数値解析における数値線形代数(すうちせんけいだいすう、英: Numerical linear algebra)とは、線形代数で現れる問題(行列積、行列指数関数、連立方程式や固有値・特異値問題)の計算・求解を行うアルゴリズムを創出するための学問である[1][2][3]。最適化問題・有限差分法・有限要素法などに応用されている[1]。
現代の主流
現代においては
などをはじめとした高速・高精度解法(反復法)の研究が主流になっている[2][5][6]。
クリロフ部分空間法
数値線形代数で現れる反復法の中には、クリロフ部分空間に理論的基盤を持つものが少なからず存在する。これらはクリロフ部分空間法(英: Krylov subspace methods)と総称され、数値線形代数において最も成功した手法とされている[6]。主なクリロフ部分空間法として以下が知られている(共役勾配法については後述する)。
共役勾配法
共役勾配法(英: conjugate gradient method)は Hestenes-Stiefel によって開発された連立方程式の数値解法であり、係数行列が正定値対称行列であるときに適用できる[A 31]。この方法はガウス=ザイデル法、ヤコビ法、SOR法よりも収束が速いとされることから、1980年代以降から様々な亜種が開発されたり、非対称行列への適用が試みられているが、前処理行列の取り方が問題によって異なるために決定版と言える解法がまだ存在してない[1][2][6]。
以下、代表的な亜種を挙げる。
- CGS (conjugate gradient squared method)[A 32]
- PCG (preconditioned conjugate gradient method、MATLABで利用可能)
- SCG (scaled conjugate gradient)[A 33]
- ICCG (incomplete Cholesky conjugate gradient method、不完全コレスキー分解付共役勾配法)[6]
- COCG (conjugate orthogonal conjugate gradient method)[A 34]
- GPBiCG[A 35][A 36]
- GPBiCG[A 37]
- STAB系の手法
- Block化した手法
可積分系・力学系と数値線形代数の間に関係があるという事実が注目されている[8][9][C 1][C 2]。具体的には戸田格子(ソリトン方程式の一つ)とQR法[8][9]・qd法[C 3]、可積分系と特異値分解[8][9][C 4][C 5]との関係が明らかになった。これらを背景に、離散可積分系から数値線形代数に有用な反復法を開発する試みが行われている。例えば
等が研究されている[8][9]。
戸川隼人. (1977). 共役勾配法. シリーズ新しい応用の数学.
反復法・高精度計算に関する論文
Fernando, K. V., & Parlett, B. N. (1994). Accurate singular values and differential qd algorithms. en:Numerische Mathematik, 67(2), 191-229.
U. von Matt, The orthogonal QD algorithm, SIAM J. Sci. Comput., 18 (1997), 1163–1186.
Dhillon, I. S., Parlett, B. N., & Vömel, C. (2006). The design and implementation of the MRRR algorithm. ACM Transactions on Mathematical Software (TOMS), 32(4), 533-560.
Abe, K., Zhang, S. L., & Mitsui, T. (1997). MRTR method: An iterative method based on the three-term recurrence formula of CG-type for nonsymmetric matrix. JSIAM, 7, 37-50.
Sakurai, T., & Tadano, H. (2007). CIRR: a Rayleigh-Ritz type method with contour integral for generalized eigenvalue problems. Hokkaido mathematical journal, 36(4), 745-757.
Tsutomu Ikegami, Tetsuya Sakurai and Umpei Nagashima: A Filter Diagonalization for Generalized Eigenvalue Problems Based on the Sakurai-Sugiura Projection Method, J. Compu. Appl. Math., Vol.233, No.8, pp.1927–1936 (2010).
Anthony P. Austin and Lloyd N. Trefethen: Computing Eigenvalues of Real Symmetric Matrices with Rational Filters in Real Arithmetic, SIAM J. Sci. Comput, Vol.37, No.3, pp.A1365–A1387 (2015).
Hiroshi Murakami: Filter Diagonalization Method by Using a Polynomial of a Resolvent as the Filter for a Real Symmetric-Definite Generalized Eigenproblem, in proceedings of EPASA2015, Springer, LNCSE-117, pp.205–232 (2018).
Hiroshi Murakami: Filters Consist of a Few Resolvents to Solve Real Symmetric-Definite Generalized Eigenproblems, Japan J. Indust. Appl. Math., Vol.36, No.2, pp.579–618 (July 2019).
