Theorem G.5 — one spike and one dip, with the tail pooled into one existing symbol
OpenSpectralFeedback.Supplement.theorem_G_5_spike_and_diplinear-algebramatrix-analysisoptimizationp2o-gran-per-chapterp2o-plan-paperp2o-v1quantum-information
Let and (standing in §G.2). Suppose the active spectrum of (its leading eigenvalues, padded with zeros and not renormalized) is
Then
Eigenvalues outside the active space are pooled into one existing coarse symbol: for every decreasing spectral basis there is an optimal -atom decomposition of the active matrix such that adding to any one of its atoms gives an optimal decomposition of with the same atoms.
This is the appendix form of Theorem 7.1. It holds for subnormalized and adds the completion clause.
Differences from the source:
- encoding: the value G_{m,s}(ρ) → IsGreatest of the published rewardSet with F = χ_s; reason: the base mission's convention
- encoding: the (unnormalized) active spectrum λ_1, …, λ_p, zero-padded → eigDesc ρ i for i < p (0-based); reason: the published eigDesc pads with zeros beyond d; same numbers
- reading: eigenvalues outside the active space are pooled into one existing coarse symbol → for every U, ∃ optimal active decomposition such that adding ρ − activePart ρ U (min d (m·s)) to any atom h₀ is optimal for ρ; reason: the sentence names no atom; the stronger reading (every choice works) is stated
Preamble
import Mathlib import Definitions.Def_SpectralFeedback_rewardSet import Definitions.Def_SpectralFeedback_pairAtom import Definitions.Def_SpectralFeedback_Supplement_chi open scoped ComplexOrder
Formal statement
namespace SpectralFeedback
namespace Supplement
/-- **Theorem G.5** (One spike and one dip; Qi, *Spectral Structure and Exact Arithmetic of
Finite-Feedback Measurement Discrimination*, manuscript (Nov. 2026), p.102 (PDF p.104)). Standing
(§G.2): `p = ms` with `m ≥ 2`, `s ≥ 2`; the active spectrum is padded with zeros and not renormalized.
Suppose the active spectrum of `ρ` is `(b + a, b, …, b, b − c)` (`p − 2` middle entries), `a ≥ 0`,
`0 ≤ c ≤ b`. Then `G_{m,s}(ρ) = χ_s(mb + a − c, mb, …, mb)` (`s − 1` entries `mb`). Eigenvalues
outside the active space are pooled into one existing coarse symbol.
**Formalization Note** `G_{m,s}(·) = v` is `IsGreatest (rewardSet s (chi s) m ·) v`; `ρ` is any
complex positive semidefinite matrix. The active spectrum is `eigDesc ρ i`, `i < p` (0-based,
zero-padded). The pooling sentence is: for every decreasing spectral basis `U`, there is an optimal
`m`-atom decomposition of the active matrix `ρ_{[min(d,p)]}` (`activePart ρ U (min d (m * s))`)
such that adding the outside eigenvalues `ρ − ρ_{[min(d,p)]}` to any one of its atoms gives an
optimal decomposition of `ρ` with the same `m` atoms. -/
theorem theorem_G_5_spike_and_dip (m s : ℕ) (hm : 2 ≤ m) (hs : 2 ≤ s) {d : ℕ}
(ρ : Matrix (Fin d) (Fin d) ℂ) (hρ : ρ.PosSemidef) (a b c : ℝ) (ha : 0 ≤ a) (hc : 0 ≤ c)
(hcb : c ≤ b)
(hspec : eigDesc ρ 0 = b + a ∧ (∀ i, 1 ≤ i → i ≤ m * s - 2 → eigDesc ρ i = b) ∧
eigDesc ρ (m * s - 1) = b - c) :
IsGreatest (rewardSet s (chi s) m ρ)
(chi s (fun i => if (i : ℕ) = 0 then (m : ℝ) * b + a - c else (m : ℝ) * b)) ∧
∀ U : Matrix (Fin d) (Fin d) ℂ, IsSortedEigenbasis ρ U →
∃ A : Fin m → Matrix (Fin d) (Fin d) ℂ, (∀ h, (A h).PosSemidef) ∧
∑ h, A h = activePart ρ U (min d (m * s)) ∧
∑ h, reward s (chi s) (A h) =
chi s (fun i => if (i : ℕ) = 0 then (m : ℝ) * b + a - c else (m : ℝ) * b) ∧
∀ h₀ : Fin m,
(∀ h, (Function.update A h₀ (A h₀ + (ρ - activePart ρ U (min d (m * s)))) h).PosSemidef) ∧
∑ h, Function.update A h₀ (A h₀ + (ρ - activePart ρ U (min d (m * s)))) h = ρ ∧
∑ h, reward s (chi s)
(Function.update A h₀ (A h₀ + (ρ - activePart ρ U (min d (m * s)))) h) =
chi s (fun i => if (i : ℕ) = 0 then (m : ℝ) * b + a - c else (m : ℝ) * b) := by sorry
end Supplement
end SpectralFeedback
Source
Qi, Spectral Structure and Exact Arithmetic of Finite-Feedback Measurement Discrimination, manuscript (Nov. 2026; PDF sha256 a5b5cf40…eabde9cf), Theorem G.5, p.102 (PDF p.104); standing assumptions §G.2, p.102