Appendix C construction: invariant orbit basis and projector marginals, Eq.(C1)
ProvedQAlgorithms.Papers.Independent.Extend50.P29.invariant_orbit_projectorLet be finite, , and . On the regular group basis define
Then , its range is exactly the subspace invariant under the diagonal left regular action, and is a Hermitian idempotent. For every one of the registers,
This identifies the invariant-orbit probability space used in Appendix C. The projector is unnormalized: , while is the mean density matrix of a Haar invariant state.
import Definitions.Def_QAlgorithms_Papers_Independent_Extend50_P29
noncomputable section
open MeasureTheory MeasureTheory.Measure
open scoped Matrix ComplexConjugate ComplexOrder Matrix.Norms.L2Operator
attribute [local instance] Classical.propDecidable Classical.decEq
namespace QAlgorithms.Papers.Independent.Extend50.P29
open QAlgorithms.Papers.Independent.Information100.P32 (evolution)
open QAlgorithms.Papers.Independent.Next46.P39 (inverseSqrt rectAct)
open QAlgorithms.Papers.Independent.Support21.P09 (groupLeftRegular)
/-- PDF p.13, Eq.(C1): M is an isometry onto exactly the invariant subspace;
its projector becomes identity after tracing any one register (no density normalization). -/
theorem invariant_orbit_projector
(G : Type) [Group G] [Fintype G] {n : ℕ} (hn : 1 ≤ n) :
(invariantOrbitMap G n)ᴴ * invariantOrbitMap G n = 1 ∧
(∀ ψ : EuclideanSpace ℂ (Fin (n + 1) → G),
RegularInvariant ψ ↔ ∃ φ : EuclideanSpace ℂ (Fin n → G), rectAct (invariantOrbitMap G n) φ = ψ) ∧
(orbitProjector G n)ᴴ = orbitProjector G n ∧
orbitProjector G n * orbitProjector G n = orbitProjector G n ∧
(∀ j : Fin (n + 1), dependentRegisterTrace (H := fun _ : Fin (n + 1) => G)
(Finset.univ.erase j) (orbitProjector G n) = 1) := by sorry
end QAlgorithms.Papers.Independent.Extend50.P29
end
Read-back
What the Lean code literally says, in plain math · gpt-6.1-sol
For every finite group and natural number , let be the matrix indexed by , , with entry when for all , and zero otherwise, and put . Then ; for every vector indexed by , simultaneous left multiplication leaves invariant if and only if there exists a vector indexed by such that , where simultaneous invariance means for every , with ; and ; and for every register , the partial trace of over register is the identity on the remaining registers. This partial trace sums matrix entries over equal values of the discarded register. The vectors need not be normalized, and the trivial group is included.
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.