Krasnoselskii, M. and Krein, S. (1952). An iteration process with minimal residuals. Mat. Sb. N.S., 31, 315–334.
Eduard Stiefel, Commentarii Mathematici Helvetici 1952; 29(1):157–179.
Freund, R. and Nachtigal, N. "QMR: A Quasi-Minimal Residual Method for Non-Hermitian Linear Systems." Numer. Math. 60, 315-339, 1991.
Freund, R. and Nachtigal, N. "An Implementation of the QMR Method Based on Coupled Two-Term Recurrences." SIAM J. Sci. Statist. Comput. 15, 313-337, 1994.
R. Freund: Conjugate Gradient-Type Methods for Linear Systems with Complex Symmetric Coefficient Matrices, SIAM Journal on Scientific and Statistical Computing, Vol. 13, No. 1, pp. 425–448 (1992).
X.-M. Gu, T.-Z. Huang, L. Li, H.-B. Li, T. Sogabe and M. Clemens: Quasi-Minimal Residual Variants of the COCG and COCR Methods for Complex Symmetric Linear Systems in Electromagnetic Simulations, IEEE Transactions on Microwave Theory and Techniques, Vol. 62, No. 12, pp. 2859–2867 (2014).
Hestenes, M. R., & Stiefel, E. (1952). Methods of conjugate gradients for solving linear systems. Washington, DC: NBS.
M. F. Moller. A scaled conjugate gradient algorithm for fast supervised learning. Neural Networks, 6:525--533, 1993.
H. A. van der Vorst and J. B. M. Melissen: A Petrov-Galerkin type method for solving , where is symmetric complex, IEEE Trans. Mag., Vol. 26, No. 2, pp. 706–708 (1990).
S. L. Zhang "GPBiCG: Generalized Product-type Methods Based on Bi-CG for Solving
Nonsymmetric Linear Systems", SIAM J. Sci. Stat. Comput. , vol.18, no.2, pp.537-551,
March 1997.
張紹良, 藤野清次,ランチョス・プロセスに基づく積型反復解法,日本応用数理学会論文誌,5(1995), 343-360.
Van der Vorst, H. A. (1992). Bi-CGSTAB: A fast and smoothly converging variant of Bi-CG for the solution of nonsymmetric linear systems. SIAM Journal on scientific and Statistical Computing, 13(2), 631-644.
M. H.Gutknecht,Variants of BiCGSTAB for Matrices with Complex Spectrum,SIAM J. Sci. Statist. Comput.,14(1993), 1020-1033.
Tony F. Chan, Efstratios Gallopoulos, Valeria Simoncini, Tedd Szeto and Charles H. Tong, A quasi-minimal residual variant of the Bi-CGSTAB algorithm for nonsymmetric systems, en:SIAM Journal on Scientific Computing 1994; 15(2):338–347.
D. P. O’Leary: The block conjugate gradient algorithm and related methods, Linear Algebra Appl., 29 (1980), 293–322.
A. A. Nikishin, A. Yu. Yeremin: Variable block CG algorithms for solving large sparse symmetric positive definite linear systems on parallel computers, I: general iterative methods, SIAM J. Matrix Anal. Appl., 16 (1995), 1135–1153.
A. El Guennouni, K. Jbilou, H. Sadok: A block version of BiCGSTAB for linear systems with multiple right-hand sides, Electron. Trans. Numer. Anal, 16 (2003), 129–142.
H. Tadano, T. Sakurai, Y. Kuramashi: Block BiCGGR: a new block Krylov subspace method for computing high accuracy solutions: JSIAM Lett., 1 (2009), 44–47.
Tadano, H. (2019). Development of the Block BiCGGR2 method for linear systems with multiple right-hand sides. Japan Journal of Industrial and Applied Mathematics, 1-15.
Tadano, H., & Kuramoto, R. (2019). Accuracy improvement of the Block BiCGSTAB method for linear systems with multiple right-hand sides by group-wise updating technique. Journal of Advanced Simulation in Science and Engineering, 6(1), 100-117.
Bini, D. A., Higham, N. J., & Meini, B. (2005). Algorithms for the matrix pth root. Numerical Algorithms, 39(4), 349-378.
