Luminary Research Archive
Luminary Research ArchiveIndependent scholarship
Luminary Portal

All Research

Research

154 papers
Browse papers
  1. Finite certificates for the lower edge of a weighted Weil operatorScott SchubrEdition 2 · 2026-10-27 Public
  2. Effective Prescribed-Gradient Stability for Homogeneous Real-Stable PolynomialsScott SchubrEdition 2 · 2026-10-27 Public
  3. Prescribed-gradient minimizers and coefficient stability for homogeneous real-stable polynomialsScott SchubrEdition 4 · 2026-10-27 Public
  4. Finite theta interpolation and eventual Laguerre failureScott SchubrEdition 2 · 2026-10-27 Public
  5. Uniform permanent-derivative compensation on the six-row conjugate-disjoint Gram locusScott SchubrEdition 2 · 2026-10-27 Public
  6. Certified complementary bounds for an odd Weil receiverScott SchubrEdition 1 · 2026-10-27 Public
  7. RH, Approximately: Certified Lower Bounds for Odd Weil PositivityScott SchubrEdition 1 · 2026-10-27 Public
  8. Sharp regularity criteria for arithmetic cutoff functionalsScott SchubrEdition 1 · 2026-10-26 Public
  9. Optimal energy for Dirichlet repairs in divisor bandsScott SchubrEdition 1 · 2026-10-26 Public
  10. Certified negative-band approximation for odd Weil formsScott SchubrEdition 1 · 2026-10-26 Public
PublicEdition 2

Trace-perturbed minimizers of the mixed discriminant

Scott Schubr

Original PDF · First-page preview

Loading original first page…

Abstract

For each fixed matrix size, we prove that the scalar tuple uniquely minimizes the mixed discriminant of positive semidefinite Hermitian matrices whose sum is the identity and whose individual traces are sufficiently close to one. This answers Question 1.7 of Gurvits--Leake, arXiv:2609.21083v2. The argument combines Gurvits's doubly stochastic uniqueness theorem with a positive constrained Hessian and compactness. We also record the exact region of positive definiteness of that Hessian at a scalar tuple. The global neighborhood in the theorem is qualitative and may depend on the size.

Read the full paper in the original PDF.