Theorem C.5 — the overlap regime : closed form of
OpenSpectralFeedback.Supplement.theorem_C_5_overlap_regimelinear-algebramatrix-analysisoptimizationp2o-gran-per-chapterp2o-plan-paperp2o-v1quantum-information
Let (standing in §C.5) and , the value at the maximally mixed state. If and , then
This value is strictly below the rank ceiling when , and equals at both endpoints and .
The formula interpolates between the disjoint-support regime and the saturated regime.
Differences from the source:
- encoding: the value G_{m,r}(ρ) (a maximum, (C.6)) → IsGreatest of the published rewardSet with F = χ_r; reason: the base mission's convention; IsGreatest asserts the maximum exists and equals the value, never a junk sSup
- encoding: I_d/d → (d : ℂ)⁻¹ • 1 on ℂ^d; reason: same matrix
Preamble
import Mathlib import Definitions.Def_SpectralFeedback_rewardSet import Definitions.Def_SpectralFeedback_Supplement_chi open scoped ComplexOrder
Formal statement
namespace SpectralFeedback
namespace Supplement
/-- **Theorem C.5** (The overlap regime; Qi, *Spectral Structure and Exact Arithmetic of
Finite-Feedback Measurement Discrimination*, manuscript (Nov. 2026), p.30 (PDF p.32)). With
`F_m(d, r) = G_{m,r}(I_d/d)` and the standing `2 ≤ r < d` of §C.5: if `m ≥ 2` and
`(m − 1)r ≤ d ≤ mr`, then
`F_m(d, r) = (d − 2m + 1 + √((d − 2m + 1)² + 4m(m − 1)(r − 1)))/(2d)` (C.17). It is strictly below
`(r − 1)/r` when `(m − 1)r < d < mr`, and equals that ceiling at both endpoints.
**Formalization Note** `I_d/d` is `(d : ℂ)⁻¹ • 1` on `ℂ^d`; `F_m(d, r) = v` is
`IsGreatest (rewardSet r (chi r) m (I_d/d)) v`. -/
theorem theorem_C_5_overlap_regime (m r d : ℕ) (hm : 2 ≤ m) (hr : 2 ≤ r) (hrd : r < d)
(h1 : (m - 1) * r ≤ d) (h2 : d ≤ m * r) :
IsGreatest (rewardSet r (chi r) m (((d : ℂ)⁻¹) • (1 : Matrix (Fin d) (Fin d) ℂ)))
(((d : ℝ) - 2 * m + 1 +
Real.sqrt (((d : ℝ) - 2 * m + 1) ^ 2 + 4 * m * ((m : ℝ) - 1) * ((r : ℝ) - 1))) / (2 * d)) ∧
((m - 1) * r < d → d < m * r →
((d : ℝ) - 2 * m + 1 +
Real.sqrt (((d : ℝ) - 2 * m + 1) ^ 2 + 4 * m * ((m : ℝ) - 1) * ((r : ℝ) - 1))) / (2 * d) <
((r : ℝ) - 1) / r) ∧
(d = (m - 1) * r ∨ d = m * r →
((d : ℝ) - 2 * m + 1 +
Real.sqrt (((d : ℝ) - 2 * m + 1) ^ 2 + 4 * m * ((m : ℝ) - 1) * ((r : ℝ) - 1))) / (2 * d) =
((r : ℝ) - 1) / r) := 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 C.5, (C.17), p.30 (PDF p.32); standing assumptions §C.5, p.30