Section 7.2: wasting and blocking alone do not suffice
ProvedQAlgorithms.Papers.Independent.Gap20.P07.wastingBlocking_counterexampleThere exist positive block lengths with and positive errors such that the finite blocking envelope
is monotone under resource wasting and subadditive. The infimum is extended nonnegative; its value for no feasible packing is , before clipping. Along the specified diagonal blocks , so rate one is sporadically achievable, but the extended supremum of positive rates satisfying the all-sequences definition is zero. Moreover there are positive sequences for which
This is an abstract error function, not the error of a quantum channel. It shows that wasting and blocking alone do not prove the sporadic-capacity equivalence.
import Definitions.Def_QAlgorithms_Papers_Independent_Gap20_P07
noncomputable section
open Filter MeasureTheory
open scoped Topology ENNReal NNReal Matrix ComplexOrder ComplexConjugate
attribute [local instance] Classical.propDecidable
attribute [local instance] Classical.decEq
attribute [local instance] IsStarNormal.instContinuousFunctionalCalculus
namespace QAlgorithms.Papers.Independent.Gap20.P07
open QAlgorithms
open QAlgorithms.RealRegisters (Hilbert)
open QAlgorithms.Papers.Independent.Dependency50.P33
(State IsTraceClass IsTraceClassChannel traceSize rankOne tensorApply)
open QAlgorithms.Papers.Independent.Frontier30.P14
(Operation isometryCoeff conjugateKernel)
open QAlgorithms.Papers.Independent.Claimed50.P35
(stabilizedTraceClassNorm traceClassTensorMap traceClassChannelPower
transpose transposeBound TraceLinear tracePair MeasurePrepare)
open QAlgorithms.Papers.Independent.Gap48.P29 (fe FiniteCPMap)
/-- PDF p.27, (102)--(103): an abstract clipped blocking envelope, not an error
arising from quantum channels, has sporadic rate one and capacity zero. The
rounding-sensitive bad subsequence is asserted with liminf >=1/2. -/
theorem wastingBlocking_counterexample :
∃ (N : ℕ → ℕ+) (epsilon : ℕ → ℝ≥0),
Tendsto (fun a => (N a : ℕ)) atTop atTop ∧
Tendsto (fun a => (N (a + 1) : ℝ) / (N a : ℝ)) atTop atTop ∧
(∀ a, 0 < epsilon a ∧ (epsilon a : ℝ) =
Real.sqrt ((N a : ℝ) / (N (a + 1) : ℝ)) / 2) ∧
Tendsto epsilon atTop (𝓝 0) ∧
ErrorMonotone (blockingEnvelope N epsilon) ∧
ErrorSubadditive (blockingEnvelope N epsilon) ∧
Tendsto (fun a => blockingEnvelope N epsilon (N a) (N a)) atTop (𝓝 0) ∧
Sporadic (blockingEnvelope N epsilon) 1 ∧
errorCapacity (blockingEnvelope N epsilon) = 0 ∧
∃ n m : ℕ → ℕ+, Tendsto (fun a => (n a : ℕ)) atTop atTop ∧
Tendsto (rateRatio n m) atTop (𝓝 0) ∧
(1 / 2 : ℝ≥0∞) ≤ Filter.liminf
(fun a => blockingEnvelope N epsilon (n a) (m a)) atTop := by sorry
end QAlgorithms.Papers.Independent.Gap20.P07
end
Read-back
What the Lean code literally says, in plain math · gpt-6.1-sol
There exist a sequence of positive integers and a sequence of nonnegative real numbers such that , , and, for every , and , with . For positive integers , define the extended nonnegative real number
List entries may repeat and are counted with multiplicity; the empty list cannot satisfy the constraint because , and an empty feasible set has infimum , giving . This function satisfies, for all positive integers ,
Moreover, , and there exist sequences of positive integers with , , and , where ratios and their limits are taken in . Nevertheless,
with the supremum taken in and hence equal to for an empty set. Finally, there exist positive-integer sequences such that , , and ; no separate requirement that tend to infinity is imposed.
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.