Fast Interpolation-Based Globality Certificates for Computing Kreiss Constants and the Distance to Uncontrollability

We propose a new approach to computing global minimizers of singular value functions in two real variables. Specifically, we present new algorithms to compute the Kreiss constant of a matrix and the distance to uncontrollability of a linear control system, both to arbitrary accuracy. Previous state-of-the-art methods for these two quantities rely on 2D level-set tests that are based on solving large eigenvalue problems. Consequently, these methods are costly, i.e., $\mathcal{O}(n^6)$ work using dense eigensolvers, and often multiple tests are needed before convergence. Divide-and-conquer techniques have been proposed that reduce the work complexity to $\mathcal{O}(n^4)$ on average and $\mathcal{O}(n^5)$ in the worst case, but these variants are nevertheless still very expensive and can be numerically unreliable. In contrast, our new interpolation-based globality certificates perform level-set tests by building interpolant approximations to certain one-variable continuous functions that are both relatively cheap and numerically robust to evaluate. Our new approach has an $\mathcal{O}(kn^3)$ work complexity and uses $\mathcal{O}(n^2)$ memory, where $k$ is the number of function evaluations necessary to build the interpolants. Not only is this interpolation process mostly “embarrassingly parallel," but also low-fidelity approximations typically suffice for all but the final interpolant, which must be built to high accuracy. Even without taking advantage of the aforementioned parallelism, $k$ is sufficiently small that our new approach is generally orders of magnitude faster than the previous state of the art.

  • 1.  P. Benner R. Byers V. Mehrmann and  H. Xu , Numerical computation of deflating subspaces of skew-Hamiltonian/Hamiltonian pencils , SIAM J. Matrix Anal. Appl. , 24 ( 2002 ), pp. 165 -- 190 , https://doi.org/10.1137/S0895479800367439. LinkISIGoogle Scholar

  • 2.  P. Benner, M. Köhler, and J. Saak, Fast approximate solution of the non-symmetric generalized eigenvalue problem on multicore architectures, in Parallel Computing: Accelerating Computational Science and Engineering (CSE), M. Bader, A. Bodeand, H.-J. Bungartz, M. Gerndt, G. R. Joubert, and F. Peters, eds., Adv. Parallel Comput. 25, IOS Press, 2014, pp. 143--152, https://doi.org/10.3233/978-1-61499-381-0-143. Google Scholar

  • 3.  P. Benner and  T. Mitchell , Extended and improved criss-cross algorithms for computing the spectral value set abscissa and radius , SIAM J. Matrix Anal. Appl. , 40 ( 2019 ), pp. 1325 -- 1352 , https://doi.org/10.1137/19M1246213. LinkISIGoogle Scholar

  • 4.  P. Benner V. Sima and  M. Voigt , Algorithm 961: Fortran 77 subroutines for the solution of skew-Hamiltonian/Hamiltonian eigenproblems , ACM Trans. Math. Software , 42 ( 2016 ), 24 , https://doi.org/10.1145/2818313. CrossrefISIGoogle Scholar

  • 5.  J. V. Burke A. S. Lewis and  M. L. Overton , Robust stability and a criss-cross algorithm for pseudospectra , IMA J. Numer. Anal. , 23 ( 2003 ), pp. 359 -- 375 , https://doi.org/10.1093/imanum/23.3.359. CrossrefISIGoogle Scholar

  • 6.  J. V. Burke A. S. Lewis and  M. L. Overton , Pseudospectral components and the distance to uncontrollability , SIAM J. Matrix Anal. Appl. , 26 ( 2004 ), pp. 350 -- 361 , https://doi.org/10.1137/S0895479803433313. LinkISIGoogle Scholar

  • 7.  R. Byers , A bisection method for measuring the distance of a stable matrix to the unstable matrices , SIAM J. Sci. Statist. Comput. , 9 ( 1988 ), pp. 875 -- 881 , https://doi.org/10.1137/0909059. LinkISIGoogle Scholar

  • 8.  R. Byers , Detecting nearly uncontrollable pairs, in Signal Processing, Scattering and Operator Theory, and Numerical Methods, Vol. III, M. A. Kaashoek, J. H. Schuppen, and A. C. M. Ran, eds., Birkhäuser, Boston , MA , 1990 , pp. 447 -- 457 . Google Scholar

  • 9.  T. A. Driscoll, N. Hale, and L. N. Trefethen, Chebfun Guide, Pafnuty, Oxford, UK, 2014, http://www.chebfun.org/docs/guide/. Google Scholar

