Home Effective highly accurate time integrators for linear Klein–Gordon equations across the scales
Article
Licensed
Unlicensed Requires Authentication

Effective highly accurate time integrators for linear Klein–Gordon equations across the scales

  • Karolina Kropielnicka ORCID logo , Karolina Lademann ORCID logo EMAIL logo and Katharina Schratz
Published/Copyright: September 11, 2024

Abstract

We propose an efficient approach for time integration of Klein–Gordon equations with highly oscillatory in time input terms. The new methods are highly accurate in the entire range, from slowly varying up to highly oscillatory regimes. Our approach is based on splitting methods tailored to the structure of the input term which allows us to resolve the oscillations in the system uniformly in all frequencies, while the error constant does not grow as the oscillations increase. Numerical experiments highlight our theoretical findings and demonstrate the efficiency of the new schemes.

MSC 2010 Classification: 65-02

Corresponding author: Karolina Lademann, Institute of Mathematics, Physics and Computer Science, University of Gdańsk, Gdańsk, Poland, E-mail: 

Funding source: European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme

Award Identifier / Grant number: 850941

Funding source: National Science Centre (NCN)

Award Identifier / Grant number: 2019/34/E/ST1/00390

Funding source: Institute of Mathematics of the Polish Academy of Sciences for the years 2021–2023

Award Identifier / Grant number: 663281

Funding source: Academic Computer Center in Gdańsk

Acknowledgment

We are grateful to Arieh Iserles for his friendship, encouragement and invaluable editorial support. The authors with to thank the Isaac Newton Institute for Mathematical Sciences for support and hospitality during the programme “Geometry, compatibility and structure preservation in computational differential equations”, supported by EPSRC grant EP/R014604/1, where this work has been initiated.

  1. Research ethics: The local Institutional Review Board deemed the study exempt from review.

  2. Author contributions: All authors have accepted responsibility for the entire content of this manuscript and approved its submission.

  3. Competing interests: Authors state no conflict of interest.

  4. Research funding: The work of Katharina Schratz in this project was funded by the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement No.850941). The work of Karolina Kropielnicka and Karolina Lademann in this project was funded by the National Science Centre (NCN) project no. 2019/34/E/ ST1/00390. Numerical simulations were carried out by Karolina Lademann at the Academic Computer Center in Gdańsk (CI TASK). This work was partially financed by Simons Foundation Award No. 663281 granted to the Institute of Mathematics of the Polish Academy of Sciences for the years 2021–2023.

  5. Data availability: Not applicable.

Appendix A: Simplification in Equations (2.4)(2.7)

Let us recall that f ( t ) = α ( t ) + | n | N a n ( t ) e i ω n t and

A ( t 1 ) = 0 1 Δ + f ( t 1 ) 0 , [ A ( t 2 ) , A ( t 1 ) ] = f ( t 1 ) f ( t 2 ) 0 0 f ( t 2 ) f ( t 1 ) .

In the following part we are calculating twofold and threefold nested commutators:

A ( t 1 ) , A ( t 2 ) , A ( t 3 ) = 0 1 Δ + f ( t 1 ) 0 f ( t 3 ) f ( t 2 ) 0 0 f ( t 2 ) f ( t 3 ) f ( t 3 ) f ( t 2 ) 0 0 f ( t 2 ) f ( t 3 ) 0 1 Δ + f ( t 1 ) 0 = 0 f ( t 2 ) f ( t 3 ) Δ + f ( t 1 ) f ( t 3 ) f ( t 2 ) 0 0 f ( t 3 ) f ( t 2 ) f ( t 2 ) f ( t 3 ) Δ + f ( t 1 ) 0 = 0 H 1 H 2 0 ,

where

H 1 = 2 f ( t 2 ) f ( t 3 ) , H 2 = Δ + f ( t 1 ) f ( t 3 ) f ( t 2 ) f ( t 2 ) f ( t 3 ) Δ + f ( t 1 ) = Δ f ( t 3 ) + f ( t 3 ) Δ Δ f ( t 2 ) f ( t 2 ) Δ + 2 f ( t 1 ) f ( t 3 ) 2 f ( t 1 ) f ( t 2 ) .

Analogously

A ( t 3 ) , A ( t 2 ) , A ( t 1 ) = 0 H 3 H 4 0 ,

where

H 3 = 2 f ( t 2 ) f ( t 1 ) ,

H 4 = Δ f ( t 1 ) + f ( t 1 ) Δ Δ f ( t 2 ) f ( t 2 ) Δ + 2 f ( t 3 ) f ( t 1 ) 2 f ( t 3 ) f ( t 2 ) .

