Abstract
Data-driven scenarios in analysis and modeling of complex dynamical systems pose formidable challenges to computational science and motivate development of new numerical methods with ever increasing level of sophistication and complexity. An excellent example is computational fluid dynamics (CFD), where the dynamic mode decomposition (DMD) and its enhancement, the sparsity promoting DMD (DMDSP) have emerged as tools of trade for analysis of flow field data, with a host of applications. The problems of high dimension, noisy data, theoretical foundation in connection with the Koopman operator and ergodic theory, and potential applicability in broad spectrum of engineering problems have triggered extensive research and resulted in many theoretical and computational advancements of the DMD framework. This chapter revisits the DMD from the core numerical linear algebra perspective. Recent results on improving numerical robustness and functionality of DMD are reviewed and supplemented with new insights. Further, a new variation of the algorithm is proposed and used as a case study for development of a numerical algorithm in the DMD framework.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
Both MATLAB and NumPy actually use LAPACK as a computing engine for most matrix operations, in particular for the SVD and computing eigenvalues and eigenvectors of general matrices.
- 2.
In MATLAB, the function \(\texttt {sqrtm(.)}\) computes the matrix principal square root.
- 3.
- 4.
See [66] for a definition and basis properties of Krylov decomposition.
- 5.
This is sometimes called mixed stability: backward error corresponds to slightly changed output.
- 6.
For an earlier, extended version see [21].
- 7.
In an example constructed by Kahan [47], the pivoting fails to reveal one small singular value.
References
Anantharamu, S., Mahesh, K.: A parallel Dynamic Mode Decomposition algorithm using modified Full Orthogonalization Arnoldi for large sequential snapshots. ArXiv e-prints (2018)
Anderson, E., Bai, Z., Bischof, C., Blackford, L.S., Demmel, J., Dongarra, J.J., Du Croz, J., Hammarling, S., Greenbaum, A., McKenney, A., Sorensen, D.: LAPACK Users’ Guide, 3rd edn. Society for Industrial and Applied Mathematics, Philadelphia (1999)
Andrew, R., Dingle, N.: Implementing QR factorization updating algorithms on GPUs. Parallel Comput. 40(7), 161–172. http://www.sciencedirect.com/science/article/pii/S0167819114000337 (2014). https://doi.org/10.1016/j.parco.2014.03.003. (7th Workshop on Parallel Matrix Algorithms and Applications)
Arbabi, H., Mezić, I.: Ergodic theory, dynamic mode decomposition, and computation of spectral properties of the Koopman operator. SIAM J. Appl. Dyn. Syst. 16(4), 2096–2126 (2017). https://doi.org/10.1137/17M1125236
Bagheri, S.: Koopman-mode decomposition of the cylinder wake. J. Fluid Mech. 726, 596–623 (2013). https://doi.org/10.1017/jfm.2013.249
Berger, E., Sastuba, M., Vogt, D., Jung, B., Amor, H.B.: Estimation of perturbations in robotic behavior using dynamic mode decomposition. Adv. Robot. 29(5), 331–343 (2015)
Björck, A.: Numerical Methods in Matrix Computations. Springer, Berlin (2015)
Brgisser, P., Cucker, F.: Condition: The Geometry of Numerical Algorithms. Springer Publishing Company, Incorporated, Berlin (2013)
Brunton, B.W., Johnson, L.A., Ojemann, J.G., Kutz, J.N.: Extracting spatialtemporal coherent patterns in large-scale neural recordings using dynamic mode decomposition. J. Neurosci. Methods 258, 1–15. http://www.sciencedirect.com/science/article/pii/S0165027015003829 (2016). https://doi.org/10.1016/j.jneumeth.2015.10.010
Businger, P.A., Golub, G.H.: Linear least squares solutions by Householder transformations. Numerische Mathematik 7, 269–276 (1965)
