Commuting unitary swivels make the powered Schatten objective nonincreasing
OpenQAlgorithms.Papers.Independent.Intakethird100.P91.swivelMonotonicityLet be integers and let act on the common space , with products in the displayed order. For , write for the unnormalized Schatten norm. On a positive eigenvalue , acts by , and it acts by zero on the kernel. No commutation between different factors or trace normalization is assumed.
Assume every is positive semidefinite. Let be the set of all tuples of full-space unitaries satisfying for every . For every pair of finite real exponents , both displayed maxima are attained and
This is the monotonicity of the optimized powered norm, including singular and zero factors.
Formalization Note The theorem asserts greatest elements of the two explicitly constructed sets of objective values and compares them. Greatest-element membership records attainment; there is no default real supremum.
import Definitions.Def_QAlgorithms_Papers_Independent_Intakethird100_P91
noncomputable section
open scoped Matrix ComplexConjugate ComplexOrder ENNReal Matrix.Norms.L2Operator
namespace QAlgorithms.Papers.Independent.Intakethird100.P91
open QAlgorithms.Papers.Independent.Information100.P50 (schattenNorm)
open QAlgorithms.Papers.Independent.Dependency50.P21
(extendedLog upperLogIntegral betaDensity betaZero interpolationExponent)
open QAlgorithms.Papers.Independent.Bridge50.P29
(positiveComplexPower orderedProduct)
/-- Theorem 1, Eq.(5), arXiv:1610.01262v1, PDF p.3 (abstract p.1).
For every finite real 1 <= q <= p, both maxima are attained and the
p-objective is at most the q-objective. IsGreatest includes attainment,
so this states literal maxima, rather than an unattained supremum bound.
The dimension and sequence are nonzero; matrices may be singular or zero
and need not have trace one. Positive powers use the support-zero convention. -/
theorem swivelMonotonicity
(d L : ℕ) (hd : 0 < d) (hL : 0 < L)
(C : Fin L → Matrix (Fin d) (Fin d) ℂ)
(hC : ∀ i, (C i).PosSemidef)
(p q : ℝ) (hq : 1 ≤ q) (hqp : q ≤ p) :
∃ mp mq : ℝ,
IsGreatest (swivelValues C p) mp ∧
IsGreatest (swivelValues C q) mq ∧ mp ≤ mq := by sorry
end QAlgorithms.Papers.Independent.Intakethird100.P91
end
Read-back
What the Lean code literally says, in plain math · gpt-6.1-sol
For every pair of positive natural numbers , every family of positive semidefinite complex matrices, and all real numbers satisfying , there exist real numbers such that is an attained greatest value of
over all families of unitary matrices satisfying , is an attained greatest value of the analogous expression with replaced by , and . “Attained greatest value” means both that the number belongs to the indicated set of values and that every value in that set is at most that number. Products are in increasing index order. For ,
and is obtained spectrally by replacing each positive eigenvalue with and each zero eigenvalue with . Singular and zero are allowed, and the case is included; and are excluded by the hypotheses.
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.