  • 10.  R. Eising , Between controllable and uncontrollable , Systems Control Lett. , 4 ( 1984 ), pp. 263 -- 264 , https://doi.org/10.1016/S0167-6911(84)80035-3. CrossrefISIGoogle Scholar

  • 11.  M. Embree and  B. Keeler , Pseudospectra of matrix pencils for transient analysis of differential-algebraic equations , SIAM J. Matrix Anal. Appl. , 38 ( 2017 ), pp. 1028 -- 1054 , https://doi.org/10.1137/15M1055012. LinkISIGoogle Scholar

  • 12.  A. Fazzi N. Guglielmi and  I. Markovsky , Computing common factors of matrix polynomials with applications in system and control theory, in Proceedings of the 58th IEEE Conference on Decision and Control (CDC), Nice , France , 2019 , pp. 7721 -- 7726 , https://doi.org/10.1109/CDC40024.2019.9030137. Google Scholar

  • 13.  M. Gao and  M. Neumann , A global minimum search algorithm for estimating the distance to uncontrollability , Linear Algebra Appl. , 188/189 ( 1993 ), pp. 305 -- 350 , https://doi.org/10.1016/0024-3795(93)90472-Z. CrossrefISIGoogle Scholar

  • 14.  M. Gu , New methods for estimating the distance to uncontrollability , SIAM J. Matrix Anal. Appl. , 21 ( 2000 ), pp. 989 -- 1003 , https://doi.org/10.1137/S0895479897328856. LinkISIGoogle Scholar

  • 15.  M. Gu E. Mengi M. L. Overton J. Xia and  J. Zhu , Fast methods for estimating the distance to uncontrollability , SIAM J. Matrix Anal. Appl. , 28 ( 2006 ), pp. 477 -- 502 , https://doi.org/10.1137/05063060X. LinkISIGoogle Scholar

  • 16.  M. Gu and  M. L. Overton , An algorithm to compute ${Sep}_\lambda$ , SIAM J. Matrix Anal. Appl. , 28 ( 2006 ), pp. 348 -- 359 , https://doi.org/10.1137/050622584. LinkISIGoogle Scholar

  • 17.  H.-O. Kreiss and  Über ̈r Differenzengleichungen die partielle Differentialgleichungen approximieren , BIT , 2 ( 1962 ), pp. 153 -- 181 , https://doi.org/10.1007/bf01957330. CrossrefGoogle Scholar

  • 18.  I. Markovsky A. Fazzi and  N. Guglielmi , Applications of polynomial common factor computation in signal processing, in Latent Variable Analysis and Signal Separation, Y. Deville, S. Gannot, R. Mason, M. D. Plumbley, and D. Ward, eds., Lecture Notes in Comput. Sci. 10891, Springer , Cham , 2018 , pp. 99 -- 106 , https://doi.org/10.1007/978-3-319-93764-9_10. Google Scholar

  • 19.  E. Mengi, Measures for Robust Stability and Controllability, Ph.D. thesis, New York University, New York, 2006, https://cs.nyu.edu/media/publications/mengi_emre.pdf. Google Scholar

  • 20.  T. Mitchell, ROSTAPACK: RObust STAbility PACKage, http://timmitchell.com/software/ROSTAPACK. Google Scholar

  • 21.  T. Mitchell, Fast Computation of sep$_\lambda$ via Interpolation-Based Globality Certificates, preprint, https://arxiv.org/abs/1911.05136, 2019. Google Scholar

  • 22.  T. Mitchell and  Computing Kreiss constant of a matrix , SIAM J. Matrix Anal. Appl. , 41 ( 2020 ), pp. 1944 -- 1975 , https://doi.org/10.1137/19M1275127. LinkISIGoogle Scholar

  • 23.  C. Paige , Properties of numerical algorithms related to computing controllability , IEEE Trans. Automat. Control , 26 ( 1981 ), pp. 130 -- 138 . CrossrefISIGoogle Scholar

  • 24.  L. N. Trefethen and M. Embree, Spectra and Pseudospectra: The Behavior of Nonnormal Matrices and Operators, Princeton University Press, Princeton, NJ, 2005. Google Scholar

  • 25.  T. G. Wright, EigTool, http://www.comlab.ox.ac.uk/pseudospectra/eigtool/, 2002. Google Scholar

  • 26.  T. G. Wright and  L. N. Trefethen , Pseudospectra of rectangular matrices , IMA J. Numer. Anal. , 22 ( 2002 ), pp. 501 -- 519 , https://doi.org/10.1093/imanum/22.4.501. CrossrefISIGoogle Scholar