Deadman, E., Higham, N. J., & Ralha, R. (2012, June). Blocked Schur algorithms for computing the matrix square root. In International Workshop on Applied Parallel Computing (pp. 171-182). Springer, Berlin, Heidelberg.
Hargreaves, G. I., & Higham, N. J. (2005). Efficient algorithms for the matrix cosine and sine. Numerical Algorithms, 40(4), 383-400.
精度保証に関する論文
Yamamoto, T. (1984). Error bounds for approximate solutions of systems of equations. Japan Journal of Applied Mathematics, 1(1), 157.
Oishi, S., & Rump, S. M. (2002). Fast verification of solutions of matrix equations. en:Numerische Mathematik, 90(4), 755-773.
森倉悠介, 椋木大地, 深谷猛, 山中脩也, & 大石進一. (2016). 大規模並列計算機における連立一次方程式の精度保証付き数値計算に対する性能評価. 研究報告ハイパフォーマンスコンピューティング (HPC), 2016(1), 1-7.
Kobayashi, Y., Ogita, T., & Ozaki, K. (2017). Acceleration of a preconditioning method for ill-conditioned dense linear systems by use of a BLAS-based method. Reliable Computing, 25, 15-23.
Kobayashi, Y., & Ogita, T. (2016). Accurate and efficient algorithm for solving ill-conditioned linear systems by preconditioning methods. Nonlinear Theory and Its Applications, IEICE, 7(3), 374-385.
A fast and efficient algorithm for solving ill-conditioned linear systems (JSIAM Letters Vol.7 (2015) pp.1-4) Yuka Kobayashi, Takeshi Ogita.
悪条件連立一次方程式の精度保証付き数値計算法, 日本応用数理学会論文誌 Vol.15, No.3, 2005, pp.269-286, 太田貴久, 荻田武史, Siegfried M. Rump, 大石進一.
Yanagisawa, Y., Ogita, T., & Oishi, S. (2014). Convergence analysis of an algorithm for accurate inverse Cholesky factorization. Japan Journal of Industrial and Applied Mathematics, 31(3), 461-482.
Yanagisawa, Y., Ogita, T., & Oishi, S. I. (2014). A modified algorithm for accurate inverse Cholesky factorization. Nonlinear Theory and Its Applications, IEICE, 5(1), 35-46.
Yanagisawa, Y., & Ogita, T. (2013). Convergence analysis of accurate inverse Cholesky factorization. JSIAM Letters, 5, 25-28.
Minamihata, A., Sekine, K., Ogita, T., Rump, S. M., & Oishi, S. I. (2015). Improved error bounds for linear systems with H-matrices. Nonlinear Theory and Its Applications, IEICE, 6(3), 377-382.
Yamamoto, T. (1982). Error bounds for computed eigenvalues and eigenvectors. II. en:Numerische Mathematik, 40(2), 201-206.
Mayer, G. (1994). Result verification for eigenvectors and eigenvalues. Topics in Validated Computations, Elsevier, Amsterdam, 209-276.
Rump, S. M. (1994). Verification methods for dense and sparse systems of equations.
Rump, S. M. (2001). Computational error bounds for multiple or nearly multiple eigenvalues. Linear Algebra and its Applications, 324(1-3), 209-226.
Yamamoto, N. (2001). A simple method for error bounds of eigenvalues of symmetric matrices. Linear Algebra and its Applications, 324(1-3), 227-234.
Elsner, L., & Friedland, S. (1995). Singular values, doubly stochastic matrices, and applications. Linear Algebra and Its Applications, 220, 161-169.
Ipsen, I. C. (1998). Relative perturbation results for matrix eigenvalues and singular values. en:Acta numerica, 7, 151-201.
Deif, A. S. (1990). Realistic apriori and a posteriori error bounds for computed eigenvalues. IMA journal of numerical analysis, 10(3), 323-329.
丸山晃佐, 荻田武史, 中谷祐介, & 大石進一. (2004). 実対称定値一般化固有値問題のすべての固有値の精度保証付き数値計算法. 電子情報通信学会論文誌 A, 87(8), 1111-1119.
Alefeld, G., Gienger, A., & Mayer, G. (1994). Numerical validation for an inverse matrix eigenvalue problem. Computing, 53(3-4), 311-322.
