (1997)
Wagner, C., Wittum, G.: Adaptive filtering. Numer. Math. 78(2), 305–328 (1997)
Acknowledgements
S. Cipolla acknowledges Dr. F. Durastante for pointing out Lemma 3.1 in [39] and for fruitful discussions on its generalization (Lemma 3). Moreover, he would like to thank Professor M.A. Botchev because the original idea of Theorem 1 was born during his lectures in the Rome-Moscow school of Matrix Methods and Applied Linear Algebra (2016 Edition). We acknowledge the anonymous referee for pointing relevant typos and mistakes and for his/her remarks which improved significantly the quality of the manuscript and leaded, in particular, to the statement of the result in Corollary 1. S.C. and C.D.F. are members of the INdAM Research group GNCS, which partially supported this work. The work of C.D.F. was partially supported by MIUR Excellence Department Project awarded to the Department of Mathematics, University of Rome Tor Vergata, CUP E83C18000100006.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
About this article
Cite this article
Cipolla, S., Di Fiore, C. & Zellini, P. Low complexity matrix projections preserving actions on vectors. Calcolo 56, 8 (2019). https://doi.org/10.1007/s10092-019-0305-8
Received:
Accepted:
Published:
DOI: https://doi.org/10.1007/s10092-019-0305-8
Keywords
- Arnoldi method
- Block-Krylov spaces
- Direction preserving projections
- Matrix projections
- Unitary decomposition by householder matrices