Skip to main content

Advertisement

SpringerLink
Log in

Highly scalable hybrid domain decomposition method for the solution of huge scalar variational inequalities

  • Original Paper
  • Published:
Numerical Algorithms Aims and scope Submit manuscript

Abstract

The unpreconditioned hybrid domain decomposition method was recently shown to be a competitive solver for linear elliptic PDE problems discretized by structured grids. Here, we plug H-TFETI-DP (hybrid total finite element tearing and interconnecting dual primal) method into the solution of huge boundary elliptic variational inequalities. We decompose the domain into subdomains that are discretized and then interconnected partly by Lagrange multipliers and partly by edge averages. After eliminating the primal variables, we get a quadratic programming problem with a well-conditioned Hessian and bound and equality constraints that is effectively solvable by specialized algorithms. We prove that the procedure enjoys optimal, i.e., asymptotically linear complexity. The analysis uses recently established bounds on the spectrum of the Schur complements of the clusters interconnected by edge/face averages. The results extend the scope of scalability of massively parallel algorithms for the solution of variational inequalities and show the outstanding efficiency of the H-TFETI-DP coarse grid split between the primal and dual variables.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6

Similar content being viewed by others

References

  1. Axelsson, O.: Iterative Solution Methods. Cambridge University Press, Cambridge (1994)

    Book  Google Scholar 

  2. Brenner, S. C.: The condition number of the Schur complement. Numer. Math. 83, 187–203 (1999)

    Article  MathSciNet  Google Scholar 

  3. Brzobohatý, T., Dostál, Z., Kozubek, T., Kovář, P., Markopoulos, A.: Cholesky decomposition with fixing nodes to the stable computation of a generalized inverse of the stiffness matrix of a floating structure. Int. J. Numer. Methods Eng. 88(5), 493–509 (2011)

    Article  MathSciNet  Google Scholar 

  4. Brzobohatý, T., Jarošová, M., Kozubek, T., Menšík, M., Markopoulos, A.: The hybrid total FETI method. In: Proceedings of the Third International Conference on Parallel, Distributed, Grid, and Cloud Computing for Engineering Civil-Comp (2013)

  5. Conn, A. R., Gould, N. I. M., Toint, P. L.: A globally convergent augmented Lagrangian algorithm for optimization with general constraints and simple bounds. SIAM J. Numer. Anal. 28, 545–572 (1991)

    Article  MathSciNet  Google Scholar 

  6. Dostál, Z.: Conjugate gradient method with preconditioning by projector. Int. J. Comput. Math. 23, 315–324 (1988)

    Article  Google Scholar 

  7. Dostál, Z.: A proportioning based algorithm with rate of convergence for bound constrained quadratic programming. Numer. Algorithms 34(2–4), 293–302 (2003)

    Article  MathSciNet  Google Scholar 

  8. Dostál, Z.: An optimal algorithm for bound and equality constrained quadratic programming problems with bounded spectrum. Computing 78, 311–328 (2006)

    Article  MathSciNet  Google Scholar 

  9. Dostál, Z.: On the decrease of a quadratic function along the projected–gradient path. ETNA 31, 25–59 (2008)

    MathSciNet  MATH  Google Scholar 

  10. Dostál, Z.: Optimal quadratic programming algorithms, with applications to variational inequalities, 1st edn. Springer, New York (2009)

    MATH  Google Scholar 

  11. Dostál, Z., Horák, D.: Theoretically supported scalable FETI for numerical solution of variational inequalities. SIAM J. Numer. Anal. 45(2), 500–513 (2007)

    Article  MathSciNet  Google Scholar 

  12. Dostál, Z., Schöberl, J.: Minimizing quadratic functions subject to bound constraints with the rate of convergence and finite termination. Comput. Opt. Appl. 30(1), 23–44 (2005)

    Article  MathSciNet  Google Scholar 

  13. Dostál, Z., Vlach, O.: An accelerated augmented Lagrangian algorithm with adaptive orthogonalization strategy for bound and equality constrained quadratic programming and its application to large-scale contact problems of elasticity. J. Comput. Appl. Math. 394(1), 113565 (2021)

