Theorem 11: pure-ensemble RSP trade-off
OpenQAlgorithms.Papers.Independent.Intakenext100.P54.pure_ensemble_tradeoffLet be a finite ensemble on a fixed space , , with nonnegative probabilities summing to one and normalized vectors. Repeated targets and zero probabilities are allowed. On length- blocks the source states and probabilities are products. Let be the infimum of nonnegative ebit rates achievable at cbit rate , where every positive error and rate slack precede a blocklength threshold, every larger blocklength admits one one-way code, and fidelity is the global ensemble average. Then
The auxiliary alphabet is finite but arbitrary. For these pure targets . At every feasible rate the information minimum is attained and the resulting operational rate pair is achievable; infeasible rates have value . All entropies are in bits. This identifies the complete one-way pure-ensemble trade-off.
import Definitions.Def_QAlgorithms_Papers_Independent_Intakenext100_P54
noncomputable section
open scoped BigOperators Matrix ComplexConjugate ComplexOrder NNReal ENNReal
open MeasureTheory ProbabilityTheory
namespace QAlgorithms.Papers.Independent.Intakenext100.P54
open QAlgorithms.Papers.Independent.Information100.P03 (PureEnsemble)
open QAlgorithms.Papers.Independent.Information100.P53 (fidelity)
open QAlgorithms.Papers.Independent.Dependency50.P07 (wordProb)
open QAlgorithms.Papers.Independent.Claimed50.P08 (classicalEntropy)
open QAlgorithms.Papers.Independent.Claimed50.P38 (tensorSuperop)
/-- Theorem 11, PDF pp.6–7. The operational curve equals the minimum over equation (5),
including the attainment assertion at feasible rates. Repeated targets are allowed. -/
theorem pure_ensemble_tradeoff {d : ℕ} (hd : 0 < d) (E : PureEnsemble (Fin d))
(R : ℝ) (hR : 0 ≤ R) :
pureOperational E R = rspSingleLetter E.weight (fun i => pureDensity (E.vector i)) R ∧
rspMinimumAttained E.weight (fun i => pureDensity (E.vector i)) R ∧
(pureOperational E R ≠ ⊤ → RSPPureAchievable E R (pureOperational E R).toReal) := by sorry
end QAlgorithms.Papers.Independent.Intakenext100.P54
end
Read-back
What the Lean code literally says, in plain math · gpt-6.1-sol
For every positive natural number , every finite ensemble with , , and unit vectors , and every real , the theorem makes the following assertions. Here denotes the matrix with entries , and , using continuous functional calculus for ; write . For a finite array with , set , , and when , otherwise . Define
where the infimum is in . Define a pair to be achievable precisely as follows: , and for every , there is a positive integer such that for every , there exist positive integers , a -outcome POVM on for each word , and maps from to matrices, each admitting a finite representation with , such that, on writing
one has
A POVM here means that all its matrices are positive semidefinite and their sum is the identity. Let be the infimum in of all achievable . The theorem asserts jointly that ; that if is finite, some finite array attains it while satisfying ; and that if is finite, the pair , with viewed as a real number, is itself achievable. Empty feasible sets have infimum , zero ensemble weights are allowed, and division by zero in the auxiliary formulas is zero. Ensemble normalization forces , so an auxiliary array cannot have .
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.