OAI/Analysis/AffineBernstein/AffineFamily.lean

GitHub source file

20 declarations, 0 contain sorry, 0 declared as axioms

  1. theorem OAI.AffineBernstein.LocalDistanceConverges.of_pointwise line 23
  2. theorem OAI.AffineBernstein.LocalDistanceConverges.homeomorph line 56
  3. theorem OAI.AffineBernstein.LocalDistanceConverges.diagonal line 100
  4. structure OAI.AffineBernstein.InAffineLimitFamily line 129

    Source comment

    Literal full-dimensional affine local-limit family of geometry.tex:87. The affine maps may vary along the approximating sequence.
    structure InAffineLimitFamily {F : Type*} [NormedAddCommGroup F] [NormedSpace ℝ F]
        (D : Set E) (C : Set F) : Prop where
    …
  5. theorem OAI.AffineBernstein.InAffineLimitFamily.nonempty line 142
  6. theorem OAI.AffineBernstein.InAffineLimitFamily.affine_image line 146
  7. theorem OAI.AffineBernstein.InAffineLimitFamily.diagonal_closure line 169
  8. theorem OAI.AffineBernstein.InAffineLimitFamily.of_affine_image line 178
  9. theorem OAI.AffineBernstein.matrixLinearEquiv_apply line 194

    Source comment

    The coordinate action of the actual matrix linear equivalence.
    theorem matrixLinearEquiv_apply {n : ℕ} (A : Matrix (Fin n) (Fin n) ℝ)
        (hA : IsUnit A.det) (x : Space n) (i : Fin n) :
        Matrix.toLinearEquiv (EuclideanSpace.basisFun (Fin n) ℝ).toBasis A hA x i =
          A.mulVec x i := by
    …

    Mentions: OAI.AffineBernstein.Space

  10. theorem OAI.AffineBernstein.convex_sum_mem_le_one line 204

    Source comment

    A convex combination allowing unused mass at the origin.
    theorem convex_sum_mem_le_one {E : Type*} [AddCommGroup E] [Module ℝ E]
        {ι : Type*} [Fintype ι] {K : Set E} (hK : Convex ℝ K) (h0 : (0 : E) ∈ K)
        (c : ι → ℝ) (v : ι → E) (hc : ∀ i, 0 ≤ c i)
        (hsum : ∑ i, c i ≤ 1) (hv : ∀ i, v i ∈ K) : (∑ i, c i • v i) ∈ K := by
    …
  11. theorem OAI.AffineBernstein.symmetric_convex_sum_mem line 222

    Source comment

    The cross-polytope spanned by any points of a symmetric convex body lies in that body. This is the inner-inclusion half of max-determinant normalization.
    theorem symmetric_convex_sum_mem {n : ℕ} {K : Set (Space n)}
        (hK : Convex ℝ K) (h0 : (0 : Space n) ∈ K)
        (hneg : ∀ x ∈ K, -x ∈ K) (v : Fin n → Space n)
        (hv : ∀ i, v i ∈ K) (c : Fin n → ℝ) (hc : ∑ i, |c i| ≤ 1) :
        (∑ i, c i • v i) ∈ K := by
    …

    Mentions: OAI.AffineBernstein.Space

  12. theorem OAI.AffineBernstein.sum_abs_coordinate_le line 239

    Source comment

    The exact ℓ¹/Euclidean comparison used for the inner ball.
    theorem sum_abs_coordinate_le {n : ℕ} (x : Space n) :
        ∑ i, |x i| ≤ Real.sqrt n * ‖x‖ := by
    …

    Mentions: OAI.AffineBernstein.Space

  13. theorem OAI.AffineBernstein.norm_le_sqrt_card_of_coordinate_le line 249

    Source comment

    The exact ℓ∞/Euclidean comparison used for the outer ball.
    theorem norm_le_sqrt_card_of_coordinate_le {n : ℕ} (x : Space n)
        (hx : ∀ i, |x i| ≤ 1) : ‖x‖ ≤ Real.sqrt n := by
    …

    Mentions: OAI.AffineBernstein.Space

  14. theorem OAI.AffineBernstein.exists_max_det_columns line 265

    Source comment

    A compact body's column determinant has a maximizer, with nonzero value whenever the body contains a ball about zero.
    theorem exists_max_det_columns {n : ℕ} {K : Set (Space n)}
        (hK : IsCompact K) {r : ℝ} (hr : 0 < r)
        (hball : Metric.closedBall (0 : Space n) r ⊆ K) :
        ∃ v : Fin n → Space n, (∀ i, v i ∈ K) ∧
          (Matrix.det (fun i j => v j i) ≠ 0) ∧
          ∀ w : Fin n → Space n, (∀ i, w i ∈ K) →
            |Matrix.det (fun i j => w j i)| ≤ |Matrix.det (fun i j => v j i)| := by
    …

