All eight optimal complementary-slackness relations
ProvedQAlgorithms.Papers.Independent.Support50.P04.optimal_complementary_slacknessLet be a CPTP channel on finite positive-dimensional complex Hilbert spaces with fixed orthonormal bases. Write and for the maximum unassisted and classical-post-processing assisted entanglement-transmission rates. A code has a positive integer message dimension , rate , and squared fidelity to the normalized maximally entangled target at least . A cpp encoder may retain sender memory; its decoder is LOCC across that memory and the receiver outputs, with its final quantum output at the receiver and no preshared entanglement. Let . Let be the unnormalized Choi matrix. Define . Set , where the comparison Choi operators satisfy , , and . In the minimization program, is a density matrix, , , , and . In the maximization program, , , , , , and . Assume the displayed feasible minimization and maximization variables are optimal: . Then all eight relations hold:
These characterize relations between actual optimal solutions, not arbitrary feasible variables.
import Definitions.Def_QAlgorithms_Papers_Independent_Support50_P04
noncomputable section
open scoped Matrix ComplexConjugate ComplexOrder
open QAlgorithms
namespace QAlgorithms.Papers.Independent.Support50.P04
variable {A B E : Type} [Fintype A] [DecidableEq A]
variable [Fintype B] [DecidableEq B] [Fintype E] [DecidableEq E]
/-- arXiv:1504.04617v2, PDF p.22, Eqs.(A9)–(A16): ALL eight relations for
optimal feasible solutions. The T_A in (A13) is visible in the PDF but lost by
extraction; implicit identity tensor factors are explicit below. -/
theorem optimal_complementary_slackness [Nonempty A] [Nonempty B]
(N : Superop A B) (hN : IsChannel N) (epsilon : ℝ)
(he0 : 0 < epsilon) (he1 : epsilon < 1)
(rho xi : Matrix A A ℂ) (Lambda Theta J R : Matrix (A × B) (A × B) ℂ)
(mu nu : ℝ) (hmin : SdpMinFeasible N epsilon rho xi Lambda Theta)
(hmax : SdpMaxFeasible N mu nu J R)
(hminopt : xi.trace.re = metaValue N epsilon)
(hmaxopt : mu * (1-epsilon) - nu = metaValue N epsilon) :
rho.trace = 1 ∧
(Lambda * QAlgorithms.Papers.Independent.Information100.P34.choi N).trace.re = 1-epsilon ∧
Matrix.kronecker rho.transpose (1 : Matrix B B ℂ) * R = Lambda * R ∧
Matrix.kronecker xi (1 : Matrix B B ℂ) * J = (Lambda + QAlgorithms.Entanglement.partialTransposeFirst Theta) * J ∧
(nu : ℂ) • rho = traceRight (QAlgorithms.Entanglement.partialTransposeFirst R) * rho ∧
QAlgorithms.Entanglement.partialTransposeFirst J * Theta = 0 ∧
traceRight J * xi = xi ∧
(mu : ℂ) • (QAlgorithms.Papers.Independent.Information100.P34.choi N * Lambda) = (J + R) * Lambda := by sorry
end QAlgorithms.Papers.Independent.Support50.P04
end
Read-back
What the Lean code literally says, in plain math · gpt-6.1-sol
Let be nonempty finite index types with decidable equality, and let be a map from complex matrices to complex matrices that is a channel: there exist a natural number and matrices , , such that and for every . Write for the matrix with entries , where is the matrix unit; write for partial transpose on , and . Let , let , let , and let . Assume that
and that
Define the real number
where is indexed by and by . Assume also that and . Then all of the following equalities hold:
The assumptions allow or , singular feasible matrices, and one-element index types; they do not impose strict positivity on these quantities.
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.