Minimum eigenvalues cannot separate faster than trace distance
ProvedQAlgorithms.Papers.Independent.Support50.P17.minimumEigenvalueSeparationLet be matrices on the same nonzero finite-dimensional complex Hilbert space, with . Let and denote their full smallest eigenvalues, including any zero eigenvalues. For , where , one has
No trace-one normalization is imposed. This supplies the eigenvalue-distance constraint needed in the entropy bounds.
import Definitions.Def_QAlgorithms_Papers_Independent_Support50_P17
noncomputable section
open MeasureTheory
open scoped Matrix ComplexOrder Matrix.Norms.L2Operator
namespace QAlgorithms.Papers.Independent.Support50.P17
attribute [local instance] Classical.propDecidable Classical.decEq
-- Reuse the exact frozen support predicate and signed, unnormalised entropy.
open QAlgorithms.Papers.Independent.Dependency50.P21 (IsSupported relativeEntropy)
/-- PDF p.6, Lemma 2: separation of the full minimum eigenvalues for equal-trace
PSD matrices, including zero eigenvalues and the one-dimensional case. -/
theorem minimumEigenvalueSeparation {I : Type*} [Fintype I] [DecidableEq I] [Nonempty I]
(A B : Matrix I I ℂ) (hA : A.PosSemidef) (hB : B.PosSemidef)
(htrace : A.trace = B.trace) :
|minEigenvalue A - minEigenvalue B| ≤ traceDistance A B := by sorry
end QAlgorithms.Papers.Independent.Support50.P17
end
Read-back
What the Lean code literally says, in plain math · gpt-6.1-sol
For every finite, nonempty index type with decidable equality and every pair of positive semidefinite complex matrices indexed by with equal complex traces,
Here denotes the smallest eigenvalue, including zero eigenvalues, and the square root is defined by continuous functional calculus. Neither matrix is required to have trace or to be invertible; zero matrices and one-dimensional spaces are included.
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.