    Mentions: OAI.AffineBernstein.Space

  15. theorem OAI.AffineBernstein.maxdet_coordinate_bound line 297

    Source comment

    Maximal determinant makes each inverse coordinate of a body point at most one. The proof is the literal one-column replacement identity.
    theorem maxdet_coordinate_bound {n : ℕ} {K : Set (Space n)}
        (v : Fin n → Space n) (hv : ∀ i, v i ∈ K)
        (hd : Matrix.det (fun i j => v j i) ≠ 0)
        (hmax : ∀ w : Fin n → Space n, (∀ i, w i ∈ K) →
          |Matrix.det (fun i j => w j i)| ≤ |Matrix.det (fun i j => v j i)|) :
        ∃ q : Space n ≃ₗ[ℝ] Space n,
          (∀ x : Space n, q x = ∑ i, x i • v i) ∧
          ∀ x ∈ K, ∀ i, |q.symm x i| ≤ 1 := by
    …

    Mentions: OAI.AffineBernstein.Space

  16. theorem OAI.AffineBernstein.symmetric_body_normalization line 339

    Source comment

    The symmetric max-determinant normalization, with the exact weaker radius n sufficient for all centered normalizations used in the source.
    theorem symmetric_body_normalization {n : ℕ} (hn : 1 ≤ n)
        {K : Set (Space n)} (hK : IsCompact K) (hconv : Convex ℝ K)
        (h0 : (0 : Space n) ∈ K) (hneg : ∀ x ∈ K, -x ∈ K)
        {r : ℝ} (hr : 0 < r) (hball : Metric.closedBall (0 : Space n) r ⊆ K) :
        ∃ A : Space n ≃L[ℝ] Space n,
          Metric.closedBall 0 1 ⊆ A '' K ∧ A '' K ⊆ Metric.closedBall 0 n := by
    …

    Mentions: OAI.AffineBernstein.Space

  17. theorem OAI.AffineBernstein.centered_body_normalization line 380

    Source comment

    Exact centered normalization in geometry.tex (2/(C+1), 2n), proved without importing John's theorem: first normalize the symmetric difference body, then use the actual balance inclusion.
    theorem centered_body_normalization {n : ℕ} (hn : 1 ≤ n)
        {K : Set (Space n)} (hK : IsCompact K) (hconv : Convex ℝ K)
        (hint : (interior K).Nonempty) (h0 : (0 : Space n) ∈ K)
        {C : ℝ} (hC : 1 ≤ C) (hbal : -K ⊆ C • K) :
        ∃ A : Space n ≃L[ℝ] Space n,
          Metric.closedBall 0 (2 / (C + 1)) ⊆ A '' K ∧
          A '' K ⊆ Metric.closedBall 0 (2 * n) := by
    …

    Mentions: OAI.AffineBernstein.Space

  18. theorem OAI.AffineBernstein.simplex_ball_coordinates line 451

    Source comment

    A dimension-only inscribed ball in the standard simplex. This allows noncentered model fibers to be normalized without importing John's theorem.
    theorem simplex_ball_coordinates {n : ℕ} (hn : 1 ≤ n) (z : Space n)
        (hz : z ∈ Metric.closedBall (WithLp.toLp 2 (fun _ : Fin n => (1 : ℝ) / (n + 1)))
          (1 / ((n : ℝ) + 1) ^ 2)) :
        (∀ i, 0 ≤ z i) ∧ (∑ i, z i) ≤ 1 := by
    …

    Mentions: OAI.AffineBernstein.Space

  19. theorem OAI.AffineBernstein.noncentered_body_normalization line 487

    Source comment

    A linear normalization of an arbitrary full-dimensional compact convex body containing the distinguished origin. The dimension-only outer radius is intentionally not John's sharp radius; it suffices for the compactness argument and does not alter the source's final centered normalization.
    theorem noncentered_body_normalization {n : ℕ} (hn : 1 ≤ n)
        {K : Set (Space n)} (hK : IsCompact K) (hconv : Convex ℝ K)
        (hint : (interior K).Nonempty) (h0 : (0 : Space n) ∈ K) :
        ∃ (A : Space n ≃L[ℝ] Space n) (c : Space n),
          Metric.closedBall c 1 ⊆ A '' K ∧
          A '' K ⊆ Metric.closedBall 0 (2 * ((n : ℝ) + 1) ^ 3) ∧
          ‖c‖ ≤ 2 * ((n : ℝ) + 1) ^ 3 := by
    …

    Mentions: OAI.AffineBernstein.Space

  20. theorem OAI.AffineBernstein.noncentered_body_normalization_all line 567

    Source comment

    The same intermediate normalization with the zero-dimensional convention made explicit. Its radii are only used in compactness arguments.
    theorem noncentered_body_normalization_all {n : ℕ}
        {K : Set (Space n)} (hK : IsCompact K) (hconv : Convex ℝ K)
        (hint : (interior K).Nonempty) (h0 : (0 : Space n) ∈ K) :
        ∃ (A : Space n ≃L[ℝ] Space n) (c : Space n),
          Metric.closedBall c 1 ⊆ A '' K ∧
          A '' K ⊆ Metric.closedBall 0 (2 * ((n : ℝ) + 1) ^ 3) ∧
          ‖c‖ ≤ 2 * ((n : ℝ) + 1) ^ 3 := by
    …

    Mentions: OAI.AffineBernstein.Space