Cat Le Ngo, A., See, J., Chung-Wei Phan, R.: Sparsity in Dynamics of Spontaneous Subtle Emotions: Analysis & Application. ArXiv e-prints (2016)
Chandrasekaran, S., Ipsen, I.C.F.: On rank-revealing QR factorisations. SIAM J. Matrix Anal. Appl. 15(2), 592–622 (1994)
Dawson, S.T.M., Hemati, M.S., Williams, M.O., Rowley, C.W.: Characterizing and correcting for the effect of sensor noise in the dynamic mode decomposition. Exp. Fluids 57(3), 42 (2016). https://doi.org/10.1007/s00348-016-2127-7
Demmel, J.: The geometry of iii-conditioning. J. Complex. 3(2), 201–229. http://www.sciencedirect.com/science/article/pii/0885064X87900276 (1987). https://doi.org/10.1016/0885-064X(87)90027-6
Demmel, J.: Accurate singular value decompositions of structured matrices. SIAM J. Matrix Anal. Appl. 21(2), 562–580 (1999)
Demmel, J., Grigori, L., Gu, M., Xiang, H.: Communication avoiding rank revealing qr factorization with column pivoting. SIAM J. Matrix Anal. Appl. 36(1), 55–89 (2015). https://doi.org/10.1137/13092157X
Demmel, J., Grigori, L., Hoemmen, M., Langou, J.: Communication-optimal parallel and sequential QR and LU factorizations. SIAM J. Sci. Comput. 34(1), A206–A239 (2012). https://doi.org/10.1137/080731992
Demmel, J., Gu, M., Eisenstat, S., Slapničar, I., Veselić, K., Drmač, Z.: Computing the singular value decomposition with high relative accuracy. Lin. Alg. Appl. 299, 21–80 (1999)
Drmač, Z.: Algorithm 977: A QR–preconditioned QR SVD method for computing the svd with high accuracy. ACM Trans. Math. Softw. 44(1), 11:1–11:30 (2017). https://doi.org/10.1145/3061709
Drmač, Z., Bujanović, Z.: On the failure of rank revealing QR factorization software - a case study. ACM Trans. Math. Softw. 35(2), 1–28 (2008)
Drmač, Z., Mezić, I., Mohr, R.: Data driven modal decompositions: analysis and enhancements (2017). arXiv:1708.02685
Drmač, Z., Mezić, I., Mohr, R.: Data driven Koopman spectral analysis in Vandermonde-Cauchy form via the DFT: numerical method and theoretical insights. SIAM J. Sci. Comput. 41(5), A3118–A3151 (2018). arXiv:1808.09557
Drmač, Z., Mezić, I., Mohr, R.: Data driven modal decompositions: analysis and enhancements. SIAM J. Sci. Comput. 40(4), A2253–A2285 (2018). https://doi.org/10.1137/17M1144155
Drmač, Z., Mezić, I., Mohr, R.: On least squares problems with certain Vandermonde–Khatri–Rao structure with applications to DMD (2018). arXiv:1811.12562
Drmač, Z., Veselić, K.: New fast and accurate Jacobi SVD algorithm: I. SIAM J. Matrix Anal. Appl. 29(4), 1322–1342 (2008)
Drmač, Z., Veselić, K.: New fast and accurate Jacobi SVD algorithm: II. SIAM J. Matrix Anal. Appl. 29(4), 1343–1362 (2008)
Eckart, C., Young, G.: The approximation of one matrix by another of lower rank. Psychometrika 1(3), 211–218 (1936). https://doi.org/10.1007/BF02288367
Eisenstat, S., Ipsen, I.: Relative perturbation techniques for singular value problems. SIAM J. Numer. Anal. 32(6), 1972–1988 (1995)
de la Fraga, L.G.: A very fast procedure to calculate the smallest singular value. In: 2015 Eighth International Conference on Advances in Pattern Recognition (ICAPR), pp. 1–4 (2015). https://doi.org/10.1109/ICAPR.2015.7050656
Gautschi, W.: Optimally conditioned Vandermonde matrices. Numerische Mathematik 24(1), 1–12 (1975). https://doi.org/10.1007/BF01437212
Gautschi, W.: How (un)stable are Vandermonde systems? In: R. Wong (ed.) Asymptotic and Computational Analysis. Lecture Notes in Pure and Applied Mathematics, vol. 124, pp. 193–210 (1990)
Ghosal, S., Ramanan, V., Sarkar, S., Chakravarthy, S., Sarkar, S.: Detection and analysis of combustion instability from hi-speed flame images using dynamic mode decomposition. In: ASME. Dynamic Systems and Control Conference, vol. 1 (2016). https://doi.org/10.1115/DSCC2016-9907
