SMASH: Structured matrix approximation by separation and hierarchy

This paper presents an efficient method to perform Structured Matrix Approximation by Separation and Hierarchy (SMASH), when the original dense matrix is associated with a kernel function. Given points in a domain, a tree structure is first constructed based on an adaptive partitioning of the computational domain to facilitate subsequent approximation procedures. In contrast to existing schemes based on either analytic or purely algebraic approximations, SMASH takes advantage of both approaches and greatly improves the efficiency. The algorithm follows a bottom-up traversal of the tree and is able to perform the operations associated with each node on the same level in parallel. A strong rank-revealing factorization is applied to the initial analytic approximation in the separation regime so that a special structure is incorporated into the final nested bases. As a consequence, the storage is significantly reduced on one hand and a hierarchy of the original grid is constructed on the other hand. Due to this hierarchy, nested bases at upper levels can be computed in a similar way as the leaf level operations but on coarser grids. The main advantages of SMASH include its simplicity of implementation, its flexibility to construct various hierarchical rank structures and a low storage cost. Rigorous error analysis and complexity analysis are conducted to show that this scheme is fast and stable. The efficiency and robustness of SMASH are demonstrated through various test problems arising from integral equations, structured matrices, etc.

[1]  Jianlin Xia,et al.  On the Complexity of Some Hierarchical Structured Matrix Algorithms , 2012, SIAM J. Matrix Anal. Appl..

[2]  Ronald Kriemann,et al.  Hierarchical Matrices Based on a Weak Admissibility Criterion , 2004, Computing.

[3]  Mario Bebendorf,et al.  Hierarchical Matrices: A Means to Efficiently Solve Elliptic Boundary Value Problems , 2008 .

[4]  Sven Christophersen,et al.  Approximation of integral operators by Green quadrature and nested cross approximation , 2014, Numerische Mathematik.

[5]  Victor Y. Pan Fast Approximate Computations with Cauchy Matrices, Polynomials and Rational Functions , 2014, CSR.

[6]  W. Hackbusch,et al.  Introduction to Hierarchical Matrices with Applications , 2003 .

[7]  Per-Gunnar Martinsson,et al.  A direct solver with O(N) complexity for integral equations on one-dimensional domains , 2011, 1105.5372.

[8]  JIANLIN XIA,et al.  Parallel Randomized and Matrix-Free Direct Solvers for Large Structured Dense Linear Systems , 2016, SIAM J. Sci. Comput..

[9]  Victor Y. Pan,et al.  Transformations of Matrix Structures Work Again , 2013, 1303.0353.

[10]  Christopher R. Anderson,et al.  An Implementation of the Fast Multipole Method without Multipoles , 1992, SIAM J. Sci. Comput..

[11]  Federico Poloni,et al.  Fast solution of a certain Riccati equation through Cauchy-like matrices , 2009 .

[12]  Ming Gu,et al.  Stable and Efficient Algorithms for Structured Systems of Linear Equations , 1998, SIAM J. Matrix Anal. Appl..

[13]  Victor Y. Pan,et al.  How Bad Are Vandermonde Matrices? , 2015, SIAM J. Matrix Anal. Appl..

[14]  V. Pan Structured Matrices and Polynomials , 2001 .

[15]  Raymond H. Chan,et al.  A Fast Randomized Eigensolver with Structured LDL Factorization Update , 2014, SIAM J. Matrix Anal. Appl..

[16]  W. Hackbusch,et al.  On H2-Matrices , 2000 .

[17]  R. Kress Linear Integral Equations , 1989 .

[18]  Victor Y. Pan,et al.  c ○ 2003 Society for Industrial and Applied Mathematics INVERSION OF DISPLACEMENT OPERATORS ∗ , 2022 .

[19]  S. Börm Efficient Numerical Methods for Non-local Operators , 2010 .

[20]  Victor Y. Pan,et al.  Primitive and Cynical Low Rank Approximation, Preprocessing and Extensions , 2016 .

[21]  Thomas Kailath,et al.  Pivoting and backward stability of fast algorithms for solving Cauchy linear equations , 2002 .

[22]  Jianlin Xia,et al.  Superfast and Stable Structured Solvers for Toeplitz Least Squares via Randomized Sampling , 2014, SIAM J. Matrix Anal. Appl..

[23]  Mario Bebendorf,et al.  Approximation of boundary element matrices , 2000, Numerische Mathematik.

[24]  V. Hutson Integral Equations , 1967, Nature.

[25]  Lexing Ying,et al.  Fast construction of hierarchical matrix representation from matrix-vector multiplication , 2009, J. Comput. Phys..

[26]  Sabine Le Borne,et al.  H-matrix Preconditioners in Convection-Dominated Problems , 2005, SIAM J. Matrix Anal. Appl..

[27]  Nathan Halko,et al.  Finding Structure with Randomness: Probabilistic Algorithms for Constructing Approximate Matrix Decompositions , 2009, SIAM Rev..

[28]  Victor Y. Pan,et al.  Superfast algorithms for Cauchy-like matrix computations and extensions , 2000 .

[29]  Jianlin Xia,et al.  Fast algorithms for hierarchically semiseparable matrices , 2010, Numer. Linear Algebra Appl..

[30]  S. Chandrasekaran,et al.  A fast adaptive solver for hierarchically semiseparable representations , 2005 .

[31]  Thomas Kailath,et al.  Fast Gaussian elimination with partial pivoting for matrices with displacement structure , 1995 .

[32]  J. CARRIERt,et al.  A FAST ADAPTIVE MULTIPOLE ALGORITHM FOR PARTICLE SIMULATIONS * , 2022 .