    Article  MathSciNet  Google Scholar 

  14. Dostál, Z., Gomes, F.A.M., Santos, S.A.: Duality based domain decomposition with natural coarse space for variational inequalities. J. Comput. Appl. Math. 126(1–2), 397–415 (2000)

    Article  MathSciNet  Google Scholar 

  15. Dostál, Z., Friedlander, A., Santos, SA: Augmented Lagrangians with adaptive precision control for quadratic programming with simple bounds and equality constraints. SIAM J. Optim. 13(4), 1120–1140 (2003)

    Article  MathSciNet  Google Scholar 

  16. Dostál, Z., Horák, D., Kučera, R.: Total FETI—an easier implementable variant of the FETI method for numerical solution of elliptic PDE. Commun. Numer. Methods Eng. 22, 1155–1162 (2006)

    Article  MathSciNet  Google Scholar 

  17. Dostál, Z., Kozubek, T., Vlach, O.: Reorthogonalization based stiffness preconditioning in FETI algorithms with applications to variational inequalities. Numer. Lin. Algebra Appl. 22(6), 987–998 (2015)

    Article  MathSciNet  Google Scholar 

  18. Dostál, Z., Kozubek, T., Sadowská, M., Vondrák, V.: Scalable Algorithms for Contact Problems AMM, vol. 36. Springer, New York (2016)

    Book  Google Scholar 

  19. Dostál, Z., Horák, D., Sojka, R.: On the efficient reconstruction of displacements in FETI methods for contact problems. Adv. Electr. Electron. Eng. 15, 237–241 (2017)

    Google Scholar 

  20. Dostál, Z., Horák, D., Brzobohatý, T., Vodstrčil, P.: Bounds on the spectra of Schur complements of large H-TFETI clusters for 2D Laplacian and applications. Numer. Lin. Agebra Appl. https://doi.org/10.1002/nla.2344https://doi.org/10.1002/nla.2344

  21. Dostál, Z, Brzobohatý, T, Vlach, O: Schur complement spectral bounds for large hybrid FETI-DP clusters and huge three-dimensional scalar problems. J. Numer. Math. (2021)

  22. ESPRESO—Highly Parallel Framework for Engineering Applications. http://numbox.it4i.cz

  23. Farhat, C., Roux, F.-X.: A method of finite element tearing and interconnecting and its parallel solution algorithm. Int. J. Numer. Methods Eng. 32, 1205–1227 (1991)

    Article  MathSciNet  Google Scholar 

  24. Farhat, C., Roux, F.-X.: An unconventional domain decomposition method for an efficient parallel solution of large-scale finite element systems. SIAM J. Sci. Comput. 13, 379–396 (1992)

    Article  MathSciNet  Google Scholar 

  25. Farhat, C., Mandel, J., Roux, F. -X.: Optimal convergence properties of the FETI domain decomposition method. Comput. Methods Appl. Mech. Eng. 115, 365–385 (1994)

    Article  MathSciNet  Google Scholar 

  26. Farhat, C., Lesoinne, M., Pierson, K.: A scalable dual-primal domain decomposition method. Numer. Lin. Algebra Appl. 7(7–8), 687–714 (2000)

    Article  MathSciNet  Google Scholar 

  27. Hlaváček, I., Haslinger, J., Nečas, J., Lovíšek, J.: Solution of variational inequalities in mechanics. Springer, Berlin (1988). Topics in Matrix Analysis. Cambridge University Press, Cambridge (1991)

    Book  Google Scholar 

  28. Jarošová, M., Klawonn, A., Rheinbach, O.: Projector preconditioning and transformation of basis in FETI-DP algorithms for contact problem. Math. Comput. Simul. 82(10), 1894–1907 (2012)

    Article  MathSciNet  Google Scholar 

  29. Klawonn, A., Rheinbach, O.: A parallel implementation of dual-primal FETI methodsfor three dimensional linear elasticity using a transformation of basis. SIAM J. Sci. Comput. 28(5), 1886–1906 (2006)

    Article  MathSciNet  Google Scholar 

  30. Klawonn, A., Rheinbach, O.: A hybrid approach to 3-level FETI. Proc. Appl. Math. Mech. 90(1), 10841–10843 (2008)

    Article  Google Scholar 

  31. Klawonn, A., Rheinbach, O.: Highly scalable parallel domain decomposition methods with an application to biomechanics. Z. Angew. Math. Mech. 90(1), 5–32 (2010)

