Previous |  Up |  Next

Article

Title: The least squares solution of inconsistent discretized elliptic problems using the FETI method (English)
Author: Dostál, Zdeněk
Author: Horák, David
Language: English
Journal: Applications of Mathematics
ISSN: 0862-7940 (print)
ISSN: 1572-9109 (online)
Volume: 70
Issue: 6
Year: 2025
Pages: 929-939
Summary lang: English
.
Category: math
.
Summary: The variants of FETI (finite element tearing and interconnecting) based domain decomposition methods are well-established, massively parallel algorithms for solving huge linear systems arising from the discretization of elliptic partial differential equations. Here, we adapt the FETI method for solving the large least squares problems associated with inconsistent systems of linear equations arising from the discretization of elliptic partial differential equations. We briefly review the symmetric least squares problems and the FETI method, explain how FETI can find the least squares solution, prove the optimal rate of convergence, and present the results of numerical experiments demonstrating the efficiency of the proposed method in solving the least squares problem defined by the Poisson equation with inconsistent Neumann conditions. (English)
Keyword: least square problem
Keyword: domain decomposition
Keyword: redundant multipliers
MSC: 65K15
MSC: 65Y05
MSC: 90C06
DOI: 10.21136/AM.2025.0189-25
.
Date available: 2025-12-20T06:45:14Z
Last updated: 2025-12-22
Stable URL: http://hdl.handle.net/10338.dmlcz/153229
.
Reference: [1] Arbenz, P., Drmač, Z.: On positive semidefinite matrices with known null space.SIAM J. Matrix Anal. Appl. 24 (2002), 132-149. Zbl 1032.65025, MR 1920559, 10.1137/S0895479800381331
Reference: [2] Björck, \r A.: Numerical Methods for Least Squares Problems.SIAM, Philadelphia (1996). Zbl 0847.65023, MR 1386889, 10.1137/1.9781611971484
Reference: [3] Blaheta, R., Jakl, O., Starý, J., Krečmer, K.: Schwarz DD method for analysis of geocomposites.Proceedings of the 12th International Conference on Civil, Structural and Environmental Engineering Computing Civil-Comp Press, Stirlingshire (2009), Article ID 111.
Reference: [4] Calvetti, D., Reichel, L., Zhang, Q.: Conjugate gradient algorithms for symmetric inconsistent linear systems.Proceedings of the Lanczos Centenary Conference SIAM, Philadelphia (1994), 267-272.
Reference: [5] Choi, C. T., Saunders, M. A.: Algorithm 937: MINRES-QLP for symmetric and Hermitian linear equations and least-squares problems.ACM Trans. Math. Softw. 40 (2014), Article ID 16, 12 pages \99999DOI99999 10.1145/252726 . Zbl 1305.65117, MR 3181906
Reference: [6] Dostál, Z., Brzobohatý, T., Horák, D., Kružík, J., Vlach, O.: Scalable hybrid TFETI-DP methods for large boundary variational inequalities.Domain Decomposition Methods in Science and Engineering XXVI Lecture Notes in Computational Science and Engineering 145. Springer, Cham (2022), 29-40. Zbl 07936273, MR 4703833, 10.1007/978-3-030-95025-5_3
Reference: [7] Dostál, Z., Horák, D., Brzobohatý, T., Vodstrčil, P.: Bounds on the spectra of Schur complements of large H-TFETI-DP clusters for 2D Laplacian.Numer. Linear Algebra Appl. 28 (2021), Article ID e2344, 15 pages. Zbl 1549.65522, MR 4211729, 10.1002/nla.2344
Reference: [8] Dostál, Z., Horák, D., Kučera, R.: Total FETI -- an easier, implementable variant of the FETI method for the numerical solution of elliptic PDE.Commun. Numer. Methods Eng. 22 (2006), 1155-1162. Zbl 1107.65104, MR 2282408, 10.1002/cnm.881
Reference: [9] Dostál, Z., Kozubek, T., Markopoulos, A., Menšík, M.: Cholesky decomposition of a positive semidefinite matrix with known kernel.Appl. Math. Comput. 217 (2011), 6067-6077. Zbl 1211.65034, MR 2773349, 10.1016/j.amc.2010.12.069
Reference: [10] Dostál, Z., Pospíšil, L.: Conjugate gradients for symmetric positive semidefinite least-squares problems.Int. J. Comput. Math. 95 (2018), 2229-2239. Zbl 1499.65114, MR 3844674, 10.1080/00207160.2017.1371701
Reference: [11] Farhat, C., Mandel, J., Roux, F.-X.: Optimal convergence properties of the FETI domain decomposition method.Comput. Methods Appl. Mech. Eng. 115 (1994), 365-385 \99999DOI99999 10.1016/0045-7825(94)90068-X . MR 1285024
Reference: [12] Farhat, C., Roux, F.-X.: A method of finite element tearing and interconnecting and its parallel solution algorithm.Int. J. Numer. Methods Eng. 32 (1991), 1205-1227 \99999DOI99999 10.1002/nme.1620320604 . Zbl 0758.65075, MR 3618550
Reference: [13] Farhat, C., Roux, F.-X.: An unconventional domain decomposition method for an efficient parallel solution of large-scale finite element systems.SIAM J. Sci. Stat. Comput. 13 (1992), 379-396. Zbl 0746.65086, MR 1145192, 10.1137/0913020
Reference: [14] Fong, D. C.-L., Saunders, M. A.: LSMR: An iterative algorithm for sparse least-squares problems.SIAM J. Sci. Comput. 33 (2011), 2950-2971. Zbl 1232.65052, MR 2861656, 10.1137/10079687X
Reference: [15] Karypis, G., Kumar, V.: METIS: Serial Graph Partitioning and Fill-reducing Matrix Ordering.Available at https://github.com/KarypisLab/METIS.
Reference: [16] Klawonn, A., Rheinbach, O.: A hybrid approach to 3-level FETI.PAMM, Proc. Appl. Math. Mech. 8 (2008), 10841-10843. Zbl 1392.65114, 10.1002/pamm.200810841
Reference: [17] Kruis, J.: Domain Decomposition Methods for Distributed Computing.Saxe-Cobourg Publications, Stirling (2006).
Reference: [18] Paige, C. C., Saunders, M. A.: LSQR: An algorithm for sparse linear equations and sparse least squares.ACM Trans. Math. Softw. 8 (1982), 43-71. Zbl 0478.65016, MR 0661121, 10.1145/355984.3559
Reference: [19] Saad, Y.: Iterative Methods for Sparse Linear Systems.SIAM, Philadelphia (2003). Zbl 1031.65046, MR 1990645, 10.1137/1.9780898718003
Reference: [20] Toselli, A., Widlund, O. B.: Domain Decomposition Methods: Algorithms and Theory.Springer Series on Computational Mathematics 34. Springer, Berlin (2005). Zbl 1069.65138, MR 2104179, 10.1007/b137868
Reference: [21] Xin, K., Meng, L.: Block-splitting preconditioners for indefinite least squares problem.Comput. Appl. Math. 44 (2025), Article ID 61, 17 pages. Zbl 1563.65174, MR 4834472, 10.1007/s40314-024-03016-7
.

Fulltext not available (moving wall 24 months)

Partner of
EuDML logo