For the threefold nested commutators we have

A ( t 4 ) , A ( t 1 ) , A ( t 2 ) , A ( t 3 ) = 0 1 Δ + f ( t 4 ) 0 0 H 5 H 6 0 0 H 5 H 6 0 0 1 Δ + f ( t 4 ) 0 = H 6 0 0 Δ + f ( t 4 ) H 5 H 5 Δ + f ( t 4 ) 0 0 H 6 = H 6 H 5 Δ + f ( t 4 ) 0 0 Δ + f ( t 4 ) H 5 H 6 ,

where

H 5 = 2 f ( t 2 ) f ( t 1 ) ,

H 6 = Δ f ( t 1 ) f ( t 2 ) + f ( t 1 ) f ( t 2 ) Δ + 2 f ( t 3 ) f ( t 1 ) f ( t 2 ) .

Likewise,

A ( t 1 ) , A ( t 2 ) , A ( t 3 ) , A ( t 4 ) = H 8 H 7 Δ + f ( t 1 ) 0 0 Δ + f ( t 1 ) H 7 H 8

with

H 7 = 2 f ( t 3 ) f ( t 2 ) ,

H 8 = Δ f ( t 2 ) f ( t 3 ) + f ( t 2 ) f ( t 3 ) Δ + 2 f ( t 4 ) f ( t 2 ) f ( t 3 ) .

Finally,

A ( t 1 ) , A ( t 2 ) , A ( t 3 ) , A ( t 4 ) = 0 1 Δ + f ( t 1 ) 0 0 H 9 H 10 0 0 H 9 H 10 0 0 1 Δ + f ( t 1 ) 0 = H 10 0 0 Δ + f ( t 4 ) H 9 H 9 Δ + f ( t 4 ) 0 0 H 10 = H 10 H 9 Δ + f ( t 1 ) 0 0 Δ + f ( t 1 ) H 9 H 10 ,

where

H 9 = 2 f ( t 3 ) f ( t 4 ) ,

H 10 = Δ f ( t 4 ) f ( t 3 ) + f ( t 4 ) f ( t 3 ) Δ + 2 f ( t 2 ) f ( t 4 ) f ( t 3 )

and

A ( t 2 ) , A ( t 3 ) , A ( t 4 ) , A ( t 1 ) = H 12 H 11 Δ + f ( t 2 ) 0 0 Δ + f ( t 2 ) H 11 H 12 ,

where

H 11 = 2 f ( t 4 ) f ( t 1 ) ,

H 12 = Δ f ( t 1 ) f ( t 4 ) + f ( t 1 ) f ( t 4 ) Δ + 2 f ( t 3 ) f ( t 1 ) f ( t 4 ) .

To estimate term Θ4 we need to aggregate the matrices originating in individual commutators. The result is a matrix H 1 0 0 H 2 , where

H 1 = H 6 + H 5 Δ + f ( t 4 ) + H 8 H 7 Δ + f ( t 1 ) + H 10 H 9 Δ + f ( t 4 ) + H 12 H 11 Δ + f ( t 2 ) = Δ f ( t 2 ) f ( t 1 ) + f ( t 2 ) f ( t 1 ) Δ + 2 f ( t 3 ) f ( t 2 ) f ( t 1 ) 2 f ( t 1 ) f ( t 2 ) Δ + f ( t 4 ) + Δ f ( t 2 ) f ( t 3 ) + f ( t 2 ) f ( t 3 ) Δ + 2 f ( t 4 ) f ( t 2 ) f ( t 3 ) 2 f ( t 3 ) f ( t 2 ) Δ + f ( t 1 ) + Δ f ( t 4 ) f ( t 3 ) + f ( t 4 ) f ( t 3 ) Δ + 2 f ( t 2 ) f ( t 4 ) f ( t 3 ) 2 f ( t 3 ) f ( t 4 ) Δ + f ( t 1 ) + Δ f ( t 1 ) f ( t 4 ) + f ( t 1 ) f ( t 4 ) Δ + 2 f ( t 3 ) f ( t 1 ) f ( t 4 ) 2 f ( t 4 ) f ( t 1 ) Δ + f ( t 2 ) = Δ 2 f ( t 2 ) 2 f ( t 3 ) + 3 2 f ( t 2 ) 2 f ( t 3 ) Δ + 4 f ( t 4 ) f ( t 2 ) f ( t 3 ) + 4 f ( t 1 ) f ( t 2 ) f ( t 3 )

and

H 2 = 3 Δ 2 f ( t 2 ) 2 f ( t 3 ) 2 f ( t 2 ) 2 f ( t 3 ) Δ 4 f ( t 4 ) f ( t 2 ) f ( t 3 ) 4 f ( t 1 ) f ( t 2 ) f ( t 3 ) ,

where

f ( t k ) f ( t l ) = α ( t k ) + | n | N a n ( t k ) e i ω n t k α ( t l ) + | n | N a n ( t l ) e i ω n t l = α ( t k ) α ( t l ) + α ( t k ) | n | N a n ( t l ) e i ω n t l + α ( t l ) | n | N a n ( t k ) e i ω n t k + | n | N a n ( t k ) e i ω n t k | m | N a m ( t l ) e i ω m t l .