Ginsberg, J., Mohebbi, M., Patel, R., Brammer, L., Smolinski, M., Brilliant, L.: Detecting influenza epidemics using search engine query data. Nature 457, 1012–1014. http://www.nature.com/nature/journal/v457/n7232/full/nature07634.html (2009). https://doi.org/10.1038/nature07634
Golub, G.H., Klema, V.C., Stewart, G.W.: Rank degeneracy and least squares problems. Technical Report CS-TR-76-559, Stanford, CA, USA (1976)
Gu, M., Eisenstat, S.C.: Efficient algorithms for computing a strong rank-revealing QR factorization. SIAM J. Sci. Comput. 17(4), 848–869 (1996)
Hammarling, S., Lucas, C.: Updating th QR factorization and the least squares problem. Technical Report MIMS EPrint: 2008.111, University of Manchester (2008)
Higham, D.J.: Condition numbers and their condition numbers. Linear Algebr. Appl. 214, 193–213. http://www.sciencedirect.com/science/article/pii/0024379593000669 (1995). https://doi.org/10.1016/0024-3795(93)00066-9
Higham, N.J.: Accuracy and Stability of Numerical Algorithms. SIAM, Philadelphia (1996)
Holmes, P., Lumley, J.L., Berkooz, G., Rowley, C.W.: Turbulence, Coherent Structures, Dynamical Systems and Symmetry, 2 edn. Cambridge Monographs on Mechanics. Cambridge University Press, Cambridge (2012). https://doi.org/10.1017/CBO9780511919701
Horn, R.A., Johnson, C.R.: Topics in Matrix Analysis. Cambridge University Press, Cambridge (1991)
Jia, Z.: Refined iterative algorithms based on Arnoldi’s process for large unsymmetric eigenproblems. Linear Algebr. Appl. 259, 1–23. http://www.sciencedirect.com/science/article/pii/S0024379596002388 (1997). https://doi.org/10.1016/S0024-3795(96)00238-8
Jia, Z.: Polynomial characterizations of the approximate eigenvectors by the refined Arnoldi method and an implicitly restarted refined Arnoldi algorithm. Linear Algebr. Appl. 287(1–3), 191–214. http://www.sciencedirect.com/science/article/pii/S0024379598101970 (1999). https://doi.org/10.1016/S0024-3795(98)10197-0
Jia, Z.: Residuals of refined projection methods for large matrix eigenproblems. Comput. Math. Appl. 41(7), 813–820. http://www.sciencedirect.com/science/article/pii/S0898122100003217 (2001). https://doi.org/10.1016/S0898-1221(00)00321-7
Jia, Z.: Some theoretical comparisons of refined Ritz vectors and Ritz vectors. Sci. China Ser. Math. 47, 222–233 (2004)
Jia, Z., Stewart, G.W.: An analysis of the Rayleigh-Ritz method for approximating eigenspaces. Math. Comput. 70(234), 637–647 (2001). https://doi.org/10.1090/S0025-5718-00-01208-4
Jovanović, M.R., Schmid, P.J., Nichols, J.W.: Sparsity-promoting dynamic mode decomposition. Phys. Fluids 26(2), 024103 (2014)
Kahan, W.: Numerical linear algebra. Can. Math. Bull. 9(6), 757–801 (1965)
Kalashnikova, I., Arunajatesan, S., Barone, M.F., van Bloemen Waanders, B.G., Fike, J.A.: Reduced order modeling for prediction and control of large–scale systems. Sandia Report SAND2014–4693, Sandia National Laboratories (2014)
Kutz, J.N., Brunton, S.L., Brunton, B.W., Proctor, J.L.: Dynamic Mode Decomposition: Data-Driven Modeling of Complex Systems. SIAM-Society for Industrial and Applied Mathematics, USA (2016)
Lele, S.K., Nichols, J.W.: A second golden age of aeroacoustics? Philos. Trans. R. Soc. Lond. A Math. Phys. Eng. Sci. 372(2022). http://rsta.royalsocietypublishing.org/content/372/2022/20130321 (2014). https://doi.org/10.1098/rsta.2013.0321
Li, R.: Relative perturbation theory: Ii. eigenspace and singular subspace variations. SIAM J. Matrix Anal. Appl. 20(2), 471–492 (1998). https://doi.org/10.1137/S0895479896298506
Mann, J., Kutz, J.N.: Dynamic Mode Decomposition for Financial Trading Strategies. ArXiv e-prints (2015)
The Mathworks, Inc., Natick, Massachusetts: MATLAB version 9.2.0.556344 (R2017a) (2017)
