The numerical solution of an ordinary differential equation can be interpreted as the exact solution of a nearby modified equation. Investigating the behaviour of numerical solutions by analysing the modified equation is known as backward error analysis. If the original and modified equation share structural properties, then the exact and approximate solution share geometric features such as the existence of conserved quantities. Conjugate symplectic methods preserve a modified symplectic form and a modified Hamiltonian when applied to a Hamiltonian system. We show how a blended version of variational and symplectic techniques can be used to compute modified symplectic and Hamiltonian structures. In contrast to other approaches, our backward error analysis method does not rely on an ansatz but computes the structures systematically, provided that a variational formulation of the method is known. The technique is illustrated on the example of symmetric linear multistep methods with matrix coefficients.
Citation: Robert I McLachlan, Christian Offen. Backward error analysis for conjugate symplectic methods[J]. Journal of Geometric Mechanics, 2023, 15(1): 98-115. doi: 10.3934/jgm.2023005
The numerical solution of an ordinary differential equation can be interpreted as the exact solution of a nearby modified equation. Investigating the behaviour of numerical solutions by analysing the modified equation is known as backward error analysis. If the original and modified equation share structural properties, then the exact and approximate solution share geometric features such as the existence of conserved quantities. Conjugate symplectic methods preserve a modified symplectic form and a modified Hamiltonian when applied to a Hamiltonian system. We show how a blended version of variational and symplectic techniques can be used to compute modified symplectic and Hamiltonian structures. In contrast to other approaches, our backward error analysis method does not rely on an ansatz but computes the structures systematically, provided that a variational formulation of the method is known. The technique is illustrated on the example of symmetric linear multistep methods with matrix coefficients.
| [1] |
P. Chartier, E. Faou, A. Murua. An algebraic approach to invariant preserving integators: The case of quadratic and Hamiltonian invariants. Numer. Math, 103 (2006), 575–590. https://doi.org/10.1007/s00211-006-0003-8 doi: 10.1007/s00211-006-0003-8
|
| [2] | C. L. Ellison, J. W. Burby, J. M. Finn, H. Qin, W. M. Tang, Initializing and stabilizing variational multistep algorithms for modeling dynamical systems, 2014, arXiv preprint arXiv. |
| [3] | E. Hairer, C. Lubich, G. Wanner. Geometric Numerical Integration: Structure-Preserving Algorithms for Ordinary Differential Equations. Springer Series in Com- putational Mathematics, 2013 https://doi.org/10.1007/3-540-30666-8 |
| [4] |
L. Jódar, J. L. Morera, E. Navarro, On convergent linear multistep matrix methods, Int. J. Comput. Math, 40 (1991), 211–219. https://doi.org/10.1080/00207169108804014 doi: 10.1080/00207169108804014
|
| [5] |
J. D. Lambert, S. Sigurdsson, Multistep methods with variable matrix coefficients, Siam. J. Numer. Anal, 9 (1972), 715–733. https://doi.org/10.1137/0709060 doi: 10.1137/0709060
|
| [6] |
B. Leimkuhler, S. Reich, Simulating Hamiltonian Dynamics, Cambridge University Press, 2 (2005). https://doi.org/10.1017/CBO9780511614118 doi: 10.1017/CBO9780511614118
|
| [7] |
P. Libermann, C.M. Marle, Symplectic manifolds and Poisson manifolds, In Symplectic Geometry and Analytical Mechanics., (1987) 89–184. https://doi.org/10.1007/978-94-009-3807-6-3 doi: 10.1007/978-94-009-3807-6-3
|
| [8] |
J.E. Marsden, M. West. Discrete mechanics and variational integrators. Acta Numerica, 10 (2001), 357–514. https://doi.org/10.1017/S096249290100006X doi: 10.1017/S096249290100006X
|
| [9] |
R. I. McLachlan, C. Offen, Backward error analysis for variational discretisa-tions of PDEs, J. Geom. Mech, 14 (2022), 447. https://doi.org/10.3934/jgm.2022014 doi: 10.3934/jgm.2022014
|
| [10] |
S. Ober-Blöbaum, C. Offen, Variational learning of Euler-Lagrange dynam- ics from data. J. Comput. Appl. Math, (2022), 114780. https://doi.org/10.1016/j.cam.2022.114780 doi: 10.1016/j.cam.2022.114780
|
| [11] | C. Offen, Release of GitHub repository Christian-Offen/BEAConjugateSymplectic, Available from: https://github.com/Christian-Offen/BEAConjugateSymplectic. |
| [12] | C. Offen, S. Ober-Blöbaum, Symplectic integration of learned Hamiltonian systems. Poster presented at the workshop Theory of deep learning (9 August 2021 to 13 August 2021), Isaac Newton Institute, University of Cambridge, Cambridge, UK, 2021. Available from: https://github.com/Christian-Offen/symplectic-shadow-integration/raw/master/PosterWorkshopCambridge.pdf. |
| [13] |
C. Offen, S. Ober-Blöbaum, Symplectic integration of learned Hamiltonian systems. Chaos: An Interdisciplinary Journal of Nonlinear Science, 32 (2022). https://doi.org/10.1063/5.0065913 doi: 10.1063/5.0065913
|
| [14] | S. Sigurdsson, Linear multistep methods with variable matrix coefficients, In Lecture Notes in Mathematics, (1971), 327-331. Springer Berlin Heidelberg. https://doi.org/10.1007/BFb0069468 |
| [15] |
M. Vermeeren, Modified equations for variational integrators, Numer. Math, 137 (2017), 1001–1037. https://doi.org/10.1007/s00211-017-0896-4 doi: 10.1007/s00211-017-0896-4
|
| [16] | E. T. Whittaker, S. W. McCrae, Hamiltonian systems and their integral invariants, Cambridge Mathematical Library, Cambridge University Press, https://doi.org/10.1017/CBO9780511608797.012 |
Figures(2)
Robert I McLachlan, Christian Offen. Backward error analysis for conjugate symplectic methods[J]. Journal of Geometric Mechanics, 2023, 15(1): 98-115. doi: 10.3934/jgm.2023005
DownLoad: