Inexact shift-invert Arnoldi method for evolution equations

Abstract

Linear and nonlinear evolution equations with a first order time derivative, such as the heat equation, the Burgers equation, and the reaction diffusion equation have been used to solve problems in various fields of science. Differential algebraic equations of the first order are derived after space discretization. In the simplest case, the computation of one matrix exponential with a special form is required. In the most complex case, the computation of matrix functions related to its exponentials need to be implemented repeatedly. When computing large matrix functions, the Krylov subspace methods is a viable alternative. The most well-known method is the Arnoldi method, but it may require a number of iterations depending on the condition of the matrix. As a solution to this issue, we propose the Inexact Shift-invert Arnoldi method to do this more efficiently. As the result, the numerical solution of evolution equations can be computed efficiently with this method. Pertinent numerical experiments establish the effectiveness of this proposed algorithm. References

• Benzi, M. and Boito, P., Decay properties for functions of matrices over \(C^*\)-algebras. Linear Algebra and its Applications, 456(1): 174–198, 2014. doi:10.1016/j.laa.2013.11.027
• Butcher, J. C. Numerical methods for ordinary differential equations, second edition. John Wiley and Sons, Chichester, England, 2008.
• Caratheodory, C., Uber die gegenseitige Beziehung der Rander bei der konformen Abbildung des Inneren einer Jordanschen Kurve auf einen Kreis. Mathematische Annalen, 73(2): 305–320, 1913. http://gdz.sub.uni-goettingen.de/pdfcache/PPN235181684_0073/PPN235181684_0073___LOG_0029.pdf
• Carra, E. J., Turner, I. W. and Perre, P., A variable-stepsize Jacobian-free exponential integrator for simulating transport in heterogeneous porous media: Application to wood drying. Journal of Computational Physics, 233: 66–82, 2013. doi:10.1016/j.jcp.2012.07.024
• Gang, W., Feng, T. and Yimin, W., An inexact shift-and-invert Arnoldi algorithm for Toeplitz matrix exponential. Numerical Linear Algebra with Applications, 22(4): 777–792, 2015. doi:10.1002/nla.1992
• Gallopoulos, E. and Saad, Y., Efficient solution of parabolic equations by Krylov approximation methods. SIAM Journal on Scientific Statistics, 13(5):1236–1264, 1992. doi:10.1137/0913071
• Gockler, T., Rational Krylov subspace methods for \(\phi \)-functions in exponential integrators. em Karlsruher Instituts fur Technologie, 2014, Ph.D. thesis. http://d-nb.info/1060425408/34
• Guttel, S., Rational Krylov methods for operator functions. em Technischen Universitat Bergakademie Freiberg, 2010, Ph.D. thesis. http://www.qucosa.de/fileadmin/data/qucosa/documents/2764/diss_guettel.pdf
• Hashimoto, Y. and Nodera, T., Inexact shift-invert Arnoldi method for linear evolution equations (Japanese). IPSJ Journal, 57(10): 2250–2259, 2016. https://ipsj.ixsq.nii.ac.jp/ej/?action=pages_view_main&active_action=repository_view_main_item_detail&item_id=175055&item_no=1&page_id=13&block_id=8
• Higham, N. J., The scaling and squaring method for the matrix exponential revisited. SIAM Journal on Matrix Analysis and Applications, 26(4): 1179–1193, 2005. doi:10.1137/04061101X
• Hochbruck, M., A short course on exponential integrators. Series in Contemporary Applied Mathematics, 17: 29–49, 2015. http://www.siam.org/students/g2s3/2013/lecturers/Hochbruck/Summary_Hochbruck.pdf
• Hochbruck, M. and Ostermann, A., Exponential Runge–Kutta methods for parabolic problems. Applied Numerical Mathematics, 53(2–4): 323–339, 2005. doi:10.1016/j.apnum.2004.08.005
• Hochbruck, M. and Ostermann, A., Exponential integrators. Acta Numerica, 19:209–286, 2010. doi:10.1017/S0962492910000048
• Hochbruck, M. and Lubich, C., On Krylov subspace approximations to the matrix exponential Operator. SIAM Journal on Numerical Analysis, 34(5): 1911–1925, 1997. doi:10.1137/S0036142995280572
• Hochbruck, M., Lubich, C. and Selhofer, H., Exponential integrators for large systems of differential equations. SIAM Journal on Scientific Computing, 19(5):1552–1574, 1997. doi:10.1137/S1064827595295337
• Hongqing, Z., Huazhong, S. and Meiyu, D., Numerical solutions of two-dimensional Burgers' equations by discrete Adomian decomposition method. Computers and Mathematics with Applications, 60(3): 840–848, 2010. doi:10.1016/j.camwa.2010.05.03
• Kamel, A. K., Numerical study of Fisher's reaction-diffusion equation by the Sinc collocation method. Journal of Computational and Applied Mathematics, 137(2): 245–255, 2001. doi:10.1016/S0377-0427(01)00356-9
• Lee, S., Pang, H. and Sun, H., Shift-invert Arnoldi approximation to the Toeplitz matrix exponential. SIAM Journal on Scientific Computing, 32(2): 774–792, 2010. doi:10.1137/090758064
• Moret, I. and Novati, P., RD-rational approximations of the matrix exponential. BIT Numerical Mathematics, 44(3): 595–615, 2004. doi:10.1023/B:BITN.0000046805.27551.3b
• Moler, C. and Van Loan, C. F., Nineteen dubious ways to compute the exponential of a matrix, Twenty-Five Years Later. SIAM Review, 45(1): 3–49, 2003. doi:10.1137/S00361445024180
• Novati, P., Using the restricted-denominator rational Arnoldi method for exponential Integrators. SIAM Journal on Matrix Analysis and Applications, 32(4): 1537–1558, 2011. doi:10.1137/100814202
• Svoboda, Z., The convective-diffusion equation and its use in building physics. International Journal on Architectural Science, 1(2): 68–79, 2000. http://www.bse.polyu.edu.hk/researchCentre/Fire_Engineering/summary_of_output/journal/IJAS/V1/p.68-79.pdf
• Van der Vorst, H. A., Bi-CGSTAB: A fast and smoothly converging variant of Bi-CG for the solution of nonsymmetric linear systems. SlAM Journal on Scientific and Statistical Computing, 13(2): 631–644, 1992. doi:10.1137/0913035