Skip to content
DraconiticResearch
Search Sources API
203,234 studies31 sourcesUpdated 19 min ago
PreprintarXiv licencearXiv

Eternal Total Positivity and Matrix-Dependent Hadamard Thresholds

Abstract

We study entrywise powers of totally positive matrices. Our first question is which matrices remain totally positive under every positive Hadamard power. We call them eternally totally positive. Our second question concerns a fixed totally positive matrix: what is the smallest exponent after which all of its entrywise powers are totally positive? We call this number the Hadamard threshold of the matrix. Unlike the usual critical exponent, this threshold depends on the individual matrix. We give a complete characterization in order three, obtain explicit bounds for the transition exponent, and construct symmetric matrices that realize every threshold in [0,1). We also prove the basic transformation rules and continuity of the threshold on the normalized 3×3 totally positive cone. For higher orders, we introduce the logarithmic core and use it to expand each minor near the zero exponent. This reduces the 4×4 problem to the full determinant and the 3×3 submatrices. For symmetric 4×4 matrices, we show that a positive-definite logarithmic core makes the full determinant positive for every positive exponent. A two-parameter Hankel family then realizes every threshold in (1,2). Hence the exact range of thresholds in the symmetric order-four case is [0,1)∪(1,2); neither $1$ nor $2$ is attained. We also give a recursive symmetric ETP criterion in every order and an explicit classification for symmetric Toeplitz matrices of order four. In order five, a certified example exhibits an intermediate-power obstruction in a nonprincipal minor. An explicit order-three condition and a finite critical-point test give a complete order-five test.

Related

  1. Restricted Coprime Representations Modulo 6P: Exact CRT Formulas and Positivity Thresholds
    Andres M. Salazar
    arXivPreprintCC BY 4.0Full textAlgebra & number theoryarXiv
  2. Sharp Restricted Isometry Thresholds for Global Minima of Rank-Restricted Matrix LASSO
    Richard Y. Zhang
    arXivPreprintarXiv licenceFull textOptimisation & controlStatisticsarXiv
  3. On a positivity conjecture of Fasi and Higham for Padé approximants to the matrix exponential
    Francesco Scardino
    arXivPreprintarXiv licenceFull textAnalysis & PDEsNumerical analysis & scientific computingarXiv
  4. Strong Duality Thresholds and Minimum Curvature for the Nearest Rectangular Symplectic Matrix Problem
    Yulin Xie, Aiqing Zhu, Yifa Tang
    arXivPreprintarXiv licenceFull textNumerical analysis & scientific computingOptimisation & controlarXiv
  5. A Counterexample to an Eternal Domination Conjecture
    Tom Adamczewski, William F. Klostermeyer
    arXivPreprintCC BY 4.0Full textCombinatorics & discrete mathematicsarXiv
BibTeX
@misc{fallat2026eternal,
  title = {{Eternal Total Positivity and Matrix-Dependent Hadamard Thresholds}},
  author = {Shaun M. Fallat and Seyed Mahmoud Manjegani and Mohammadrahim Saeidi Dahaki},
  howpublished = {arXiv},
  year = {2026},
  month = oct,
  doi = {10.48550/arxiv.2610.38757},
  eprint = {2610.38757},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2610.38757},
}
RIS
TY  - UNPB
TI  - Eternal Total Positivity and Matrix-Dependent Hadamard Thresholds
AU  - Shaun M. Fallat
AU  - Seyed Mahmoud Manjegani
AU  - Mohammadrahim Saeidi Dahaki
PY  - 2026
DA  - 2026/10/28
PB  - arXiv
DO  - 10.48550/arxiv.2610.38757
UR  - https://arxiv.org/abs/2610.38757
AB  - We study entrywise powers of totally positive matrices. Our first question is which matrices remain totally positive under every positive Hadamard power. We call them eternally totally positive. Our second question concerns a fixed totally positive matrix: what is the smallest exponent after which all of its entrywise powers are totally positive? We call this number the Hadamard threshold of the matrix. Unlike the usual critical exponent, this threshold depends on the individual matrix. We give a complete characterization in order three, obtain explicit bounds for the transition exponent, and construct symmetric matrices that realize every threshold in [0,1). We also prove the basic transformation rules and continuity of the threshold on the normalized 3×3 totally positive cone. For higher orders, we introduce the logarithmic core and use it to expand each minor near the zero exponent. This reduces the 4×4 problem to the full determinant and the 3×3 submatrices. For symmetric 4×4 matrices, we show that a positive-definite logarithmic core makes the full determinant positive for every positive exponent. A two-parameter Hankel family then realizes every threshold in (1,2). Hence the exact range of thresholds in the symmetric order-four case is [0,1)∪(1,2); neither $1$ nor $2$ is attained. We also give a recursive symmetric ETP criterion in every order and an explicit classification for symmetric Toeplitz matrices of order four. In order five, a certified example exhibits an intermediate-power obstruction in a nonprincipal minor. An explicit order-three condition and a finite critical-point test give a complete order-five test.
ER  -