    Article  MathSciNet  Google Scholar 

  32. Klawonn, A., Widlund, O.: Dual–primal FETI method for linear elasticity. Commun. Pure Appl. Anal. LIX, 1523–1572 (2006)

    MathSciNet  MATH  Google Scholar 

  33. Klawonn, A., Lanser, M., Rheinbach, O.: Toward extremally scalable nonlinear domain decomposition methods for elliptic partiall differential equations. SIAM J. Sci. Comput. (37) 6, C667–C696 (2015)

    Article  Google Scholar 

  34. Kružík, J., Horák, D., Čermák, M., Pospísil, L., Pecha, M.: Active set expansion strategies in MPRGP algorithm. Adv. Eng. Softw. 149(1028), 95 (2020)

    Google Scholar 

  35. Lee, J.: Domain decomposition methods for auxiliary linear problems of an elliptic variational inequality. In: Bank, R., et al (eds.) Domain Decomposition Methods in Science and Engineering XX, Lecture Notes in Computational Science and Engineering, vol. 91, pp 319–326 (2013)

  36. Lee, J.: Two domain decomposition methods for auxiliary linear problems for a multibody variational inequality. SIAM J. Sci. Comput. 35(3), 1350–1375 (2013)

    Article  MathSciNet  Google Scholar 

  37. Li, J., Widdlund, O. B.: FETI-DP, BDDC,and block Cholesky method. Int. J. Num. Methods Eng. 66, 250–271 (2006)

    Article  MathSciNet  Google Scholar 

  38. Pechstein, C.: Finite and boundary element tearing and interconnecting solvers for multiscale problems. Springer, Heidelberg (2013)

    Book  Google Scholar 

  39. PERMON—Parallel, efficient, robust, modular, object-oriented numerical software toolbox. http://permon.vsb.cz/

  40. Toselli, A., Widlund, O. B.: Domain decomposition methods—algorithms and theory Springer Series on Computational Mathematics, vol. 34. Springer, Berlin (2005)

    Book  Google Scholar 

  41. Vodstrčil, P., Bouchala, J., Jarošová, M., Dostál, Z.: On conditioning of Schur complements of h-TFETI clusters for 2D problems governed by Laplacian. Appl. Math. 62(6), 699–718 (2017)

    Article  MathSciNet  Google Scholar 

Download references

Funding

This work was supported by The Ministry of Education, Youth, and Sports from the National Programme of Sustainability (NPS II) project “IT4Innovations excellence in science - LQ1602” and by the IT4Innovations infrastructure which is supported from the Large Infrastructures for Research, Experimental Development and Innovations project “e-INFRA CZ– LM2018140.” The second and third authors were supported by the grants 19-11441S of the Czech Science Foundation (GACR), SGS SP2021/103 of the VŠB – Technical University of Ostrava and “RRC/10/2019” Support for Science and Research in the Moravia–Silesia Region 2.

Author information

Authors and Affiliations

  1. Department of Applied Mathematics and IT4Innovations-National Supercomputing Center, VSB-Technical University of Ostrava, 17. listopadu 15, 70833, Ostrava, Czech Republic

    Zdeněk Dostál & Oldřich Vlach

  2. Department of Applied Mathematics, VSB-TU Ostrava and Institue of Geonics ASCR, Ostrava, Czech Republic

    David Horák & Jakub Kružík

  3. IT4Innovations-National Supercomputing Center, VSB-TU Ostrava, Ostrava, Czech Republic

    Tomáš Brzobohatý

Authors
  1. Zdeněk Dostál
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. David Horák
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Jakub Kružík
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Tomáš Brzobohatý
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Oldřich Vlach
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Zdeněk Dostál.

Ethics declarations

Conflict of interest

The authors no competing interests.

Additional information

Data availability

Data sharing not applicable to this article as no datasets were generated or analyzed during the current study.

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Dostál, Z., Horák, D., Kružík, J. et al. Highly scalable hybrid domain decomposition method for the solution of huge scalar variational inequalities. Numer Algor 91, 773–801 (2022). https://doi.org/10.1007/s11075-022-01281-3

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s11075-022-01281-3

Keywords

Mathematics Subject Classification (2010)

Search

Navigation