Universality for the Toda algorithm to compute the eigenvalues of a random matrix

We prove universality for the fluctuations of the halting time $T^{(1)}$ (see Introduction) for the Toda algorithm to compute the eigenvalues of real symmetric and Hermitian matrices $H$. $T^{(1)}$ controls the computation time for the largest eigenvalue of $H$ to a prescribed tolerance. The proof relies on recent results on the statistics of the eigenvalues and eigenvectors of random matrices (such as delocalization, rigidity and edge universality) in a crucial way.

[1]  M. Toda.  Vibration of a Chain with Nonlinear Interaction , 1967 .

[2]  S. Manakov.  Complete integrability and stochastization of discrete dynamical systems , 1974 .

[3]  H. Flaschka.  The Toda lattice. II. Existence of integrals , 1974 .

[4]  J. Moser,et al.  Three integrable Hamiltonian systems connected with isospectral deformations , 1975 .

[5]  Mark Adler,et al.  On a trace functional for formal pseudo-differential operators and the symplectic structure of the Korteweg-devries type equations , 1978 .

[6]  B. Kostant,et al.  The solution to a generalized Toda lattice and representation theory , 1979 .

[7]  A. J. Stam.  LIMIT THEOREMS FOR UNIFORM DISTRIBUTIONS ON SPHERES IN HIGH-DIMENSIONAL EUCLIDEAN SPACES , 1982 .

[8]  P. Deift,et al.  Ordinary differential equations and the symmetric eigenvalue problem , 1983 .

[9]  Carlos Tomei,et al.  Toda flows with infinitely many variables , 1985 .

[10]  Carlos Tomei,et al.  The toda flow on a generic orbit is integrable , 1986 .

[11]  C. Tracy,et al.  Level-spacing distributions and the Airy kernel , 1992, hep-th/9211141.

[12]  David S. Watkins.  Isospectral Flows , 1996 .

[13]  A. Soshnikov.  Universality at the Edge of the Spectrum¶in Wigner Random Matrices , 1999, math-ph/9907013.

[14]  P. Deift.  Orthogonal Polynomials and Random Matrices: A Riemann-Hilbert Approach , 2000 .

[15]  Universality at the edge of the spectrum for unitary, orthogonal, and symplectic ensembles of random matrices , 2005, math-ph/0507023.

[16]  Tiefeng Jiang,et al.  How many entries of a typical orthogonal matrix can be approximated by independent normals , 2006 .

[17]  H. Yau,et al.  Rigidity of eigenvalues of generalized Wigner matrices , 2010, 1007.4652.

[18]  T. Tao,et al.  Random Matrices: Universality of Local Eigenvalue Statistics up to the Edge , 2009, 0908.1982.

[19]  J. Ramírez,et al.  Beta ensembles, stochastic Airy spectrum, and a diffusion , 2006, math/0607331.

[20]  Joshua Correll,et al.  A neural computation model for decision-making times , 2012 .

[21]  L. Erdős,et al.  Universality for random matrices and log-gases , 2012, 1212.0839.

[22]  P. J. Forrester,et al.  Joint distribution of the first and second eigenvalues at the soft edge of unitary ensembles , 2012, 1209.2190.

[23]  T. Garel,et al.  Typical versus averaged overlap distribution in spin glasses: Evidence for droplet scaling theory , 2013, 1306.0423.

[24]  H. Yau,et al.  The Eigenvector Moment Flow and Local Quantum Unique Ergodicity , 2013, 1312.1301.

[25]  H. Yau,et al.  Edge Universality of Beta Ensembles , 2013, 1306.5728.

[26]  G. Schehr,et al.  Near-Extreme Eigenvalues and the First Gap of Hermitian Random Matrices , 2013, 1312.2966.

[27]  P. Deift,et al.  Universality in numerical computations with random data , 2014, Proceedings of the National Academy of Sciences.

[28]  Yann LeCun,et al.  Universality in halting time and its applications in optimization , 2015, ArXiv.

[29]  P. Deift,et al.  On the condition number of the critically-scaled Laguerre Unitary Ensemble , 2015, 1507.00750.

[30]  P. Deift,et al.  How long does it take to compute the eigenvalues of a random, symmetric matrix? , 2012, 1203.4635.