Theorem 5: Lieb three-matrix resolvent trace inequality
OpenQAlgorithms.Papers.Independent.Intakethird100.P51.liebThreeMatrixLet be positive definite matrices. The matrix-valued function is integrable on , and
Pairwise commutation of is sufficient for equality. It is not asserted necessary. This is Lieb’s three-matrix bound used in the proof of SSA.
import Definitions.Def_QAlgorithms_Papers_Independent_Intakethird100_P51
noncomputable section
open scoped Matrix ComplexOrder ComplexConjugate Matrix.Norms.L2Operator
open MeasureTheory
namespace QAlgorithms.Papers.Independent.Intakethird100.P51
attribute [local instance] Classical.propDecidable Classical.decEq
-- Exact interfaces from the frozen accepted library; no local redeclarations.
open QAlgorithms.Papers.Independent.Information100.P63 (Mat)
open QAlgorithms.Papers.Independent.Information100.P56 (entropy holevo IsLaw)
open QAlgorithms.Papers.Independent.Information100.P27 (average)
open QAlgorithms.Papers.Independent.Dependency50.P21 (relativeEntropy marginalAB channelAdjoint)
open QAlgorithms.Papers.Independent.Dependency50.P35 (marginalAC)
open QAlgorithms.Papers.Independent.Gap48.P19 (shannonNats)
open QAlgorithms.Papers.Independent.Gap48.P33 (measurementInformation)
section Statements
variable {A B C D J K O : Type}
variable [Fintype A] [Fintype B] [Fintype C] [Fintype D] [Fintype J] [Fintype K] [Fintype O]
variable [DecidableEq A] [DecidableEq B] [DecidableEq C] [DecidableEq D]
variable [DecidableEq J] [DecidableEq K] [DecidableEq O]
/-- Theorem 5, equation (17), PDF p.9: the factor R is inside the integral.
Integrability is a conclusion. Commutation is sufficient for equality, not necessary. -/
theorem liebThreeMatrix [Nonempty A] (R S T : Mat A) (hR : R.PosDef) (hS : S.PosDef) (hT : T.PosDef) :
IntegrableOn (fun u : ℝ => R * (S + (u : ℂ) • 1)⁻¹ * T * (S + (u : ℂ) • 1)⁻¹)
(Set.Ici 0) ∧
(NormedSpace.exp (logMatrix R - logMatrix S + logMatrix T)).trace.re ≤
(∫ u : ℝ in Set.Ici 0, R * (S + (u : ℂ) • 1)⁻¹ * T * (S + (u : ℂ) • 1)⁻¹).trace.re ∧
((R * S = S * R ∧ R * T = T * R ∧ S * T = T * S) →
(NormedSpace.exp (logMatrix R - logMatrix S + logMatrix T)).trace.re =
(∫ u : ℝ in Set.Ici 0, R * (S + (u : ℂ) • 1)⁻¹ * T * (S + (u : ℂ) • 1)⁻¹).trace.re) := by sorry
end Statements
end QAlgorithms.Papers.Independent.Intakethird100.P51
end
Read-back
What the Lean code literally says, in plain math · gpt-6.1-sol
For every nonempty finite index type with decidable equality and all positive definite complex matrices indexed by , the matrix-valued function
is Lebesgue integrable on , and
If additionally , , and , the two sides are equal. Each logarithm is defined by continuous functional calculus using the natural logarithm, and the exponential is the matrix exponential. The inverses are total matrix inverses, which are ordinary inverses here because is positive definite for . No trace normalization is required; the equality assertion is an implication from pairwise commutation, not an equivalence.
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.