Appendix B: Simplification in Strang splitting error in Section 3.2

Taking

X = 0 h 0 t 1 [ A ( t + t 2 ) , A ( t + t 1 ) ] d t 1 d t 2 = F 0 0 F ,

Y = 0 h 0 1 Δ + f ( t + t 1 ) 0 d t 1 = 0 h h Δ + F 0 ,

we have

[ Y , X ] = 0 h h Δ + F 0 F 0 0 F F 0 0 F 0 h h Δ + F 0 = 0 h F ( h Δ + F ) F 0 0 h F F ( h Δ + F ) 0 = 0 2 h F h Δ F h F Δ 2 F F 0 , [ Y , [ Y , X ] ] = 0 h h Δ + F 0 0 2 h F h Δ F h F Δ 2 F F 0 0 2 h F h Δ F h F Δ 2 F F 0 0 h h Δ + F 0 = h 2 Δ F 3 h 2 F Δ 4 h F F 0 0 3 h 2 Δ F + h 2 F Δ + 4 h F F , [ X , Y ] = [ Y , X ] = 0 2 h F h Δ F + h F Δ + 2 F F 0 , [ X , [ X , Y ] ] = F 0 0 F 0 2 h F h Δ F + h F Δ + 2 F F 0 0 2 h F h Δ F + h F Δ + 2 F F 0 F 0 0 F = 0 4 h F 2 2 h F Δ F + h F 2 Δ + h Δ F 2 + 4 F 2 F 0 .

References

[1] A. Abdulle, E. Weinan, B. Engquist, and E. Vanden-Eijnden, “The heterogeneous multiscale method,” Acta Numer., vol. 21, pp. 1–87, 2012, https://doi.org/10.1017/S0962492912000025.Search in Google Scholar

[2] A. Deaño, D. Huybrechs, and A. Iserles, Computing Highly Oscillatory Integrals, Philadelphia, PA, Society for Industrial and Applied Mathematics (SIAM), 2018.10.1137/1.9781611975123Search in Google Scholar

[3] B. Engquist, A. Fokas, E. Hairer, and A. Iserles, Highly Oscillatory Problems, ser. London Maths Soc. Lecture Note Series, vol. 366, Cambridge University Press, Cambridge, UK, 2009.10.1017/CBO9781139107136Search in Google Scholar

[4] M. Znojil, “Klein–Gordon equation with the time- and space-dependent mass:unitary evolution picture,” Tech. Rep., Nuclear Physics Institute, Czech Academy of Sciences, 2017, arXiv:1702.08493v1.Search in Google Scholar

[5] M. Znojil, “Non-Hermitian interaction representation and its use in relativistic quantum mechanics,” Ann. Phys., vol. 385, pp. 162–179, 2017, https://doi.org/10.1016/j.aop.2017.08.009.Search in Google Scholar

[6] A. Mostafazadeh, “Quantum mechanics of Klein–Gordon-type fields and quantum cosmology,” Ann. Phys., vol. 309, no. 1, pp. 1–48, 2004. https://doi.org/10.1016/j.aop.2003.08.010.Search in Google Scholar

[7] M. Znojil, “Quantization of big bang in crypto-Hermitian Heisenberg picture,” Springer Proc. Phys., vol. 184, pp. 383–399, 2016, https://doi.org/10.1007/978-3-319-31356-6_26.Search in Google Scholar

[8] J.-B. Chen and H. Liu, “Multisymplectic pseudospectral discretizations for (3 + 1)-dimensional Klein–Gordon equation,” Commun. Theor. Phys., vol. 50, no. 5, pp. 1052–1054, 2008. https://doi.org/10.1088/0253-6102/50/5/07.Search in Google Scholar

[9] W. Bao and X. Dong, “Analysis and comparison of numerical methods for the Klein–Gordon equation in the nonrelativistic limit regime,” Numer. Math., vol. 120, no. 2, pp. 189–229, 2012. https://doi.org/10.1007/s00211-011-0411-2.Search in Google Scholar

[10] E. Faou and K. Schratz, “Asymptotic preserving schemes for the Klein–Gordon equation in the non-relativistic limit regime,” Numer. Math., vol. 126, no. 3, pp. 441–469, 2014. https://doi.org/10.1007/s00211-013-0567-z.Search in Google Scholar

