The real power inequality
ProvedQAlgorithms.Papers.Independent.Extend50.P30.powerInequalityLet be real and . If , then the inequality below holds. The right-side hypothesis forces , so both real-power bases in the conclusion have their intended nonnegative values.
This gives the source's the real power inequality in the fixed computational-basis and phase conventions.
Formalization Note Pure states carry norm one; stabilizer averages use density matrices to discard global phases. All finite tensor indices are explicitly ordered by copy, then physical qudit.
import Definitions.Def_QAlgorithms_Papers_Independent_Extend50_P30
noncomputable section
open scoped BigOperators ComplexConjugate Matrix
namespace QAlgorithms.Papers.Independent.Extend50.P30
open QAlgorithms.Papers.Independent.Information100.P63 (tensorPower)
open QAlgorithms.Papers.Independent.Gap20.P06 (ketBra)
attribute [local instance] Classical.propDecidable
/-- PDF p.57, Lemma 7.4: real k>=1 and nonnegative right-hand side are both retained. -/
theorem powerInequality (k p : ℝ) (hk : 1 ≤ k) (hp : p ∈ Set.Icc 0 1)
(hrhs : 0 ≤ 4 * Real.rpow p k - 3) :
4 * Real.rpow p k - 3 ≤ Real.rpow (4 * p - 3) k := by sorry
end QAlgorithms.Papers.Independent.Extend50.P30
end
Read-back
What the Lean code literally says, in plain math · gpt-6.1-sol
For every pair of real numbers satisfying , , and , one has
where both powers are the real power function. The exponent need not be an integer. The hypotheses allow and , while cannot satisfy the additional hypothesis .
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.