Miyajima, S. (2017). Verified Solutions of Inverse Symmetric Eigenvalue Problems. Reliable Computing, 24(1), 31-44.
Demmel, J., & Kahan, W. (1990). Accurate singular values of bidiagonal matrices. SIAM Journal on Scientific and Statistical Computing, 11(5), 873-912.
Oishi, S. (2001). Fast enclosure of matrix eigenvalues and singular values via rounding mode controlled computation. Linear algebra and its Applications, 324(1-3), 133-146.
Ogita, T. (2008). Verified Numerical Computation of Matrix Determinant. SCAN’2008 El Paso, Texas September 29–October 3, 2008, 86.
有向丸めの変更を使用しないタイトな行列積の包含方法, 応用数理 Vol.21, No.3, 2011, pp.186–196, 尾崎克久, 荻田武史, 大石進一.
inya Miyajima, Fast verified computation for the minimal nonnegative solution of the nonsymmetric algebraic Riccati equation, Computational and Applied Mathematics, Volume 37, Issue 4, Pages 4599-4610, September 2018.
Shinya Miyajima, Fast verified computation for the solution of the T-congruence Sylvester equation, Japan Journal of Industrial and Applied Mathematics, Volume 35, Issue 2, Pages 541-551, July 2018.
Shinya Miyajima, Fast verified computation for the solvent of the quadratic matrix equation, The Electronic Journal of Linear Algebra, Volume 34, Pages 137-151, March 2018
Shinya Miyajima, Fast verified computation for solutions of algebraic Riccati equations arising in transport theory, Numerical Linear Algebra with Applications, Volume 24, Issue 5, Pages 1-12, October 2017.
Shinya Miyajima, Fast verified computation for solutions of continuous-time algebraic Riccati equations, Japan Journal of Industrial and Applied Mathematics, Volume 32, Issue 2, Pages 529-544, July 2015.
Miyajima, S. (2019). Verified computation of the matrix exponential. Advances in Computational Mathematics, 45(1), 137-152.
Miyajima, S. (2019). Verified computation for the matrix principal logarithm. Linear Algebra and its Applications, 569, 38-61.
可積分アルゴリズムに関する論文
Chu, M. T. (2008). Linear algebra algorithms as dynamical systems. en:Acta Numerica, 17, 1-86.
Sogo, K. (1993). Toda molecule equation and quotient-difference method. Journal of the Physical Society of Japan, 62(4), 1081-1084.
Iwasaki, M., & Nakamura, Y. (2006). Accurate computation of singular values in terms of shifted integrable schemes. Japan journal of industrial and applied mathematics, 23(3), 239.
「Verification of dLVv Transformation for Singular Vector Computation with High Accuracy」 『In Proceedings of International Conference on Parallel and Distributed Processing Techniques and Applications』 Vol.II, pp.881-887 (2006, 6) Masami Takata, Taro Konda, Kinji Kimura, Yoshimasa Nakamura.
「An Evaluation of Singular Value Computation by the Discrete Lotka-Volterra System」 『In Proceedings of International Conference on Parallel and Distributed Processing Techniques and Applications』 Vol.II, pp.410-416 (2005, 6) Masami Takata, Iwasaki Masashi, Kinji Kimura, Yoshimasa Nakamura.
Iwasaki, M., & Nakamura, Y. (2011). Positivity of dLV and mdLVs algorithms for computing singular values. Electronic Transactions on Numerical Analysis, 38, 184-201.
Fukuda, A., Ishiwata, E., Iwasaki, M., & Nakamura, Y. (2008). The discrete hungry Lotka–Volterra system and a new algorithm for computing matrix eigenvalues. Inverse Problems, 25(1), 015007.
Fukuda, A., Ishiwata, E., Yamamoto, Y., Iwasaki, M., & Nakamura, Y. (2013). Integrable discrete hungry systems and their related matrix eigenvalues. Annali di Matematica Pura ed Applicata, 192(3), 423-445.
Fukuda, A., Ishiwata, E., Iwasaki, M., & Nakamura, Y. (2009). On the qd-type discrete hungry Lotka-Volterra system and its application to the matrix eigenvalue algorithm. JSIAM Letters, 1, 36-39.