[11] F. Shakeri and M. Dehghan, “Numerical solution of the Klein–Gordon equation via He’s variational iteration method,” Nonlinear Dynam., vol. 51, nos. 1–2, pp. 89–97, 2008. https://doi.org/10.1007/s11071-006-9194-x.Search in Google Scholar

[12] E. Yusufoğlu, “The variational iteration method for studying the Klein–Gordon equation,” Appl. Math. Lett., vol. 21, no. 7, pp. 669–674, 2008. https://doi.org/10.1016/j.aml.2007.07.023.Search in Google Scholar

[13] P. Bader, S. Blanes, F. Casas, and N. Kopylov, “Novel symplectic integrators for the Klein–Gordon equation with space- and time-dependent mass,” J. Comput. Appl. Math., vol. 350, pp. 130–138, 2019, https://doi.org/10.1016/j.cam.2018.10.011.Search in Google Scholar

[14] M. Condon, K. Kropielnicka, K. Lademann, and R. Perczyński, “Asymptotic numerical solver for the linear Klein–Gordon equation with space- and time-dependent mass,” Appl. Math. Lett., vol. 115, p. 106935, 2021, https://doi.org/10.1016/j.aml.2020.106935.Search in Google Scholar

[15] K. Kropielnicka and K. Lademann, “Third order, uniform in low to high oscillatory coefficients, exponential integrators for Klein–Gordon equations,” arXiv preprint arXiv:2212.13762, 2022.Search in Google Scholar

[16] A. Iserles, K. Kropielnicka, and P. Singh, “Compact schemes for laser–matter interaction in Schrödinger equation based on effective splittings of Magnus expansion,” Comput. Phys. Commun., vol. 234, pp. 195–201, 2019, https://doi.org/10.1016/j.cpc.2018.07.010.Search in Google Scholar

[17] S. A. Chin and C. Chen, “Gradient symplectic algorithms for solving the Schrödinger equation with time-dependent potentials,” J. Chem. Phys., vol. 117, no. 4, pp. 1409–1415, 2002. https://doi.org/10.1063/1.1485725.Search in Google Scholar

[18] A. Iserles, H. Z. Munthe-Kaas, S. P. Nørsett, and A. Zanna, “Lie-group methods,” in Acta Numerica, 2000, ser. Acta Numer, vol. 9, Cambridge, Cambridge Univ. Press, pp. 215–365, 2000.10.1017/S0962492900002154Search in Google Scholar

[19] A. Iserles, S. P. Nørsett, and A. F. Rasmussen, “Time symmetry and high-order Magnus methods,” Appl. Numer. Math., vol. 39, nos. 3–4, pp. 379–401, 2001. https://doi.org/10.1016/s0168-9274(01)00088-5.Search in Google Scholar

[20] S. Blanes, F. Casas, J. A. Oteo, and J. Ros, “The Magnus expansion and some of its applications,” Phys. Rep., vol. 470, nos. 5–6, pp. 151–238, 2009. https://doi.org/10.1016/j.physrep.2008.11.001.Search in Google Scholar

[21] W. Magnus, “On the exponential solution of differential equations for a linear operator,” Commun. Pure Appl. Math., vol. 7, no. 4, pp. 649–673, 1954. https://doi.org/10.1002/cpa.3160070404.Search in Google Scholar

[22] T. Jahnke and C. Lubich, “Error bounds for exponential operator splittings,” BIT Numer. Math., vol. 40, pp. 735–744, 2000. https://doi.org/10.1023/a:1022396519656.10.1023/A:1022396519656Search in Google Scholar

[23] A. Zanna, “The Fer expansion and time-symmetry: a Strang-type approach,” Appl. Numer. Math., vol. 39, nos. 3–4, pp. 435–459, 2001. https://doi.org/10.1016/s0168-9274(01)00085-x.Search in Google Scholar

[24] S. Blanes, F. Casas, and J. Ros, “High order optimized geometric integrators for linear differential equations,” BIT Numer. Math., vol. 42, pp. 262–284, 2002.10.1023/A:1021942823832Search in Google Scholar

[25] D. A. Kopriva, Implementing Spectral Methods for Partial Differential Equations, ser. Scientific Computation, Algorithms for Scientists and Engineers, Berlin, Springer, 2009.10.1007/978-90-481-2261-5Search in Google Scholar

[26] L. N. Trefethen, Spectral Methods in MATLAB, Philadelphia, SIAM, 2000.10.1137/1.9780898719598Search in Google Scholar

Received: 2023-05-25
Accepted: 2024-06-26
Published Online: 2024-09-11
Published in Print: 2025-06-26

© 2024 Walter de Gruyter GmbH, Berlin/Boston

Downloaded on 26.11.2025 from https://www.degruyterbrill.com/document/doi/10.1515/jnma-2023-0070/pdf
Scroll to top button