Mirsky, L.: Symmetric gauge functions and unitarily invariant norms. Q. J. Math. 11(1), 50 (1960). https://doi.org/10.1093/qmath/11.1.50
Nguyen, H.D., Demmel, J.: Reproducible tall-skinny QR. In: 2015 IEEE 22nd Symposium on Computer Arithmetic, pp. 152–159 (2015). https://doi.org/10.1109/ARITH.2015.28
Oliphant, T.: Guide to NumPy. Trelgol Publishing (2006)
Pan, V.Y.: How bad are Vandermonde matrices? ArXiv e-prints http://adsabs.harvard.edu/abs/2015arXiv150402118P (2015)
Proctor, J.L., Eckhoff, P.A.: Discovering dynamic patterns from infectious disease data using dynamic mode decomposition. Int. Health 7(2), 139–145 (2015). https://doi.org/10.1093/inthealth/ihv009
Rowley, C.W.: Model reduction for fluids, using balanced proper orthogonal decomposition. Int. J. Bifur. Chaos Appl. Sci. Eng. 15(3), 997–1013 (2005)
Rowley, C.W., Colonius, T., Murray, R.M.: Model reduction for compressible flows using pod and galerkin projection. Phys. D: Nonlinear Phenom 189(1), 115–129. http://www.sciencedirect.com/science/article/pii/S0167278903003841 (2004). https://doi.org/10.1016/j.physd.2003.03.001
Rowley, C.W., Mezić, I., Bagheri, S., Schlatter, P., Henningson, D.S.: Spectral analysis of nonlinear flows. J. Fluid Mech. 641, 115–127 (2009)
Schmid, P.J.: Dynamic mode decomposition of numerical and experimental data. J. Fluid Mech. 656, 5–28. https://www.cambridge.org/core/article/dynamic-mode-decomposition-of-numerical-and-experimental-data/AA4C763B525515AD4521A6CC5E10DBD4 (2010). https://doi.org/10.1017/S0022112010001217
Schmid, P.J., Sesterhenn, J.: Dynamic mode decomposition of numerical and experimental data. In: Sixty-First Annual Meeting of the APS Division of Fluid Dynamics. San Antonio, Texas, USA (2008)
Serre, G., Lafon, P., Gloerfelt, X., Bailly, C.: Reliable reduced-order models for time-dependent linearized Euler equations. J. Comput. Phys. 231(15), 5176–5194 (2012)
van der Sluis, A.: Condition numbers and equilibration of matrices. Numerische Mathematik 14, 14–23 (1969)
Stewart, G.W.: A Krylov-Schur algorithm for large eigenproblems. SIAM J. Matrix Anal. Appl. 23(3), 601–614 (2001)
Stewart, G.W., Sun, J.G.: Matrix Perturbation Theory. Academic, Cambridge (1990)
Tu, J.H., Rowley, C.W., Luchtenburg, D.M., Brunton, S.L., Kutz, J.N.: On Dynamic Mode Decomposition: Theory and Applications. ArXiv e-prints (2013)
Volkwein, S.: Model reduction using proper orthogonal decomposition. Lecture Notes, Institute of Mathematics and Scientific Computing, University of Graz. http://www.uni-graz.at/imawww/volkwein/POD.pdf (2011)
Williams, M.O., Kevrekidis, I.G., Rowley, C.W.: A data-driven approximation of the Koopman operator: extending dynamic mode decomposition. J. Nonlinear Sci. 25(6), 1307–1346 (2015). https://doi.org/10.1007/s00332-015-9258-5
Acknowledgements
This research is supported by the DARPA Contract HR0011-16-C-0116 “On a Data-Driven, Operator-Theoretic Framework for Space–Time Analysis of Process Dynamics” and the DARPA Contract HR0011-18-9-0033 “The Physics of Artificial Intelligence”.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2020 Springer Nature Switzerland AG
About this chapter
Cite this chapter
Drmač, Z. (2020). Dynamic Mode Decomposition—A Numerical Linear Algebra Perspective. In: Mauroy, A., Mezić, I., Susuki, Y. (eds) The Koopman Operator in Systems and Control. Lecture Notes in Control and Information Sciences, vol 484. Springer, Cham. https://doi.org/10.1007/978-3-030-35713-9_7
Download citation
DOI: https://doi.org/10.1007/978-3-030-35713-9_7
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-35712-2
Online ISBN: 978-3-030-35713-9
eBook Packages: Intelligent Technologies and RoboticsIntelligent Technologies and Robotics (R0)