[33]  T. Kailath,et al.  A fast parallel Björck–Pereyra-type algorithm for solving Cauchy linear equations , 1999 .

[34]  JIANLIN XIA,et al.  MULTI-LAYER HIERARCHICAL STRUCTURES AND FACTORIZATIONS , 2016 .

[35]  Peter Benner,et al.  Computing All or Some Eigenvalues of Symmetric H-Matrices , 2012, SIAM J. Sci. Comput..

[36]  Lexing Ying,et al.  Hierarchical Interpolative Factorization for Elliptic Operators: Integral Equations , 2013, 1307.2666.

[37]  Steffen Börm,et al.  Data-sparse approximation of non-local operators by H2-matrices , 2007 .

[38]  Lars Grasedyck,et al.  Existence of a low rank or ℋ︁‐matrix approximant to the solution of a Sylvester equation , 2004, Numer. Linear Algebra Appl..

[39]  Leslie Greengard,et al.  A fast algorithm for particle simulations , 1987 .

[40]  Lothar Reichel,et al.  Factorizations of Cauchy matrices , 1997 .

[41]  Leslie Greengard,et al.  A Fast Direct Solver for Structured Linear Systems by Recursive Skeletonization , 2012, SIAM J. Sci. Comput..

[42]  E. Tyrtyshnikov Mosaic-Skeleton approximations , 1996 .

[43]  Shivkumar Chandrasekaran,et al.  A Fast ULV Decomposition Solver for Hierarchically Semiseparable Representations , 2006, SIAM J. Matrix Anal. Appl..

[44]  Steffen Börm,et al.  Data-sparse Approximation by Adaptive ℋ2-Matrices , 2002, Computing.

[45]  D. Zorin,et al.  A kernel-independent adaptive fast multipole algorithm in two and three dimensions , 2004 .

[46]  S. Goreinov,et al.  A Theory of Pseudoskeleton Approximations , 1997 .

[47]  Boris N. Khoromskij,et al.  Solution of Large Scale Algebraic Matrix Riccati Equations by Use of Hierarchical Matrices , 2003, Computing.

[48]  Per-Gunnar Martinsson,et al.  A Fast Randomized Algorithm for Computing a Hierarchically Semiseparable Representation of a Matrix , 2011, SIAM J. Matrix Anal. Appl..

[49]  Steffen Börm,et al.  Hybrid cross approximation of integral operators , 2005, Numerische Mathematik.

[50]  V. Pan Structured Matrices and Polynomials: Unified Superfast Algorithms , 2001 .

[51]  Jianlin Xia,et al.  Efficient Structured Multifrontal Factorization for General Large Sparse Matrices , 2013, SIAM J. Sci. Comput..

[52]  Pieter Ghysels,et al.  A Distributed-Memory Package for Dense Hierarchically Semi-Separable Matrix Computations Using Randomization , 2015, ACM Trans. Math. Softw..

[53]  Ming Gu,et al.  Subspace Iteration Randomization and Singular Value Problems , 2014, SIAM J. Sci. Comput..

[54]  Steffen Börm,et al.  Approximation of Integral Operators by -Matrices with Adaptive Bases , 2005, Computing.

[55]  W. S. Venturini Boundary Integral Equations , 1983 .

[56]  W. Hackbusch,et al.  Hierarchical Matrices: Algorithms and Analysis , 2015 .

[57]  Mario Bebendorf,et al.  Constructing nested bases approximations from the entries of non-local operators , 2012, Numerische Mathematik.

[58]  W. Hackbusch,et al.  H 2 -matrix approximation of integral operators by interpolation , 2002 .

[59]  Sergej Rjasanow,et al.  Adaptive Low-Rank Approximation of Collocation Matrices , 2003, Computing.

[60]  Victor Y. Pan,et al.  N A ] 5 J un 2 01 6 Fast Low-rank Approximation of a Matrix : Novel Insights , Novel Multipliers , and Extensions ∗ , 2016 .

[61]  Xiaobai Sun,et al.  A Matrix Version of the Fast Multipole Method , 2001, SIAM Rev..

[62]  Ming Gu,et al.  Efficient Algorithms for Computing a Strong Rank-Revealing QR Factorization , 1996, SIAM J. Sci. Comput..

[63]  Eugene E. Tyrtyshnikov,et al.  Matrix‐free iterative solution strategies for large dense linear systems , 1997 .

[64]  K. Atkinson The Numerical Solution of Integral Equations of the Second Kind , 1997 .

[65]  Jianlin Xia,et al.  A Superfast Structured Solver for Toeplitz Linear Systems via Randomized Sampling , 2012, SIAM J. Matrix Anal. Appl..

[66]  Victor Y. Pan,et al.  Fast approximate computations with Cauchy matrices and polynomials , 2015, Math. Comput..

[67]  V. Rokhlin,et al.  A fast direct solver for boundary integral equations in two dimensions , 2003 .

[68]  Steffen Börm Construction of Data-Sparse H2-Matrices by Hierarchical Compression , 2009, SIAM J. Sci. Comput..

[69]  Piet Hut,et al.  A hierarchical O(N log N) force-calculation algorithm , 1986, Nature.

[70]  V. Rokhlin Rapid solution of integral equations of classical potential theory , 1985 .

[71]  Eugene E. Tyrtyshnikov,et al.  Incomplete Cross Approximation in the Mosaic-Skeleton Method , 2000, Computing.

[72]  Jianlin Xia,et al.  On the Stability of Some Hierarchical Rank Structured Matrix Algorithms , 2016, SIAM J. Matrix Anal. Appl..