Yamamoto, Y., & Fukaya, T. (2010). Differential qd algorithm for totally nonnegative Hessenberg matrices: introduction of origin shifts and relationship with the discrete hungry Lotka-Volterra system. JSIAM Letters, 2, 69-72.
「An Improved Shift Strategy for the Modified Discrete Lotka-Volterra with Shift Algorithm」 『In Proceedings of 2011 International Conference on Parallel and Distributed Processing Techniques and Applications』 Vol.II, pp.720-726 (2011, 7) Masami Takata, Takumi Yamashita, Akira Ajisaka, Kinji Kimura, Yoshimasa Nakamura.
「Speed-up in mdLVs by Limitation in Computational Number of Shift」 『In Proceedings of 2009 International Conference on Parallel and Distributed Processing Techniques and Applications』 Vol.II, pp.704-710 (2009, 7) Masami Takata, Kinji Kimura, Yoshimasa Nakamura.
Takahashi, Y., Iwasaki, M., Fukuda, A., Ishiwata, E., & Nakamura, Y. (2018). Periodic convergence in the discrete hungry Toda equation. Journal of Physics A: Mathematical and Theoretical, 51(34), 344001.
Fukuda, A., Yamamoto, Y., Iwasaki, M., Ishiwata, E., & Nakamura, Y. (2013). On a shifted LR transformation derived from the discrete hungry Toda equation. Monatshefte für Mathematik, 170(1), 11-26.
Fukuda, A., Yamamoto, Y., Iwasaki, M., Ishiwata, E., & Nakamura, Y. (2012). Error analysis for matrix eigenvalue algorithm based on the discrete hungry Toda equation. Numerical Algorithms, 61(2), 243-260.
Toyokawa, H., Kimura, K., Takata, M., & Nakamura, Y. (2009). On parallelism of the I-SVD algorithm with a multi-core processor. JSIAM Letters, 1, 48-51.
洋書
- Demmel, J. W. (1997). Applied numerical linear algebra. SIAM.
- Ciarlet, P. G., Miara, B., & Thomas, J. M. (1989). Introduction to numerical linear algebra and optimization. Cambridge University Press.
- Trefethen, Lloyd; Bau III, David (1997). Numerical Linear Algebra (1st ed.). Philadelphia: SIAM. ISBN 978-0-89871-361-9.
- Golub, Gene H.; Van Loan, Charles F. (1996). Matrix Computations (3rd ed.). Baltimore: The Johns Hopkins University Press. ISBN 0-8018-5413-X.
- Matrix Iterative Analysis, Varga, Richard S., Springer, 2000.
- Saad, Yousef (2003). Iterative methods for sparse linear systems (2nd ed. ed.). SIAM. ISBN 0898715342. OCLC 51266114
- Higham, N. J. (2008). Functions of matrices: theory and computation. SIAM.
- Higham, N. J. (2002). Accuracy and stability of numerical algorithms. SIAM.
- David S. Watkins (2008), The Matrix Eigenvalue Problem: GR and Krylov Subspace Methods, SIAM.
- Liesen, J., & Strakos, Z. (2012). Krylov subspace methods: principles and analysis. OUP Oxford.
- Claude Brezinski, Gérard Meurant and Michela Redivo-Zaglia (2022). A Journey through the History of Numerical Linear Algebra. SIAM. ISBN 978-1-61197-722-6.
和書
- 仁木滉, & 河野敏行. (1998). 楽しい反復法. 共立出版.
- 河村哲也. (2005). 線形代数と数値解析, 朝倉書店.
- 杉原正顯, 室田一雄, 線形計算の数理, 岩波書店, 2009年.
- 山本哲朗. (2010). 行列解析の基礎–Advanced 線形代数, SGC ライブラリ 79. サイエンス社.
- 線形方程式の反復解法、一般社団法人 日本計算工学会 編、藤野清次 著、阿部邦美 著、杉原正顯 著、丸善出版、2013年09月。
- 「固有値分解装置、及び固有値分解法」 特許5017666(日本)、PCT/JP2007/051575
- 「特異値分解装置、及び特異値分解法」 特許5011545(日本)、PCT/JP2006/318713
科学研究費助成事業
ウィキメディア・コモンズには、
数値線形代数に関連するカテゴリがあります。