OAI/Analysis/KadisonKastler/FiniteProjectionAveraging.lean

GitHub source file

3 declarations, 0 contain sorry, 0 declared as axioms

Constructed finite-factor projection averages

  1. theorem OAI.KadisonKastler.ambient_projection_le line 21

    Source documentation

    The ambient operator order agrees with the algebraic projection order.

    theorem ambient_projection_le {p q : AlgebraProjection M} (hpq : p ≤ q) :
        p.val.val ≤ q.val.val := by
    …
  2. theorem OAI.KadisonKastler.exists_average_of_consecutive_matrix_cuts line 28

    Source documentation

    Consecutive equal-trace matrix cuts give an average within one matrix slot.

    theorem exists_average_of_consecutive_matrix_cuts (τ : TracialRepresentation M)
        {n m : ℕ} (hn : 0 < n) (hm : m < n)
        (E : MatrixUnitSystem (Fin n) (Algebra M))
        (ht : ∀ i, τ.trace (E.entry i i) = (((n : ℝ)⁻¹ : ℝ) : ℂ))
        (p : AlgebraProjection M)
        (hlo : matrixDiagonalCut E (Finset.univ.filter (fun i => i.val < m)) ≤ p)
        (hhi : p ≤ matrixDiagonalCut E
          (insert ⟨m, hm⟩ (Finset.univ.filter (fun i => i.val < m)))) :
        ∃ A : FiniteInnerAverage M,
          ‖A.apply p.val.val - (((m : ℝ) / n : ℝ) : ℂ) • (1 : H →L[ℂ] H)‖ ≤ (n : ℝ)⁻¹ := by
    …

    Mentions: OAI.KadisonKastler.Algebra, OAI.KadisonKastler.matrixDiagonalCut, OAI.KadisonKastler.FiniteInnerAverage

  3. theorem OAI.KadisonKastler.exists_projection_scalar_average_of_trace_lt_one line 70

    Source documentation

    Every projection of trace below one has scalar finite averages with arbitrarily small error, obtained by rational cuts and inner transport.

    theorem exists_projection_scalar_average_of_trace_lt_one (τ : TracialRepresentation M)
        (hfactor : ∀ c : Algebra M, (∀ a : Algebra M, a * c = c * a) →
          ∃ z : ℂ, c = z • (1 : Algebra M))
        (hdiff : ∀ p : Algebra M, IsStarProjection p → p ≠ 0 →
          ∃ q : Algebra M, IsStarProjection q ∧ q ≠ 0 ∧ q ≠ p ∧ p * q = q ∧ q * p = q)
        (p : AlgebraProjection M) (hp : (τ.trace p.val).re < 1)
        (ε : ℝ) (hε : 0 < ε) :
        ∃ (A : FiniteInnerAverage M) (z : ℂ),
          ‖A.apply p.val.val - z • (1 : H →L[ℂ] H)‖ < ε := by
    …

    Mentions: OAI.KadisonKastler.Algebra, OAI.KadisonKastler.FiniteInnerAverage