Displacement structure and eigenvalue bounds for multiband prolate matrices
Abstract
We obtain explicit non-asymptotic bounds for the eigenvalue transition of multiband prolate matrices. Let S⊂[−1/2,1/2] be a finite union of pairwise disjoint intervals, and let λ0≥⋯≥λN−1 be the eigenvalues of BN,S. If ν is the number of boundary points of S on the frequency torus, then, for every 0<eps<1/2, #{k:eps<λk<1−eps} ≤ νceillog(4N)/π2 log (4/(eps(1−eps))). Compared with the multiband estimate of Zhu and Wakin~\cite{wakin2017approximating}, this improves the dependence on eps from order 1/[eps(1−eps)] to logarithmic dependence. We also prove a second bound in which the main logarithmic term depends on the time--bandwidth product N|S| rather than only on N. We also compare S with frequency sets having fewer boundary points. Using a classical large-sieve estimate, we take into account not only the total size of the change in the frequency set, but also how that change is distributed. For sets with many small gaps, this can give a bound growing like log N for fixed eps, even when the bounds based only on the total size of the gaps grow at least linearly in N. We also obtain exponential bounds for individual eigenvalues and use them to estimate the error in approximating sampled multiband signals by leading eigenvectors. A discretization argument gives corresponding results for the continuous problem. Numerical examples compare the bounds and examine how the locations of the frequency intervals affect the eigenvalues.
Related
- Large deviations for the largest eigenvalue of a polynomial in Gaussian matrices
- Mesoscopic eigenvalue statistics for correlated random matrices
- Contour-integral and Fourier transform based multivariable quantum eigenvalue transformation for commuting matrices
- Staircase phase transitions for the largest eigenvalue of heavy-tailed sample correlation matrices
- Improved Inclusion-Exclusion Eigenvalue Localization Sets for Matrices and Matrix Polynomials
BibTeX
@misc{gomez2026displacement,
title = {{Displacement structure and eigenvalue bounds for multiband prolate matrices}},
author = {Luis Gomez and Hana Huber and Azita Mayeli},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.32082},
eprint = {2610.32082},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.32082},
}
RIS
TY - UNPB
TI - Displacement structure and eigenvalue bounds for multiband prolate matrices
AU - Luis Gomez
AU - Hana Huber
AU - Azita Mayeli
PY - 2026
DA - 2026/10/23
PB - arXiv
DO - 10.48550/arxiv.2610.32082
UR - https://arxiv.org/abs/2610.32082
AB - We obtain explicit non-asymptotic bounds for the eigenvalue transition of multiband prolate matrices. Let S⊂[−1/2,1/2] be a finite union of pairwise disjoint intervals, and let λ_0≥⋯≥λ_N−1 be the eigenvalues of B_N,S. If ν is the number of boundary points of S on the frequency torus, then, for every 0<eps<1/2, #{k:eps<λ_k<1−eps} ≤ νceillog(4N)/π^2 log (4/(eps(1−eps))). Compared with the multiband estimate of Zhu and Wakin~\cite{wakin2017approximating}, this improves the dependence on eps from order 1/[eps(1−eps)] to logarithmic dependence. We also prove a second bound in which the main logarithmic term depends on the time--bandwidth product N|S| rather than only on N. We also compare S with frequency sets having fewer boundary points. Using a classical large-sieve estimate, we take into account not only the total size of the change in the frequency set, but also how that change is distributed. For sets with many small gaps, this can give a bound growing like log N for fixed eps, even when the bounds based only on the total size of the gaps grow at least linearly in N. We also obtain exponential bounds for individual eigenvalues and use them to estimate the error in approximating sampled multiband signals by leading eigenvectors. A discretization argument gives corresponding results for the continuous problem. Numerical examples compare the bounds and examine how the locations of the frequency intervals affect the eigenvalues.
ER -