Abstract
Many matrices appearing in numerical methods for partial differential equations and integral equations are rank-structured, i.e., they contain submatrices that can be approximated by matrices of low rank. A relatively general class of rank-structured matrices are \({\mathcal {H}}^2\)-matrices: they can reach the optimal order of complexity, but are still general enough for a large number of practical applications. We consider algorithms for performing algebraic operations with \({\mathcal {H}}^2\)-matrices, i.e., for approximating the matrix product, inverse or factorizations in almost linear complexity. The new approach is based on local low-rank updates that can be performed in linear complexity. These updates can be combined with a recursive procedure to approximate the product of two \({\mathcal {H}}^2\)-matrices, and these products can be used to approximate the matrix inverse and the LR or Cholesky factorization. Numerical experiments indicate that the new algorithm leads to preconditioners that require \({\mathcal {O}}(n)\) units of storage, can be evaluated in \({\mathcal {O}}(n)\) operations, and take \({\mathcal {O}}(n \log n)\) operations to set up.
Similar content being viewed by others
References
Bebendorf, M.: Why finite element discretizations can be factored by triangular hierarchical matrices. SIAM J. Numer. Anal. 45(4), 1472–1494 (2007)
Bebendorf, M., Hackbusch, W.: Existence of \({\cal H}\)-matrix approximants to the inverse FE-matrix of elliptic operators with \(L^{\infty }\)-coefficients. Numer. Math. 95, 1–28 (2003)
Börm, S.: \({\cal H}^2\)-matrix arithmetics in linear complexity. Computing 77(1), 1–28 (2006)
Börm, S.: Adaptive variable-rank approximation of dense matrices. SIAM J. Sci. Comp. 30(1), 148–168 (2007)
Börm, S.: Data-sparse approximation of non-local operators by \({\cal H}^2\)-matrices. Lin. Algebra Appl. 422, 380–403 (2007)
Börm, S.: Approximation of solution operators of elliptic partial differential equations by \({\cal H}\)- and \({\cal H}^{2}\)-matrices. Numer. Math. 115(2), 165–193 (2010)
Börm, S.: Efficient Numerical Methods for Non-local Operators: \({\cal H}^{2}\)-Matrix Compression, Algorithms and Analysis. EMS Tracts in Mathematics, vol. 14, (2010)
Börm, S., Hackbusch, W.: Data-sparse approximation by adaptive \({\cal H}^2\)-matrices. Computing 69, 1–35 (2002)
Börm, S., Löhndorf, M., Melenk, J.M.: Approximation of integral operators by variable-order interpolation. Numer. Math. 99(4), 605–643 (2005)
Chandrasekaran, S., Gu, M., Lyons, W.: A fast adaptive solver for hierarchically semiseparable representations. Calcolo 42, 171–185 (2005)
Faustmann, M., Melenk, J.M., Praetorius, D.: Existence of \({\cal H}\)-matrix approximants to the inverses of BEM matrices: the simple-layer operator. Tech. Rep. 37, Institut für Analysis und Scientific Computing, TU Wien (2013). http://www.asc.tuwien.ac.at/preprint/2013/asc37x2013.pdf
Faustmann, M., Melenk, J.M., Praetorius, D.: \({\cal H}\)-matrix approximability of the inverse of FEM matrices. Tech. Rep. 20, Institut für Analysis und Scientific Computing, TU Wien (2013). http://www.asc.tuwien.ac.at/preprint/2013/asc20x2013.pdf
Grasedyck, L., Hackbusch, W.: Construction and arithmetics of \({\cal H}\)-matrices. Computing 70, 295–334 (2003)
Grasedyck, L., Kriemann, R., LeBorne, S.: Domain decomposition based \({\cal H}\)-LU preconditioning. Numer. Math. 112(4), 565–600 (2009)
Hackbusch, W.: A sparse matrix arithmetic based on \({\cal H}\)-matrices. Part I: introduction to \({\cal H}\)-matrices. Computing 62, 89–108 (1999)
Hackbusch, W.: Hierarchische Matrizen—Algorithmen und Analysis. Springer, Berlin (2009)
Hackbusch, W., Khoromskij, B.N.: A sparse matrix arithmetic based on \({\cal H}\)-matrices. Part II: application to multi-dimensional problems. Computing 64, 21–47 (2000)
Hackbusch, W., Khoromskij, B.N., Sauter, S.A.: On \({\cal H}^{2}\)-matrices. In: Bungartz, H., Hoppe, R., Zenger, C. (eds.) Lectures on Applied Mathematics, pp. 9–29. Springer, Berlin (2000)
Hestenes, M.R., Stiefel, E.: Methods of conjugate gradients for solving linear systems. J. Res. Nat. B. Stand. 49(6), 409–436 (1952)
Lintner, M.: The eigenvalue problem for the 2d Laplacian in \({\cal H}\)-matrix arithmetic and application to the heat and wave equation. Computing 72, 293–323 (2004)
Martinsson, P.G.: A fast direct solver for a class of elliptic partial differential equations. J. Sci. Comput. 38, 316–330 (2008)
Martinsson, P.G., Rokhlin, V., Tygert, M.: A fast algorithm for the inversion of general Toeplitz matrices. Comp. Math. Appl. 50, 741–752 (2005)
Xia, J., Chandrasekaran, S., Gu, M., Li, X.: Superfast multifrontal method for large structured linear systems of equations. SIAM J. Matrix Anal. Appl. 31(3), 1382–1411 (2009)
Xia, J., Chandrasekaran, S., Gu, M., Li, X.S.: Fast algorithms for hierarchically semiseparable matrices. Numer. Lin. Alg. Appl. (2009). doi:10.1002/nla.691
Acknowledgments
A major part of this research was funded by the Deutsche Forschungsgemeinschaft in the context of project BO 3289/4-1.
Author information
Authors and Affiliations
Corresponding author
Additional information
Communicated by Gabriel Wittum.
Rights and permissions
About this article
Cite this article
Börm, S., Reimer, K. Efficient arithmetic operations for rank-structured matrices based on hierarchical low-rank updates. Comput. Visual Sci. 16, 247–258 (2013). https://doi.org/10.1007/s00791-015-0233-3
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00791-015-0233-3