§6.1.1: weak duality and Slater strong duality
OpenQAlgorithms.Papers.Independent.Intakenext100.P07.semidefinite_dualityAll quantum systems below are finite-dimensional complex Hilbert spaces of positive dimension, and logarithms are binary. A normalized state is positive semidefinite with trace one; a subnormalized state has trace at most one. Marginals are partial traces. Here is squared algebraic fidelity, is squared generalized fidelity, is purified distance, and . The symbols and denote conditional min- and max-entropy, with values and at zero. Smoothing optimizes over all subnormalized states in the purified-distance ball; every entropy radius satisfies .
For the source SDP, primal extended infimum a≥dual extended supremum b; if a is finite and there exists a strictly positive dual variable Y with E†Y<K, then a=b and the primal infimum is attained.
This result records the source’s operational or structural relation, with the stated endpoint and attainment clauses retained.
import Definitions.Def_QAlgorithms_Papers_Independent_Intakenext100_P07
noncomputable section
open Filter
open scoped BigOperators Matrix ComplexConjugate ComplexOrder Matrix.Norms.L2Operator Topology
namespace QAlgorithms.Papers.Independent.Intakenext100.P07
attribute [local instance] Classical.propDecidable Classical.decEq
variable {I J A B C D X Y Z : Type*}
variable [Fintype I] [DecidableEq I] [Fintype J] [DecidableEq J]
variable [Fintype A] [DecidableEq A] [Fintype B] [DecidableEq B]
variable [Fintype C] [DecidableEq C] [Fintype D] [DecidableEq D]
variable [Fintype X] [DecidableEq X] [Fintype Y] [DecidableEq Y]
variable [Fintype Z] [DecidableEq Z]
/-- NODE paper:1504.00233#6.sdpduality
§6.1.1: weak duality and Slater strong duality, arXiv:1504.00233v5, PDF p.92. -/
theorem semidefinite_duality [Nonempty I] [Nonempty J] (E : paperMat I →ₗ[ℂ] paperMat J) (K : paperMat I) (L : paperMat J)
(hK : K.IsHermitian) (hL : L.IsHermitian) (hE : ∀ M : paperMat I, M.IsHermitian → (E M).IsHermitian) :
sdpDual E K L ≤ sdpPrimal E K L ∧
((∃ a : ℝ, sdpPrimal E K L = (a:EReal)) →
(∃ Y : paperMat J, Y.PosDef ∧ (K-paperAdjointAction E Y).PosDef) →
sdpPrimal E K L = sdpDual E K L ∧ ∃ X : paperMat I, X.PosSemidef ∧ paperBelow L (E X) ∧
((K*X).trace.re:EReal) = sdpPrimal E K L) := by sorry
end QAlgorithms.Papers.Independent.Intakenext100.P07
end
Read-back
What the Lean code literally says, in plain math · gpt-6.1-sol
Let be nonempty finite types with decidable equality, let be a complex-linear map from matrices on to matrices on , and let be Hermitian matrices on , respectively. Assume that is Hermitian whenever is Hermitian. Define the matrix entrywise by
where has a single entry at , and define extended-real optimization values
Then . If equals some finite real number and there exists a matrix such that both and are positive definite, then and there exists with and , with the latter real value embedded into the extended reals. An empty primal feasible set has infimum , and an empty dual feasible set has supremum ; neither feasibility nor finiteness is assumed for the unconditional inequality.
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.