Rupert Counterexample

6. The Local Theorem🔗

Lemma6.1
Group: Linear-algebra lemmas for local geometry. (5)
Group member previews
Preview
Definition 6.2
Loading preview
Group member preview content is loaded from the rendered-fragment cache.
uses 0used by 1L∃∀N

For any P \in \mathbb{R}^3 one has \|M(\theta, \phi) P\|^2=\|P\|^2-\langle X(\theta,\varphi),P\rangle^2.

Lean code for Lemma6.11 theorem
  • complete
    theorem Local.pythagoras {θ φ : } (P : Euc(3)) :
      (rotM θ φ) P ^ 2 = P ^ 2 - inner  (vecX θ φ) P ^ 2
    theorem Local.pythagoras {θ φ : } (P : Euc(3)) :
      (rotM θ φ) P ^ 2 =
        P ^ 2 - inner  (vecX θ φ) P ^ 2
    [SY25] Lemma 21.
    
    `rotM θ φ` consists of the first two rows of a rotation whose third row is
    `vecX θ φ`, so this is Parseval for that rotated orthonormal basis. 
Proof for Lemma 6.1
uses 0

See polyhedron.without.rupert, Lemma 21.

Definition6.2
Group: Linear-algebra lemmas for local geometry. (5)
Group member previews
Preview
Lemma 6.1
Loading preview
Group member preview content is loaded from the rendered-fragment cache.
uses 0used by 1L∃∀N

Given v_1, \dots, v_n \in \R^n write \mathrm{span}^+(v_1,\dots,v_n) for the set (simplicial cone) in \R^n defined by

\mathrm{span}^+(v_1,\dots,v_n) = \Big\{ w \in \R^n \colon \exists \lambda_1,\dots,\lambda_n > 0 \text{ s.t. } w = \sum_{i=1}^n \lambda_i v_i \Big\}, $$

which is the natural restriction of \mathrm{span}(v_1,\dots,v_n) to positive weights.

Lean code for Definition6.21 definition
  • complete
    def Local.Spanp {n : } (v : Fin n  Euc(n)) : Set Euc(n)
    def Local.Spanp {n : } (v : Fin n  Euc(n)) :
      Set Euc(n)
    The positive cone of a finite collection of vectors 
Lemma6.3
Group: Linear-algebra lemmas for local geometry. (5)
Group member previews
Preview
Lemma 6.1
Loading preview
Group member preview content is loaded from the rendered-fragment cache.
uses 0used by 1L∃∀N

Let V_1,V_2,V_3,Y,Z \in \mathbb{R}^3 with \|Y \|=\| Z \| and Y,Z \in \mathrm{span}^+(V_1,V_2,V_3). Then at least one of the following inequalities fails:

  • \langle V_1, Y \rangle > \langle V_1, Z \rangle

  • \langle V_2, Y \rangle > \langle V_2, Z \rangle

  • \langle V_3, Y \rangle > \langle V_3, Z \rangle

Lean code for Lemma6.31 theorem
  • complete
    theorem Local.langles {Y Z : Euc(3)} {V : Fin 3  Euc(3)} (hYZ : Y = Z)
      (hY : Y  Local.Spanp V) (hZ : Z  Local.Spanp V) :
      inner  (V 0) Y  inner  (V 0) Z 
        inner  (V 1) Y  inner  (V 1) Z 
          inner  (V 2) Y  inner  (V 2) Z
    theorem Local.langles {Y Z : Euc(3)}
      {V : Fin 3  Euc(3)} (hYZ : Y = Z)
      (hY : Y  Local.Spanp V)
      (hZ : Z  Local.Spanp V) :
      inner  (V 0) Y  inner  (V 0) Z 
        inner  (V 1) Y  inner  (V 1) Z 
          inner  (V 2) Y  inner  (V 2) Z
    [SY25] Lemma 23 
Proof for Lemma 6.3
uses 0

See polyhedron.without.rupert, Lemma 23.

Lemma6.4
Group: Linear-algebra lemmas for local geometry. (5)
Group member previews
Preview
Lemma 6.1
Loading preview
Group member preview content is loaded from the rendered-fragment cache.
uses 0used by 1L∃∀N

For A,\overline{A},B,\overline{B}\in \mathbb{R}^{n\times n} and P_1,P_2\in \mathbb{R}^n it holds that

|\langle AP_1,BP_2\rangle-\langle \overline{A}P_1,\overline{B}P_2\rangle| \leq \|P_1\|\,\|P_2\|\,\big( \|A-\overline{A}\|\,\|\overline{B}\| + \|\overline{A}\|\,\|B-\overline{B}\|+\|A-\overline{A}\|\,\|B-\overline{B}\|\big).

Lean code for Lemma6.41 theorem
  • complete
    theorem Local.abs_sub_inner_bars_le {m n : } (A B A_ B_ : Euc(m) →L[] Euc(n))
      (P₁ P₂ : Euc(m)) :
      |inner  (A P₁) (B P₂) - inner  (A_ P₁) (B_ P₂)| 
        P₁ * P₂ *
          (A - A_ * B_ + A_ * B - B_ + A - A_ * B - B_)
    theorem Local.abs_sub_inner_bars_le {m n : }
      (A B A_ B_ : Euc(m) →L[] Euc(n))
      (P₁ P₂ : Euc(m)) :
      |inner  (A P₁) (B P₂) -
            inner  (A_ P₁) (B_ P₂)| 
        P₁ * P₂ *
          (A - A_ * B_ + A_ * B - B_ +
            A - A_ * B - B_)
    [SY25] Lemma 24 
Proof for Lemma 6.4
uses 0

See polyhedron.without.rupert, Lemma 24.

Lemma6.5
Group: Linear-algebra lemmas for local geometry. (5)
Group member previews
Preview
Lemma 6.1
Loading preview
Group member preview content is loaded from the rendered-fragment cache.
uses 0used by 1L∃∀N

For A,B\in \mathbb{R}^{n\times n} and P_1,P_2\in \mathbb{R}^n one has

|\langle AP_1,AP_2\rangle-\langle BP_1,BP_2\rangle| \leq \|P_1\|\,\|P_2\|\,\|A-B\|\,(\|A\|+\|B\| + \|A-B\|).

Lean code for Lemma6.51 theorem
  • complete
    theorem Local.abs_sub_inner_le {m n : } (A B : Euc(m) →L[] Euc(n))
      (P₁ P₂ : Euc(m)) :
      |inner  (A P₁) (A P₂) - inner  (B P₁) (B P₂)| 
        P₁ * P₂ * A - B * (A + B + A - B)
    theorem Local.abs_sub_inner_le {m n : }
      (A B : Euc(m) →L[] Euc(n))
      (P₁ P₂ : Euc(m)) :
      |inner  (A P₁) (A P₂) -
            inner  (B P₁) (B P₂)| 
        P₁ * P₂ * A - B *
          (A + B + A - B)
    [SY25] Lemma 25 
Proof for Lemma 6.5
uses 0

See polyhedron.without.rupert, Lemma 25.

Lemma6.6
Group: Linear-algebra lemmas for local geometry. (5)
Group member previews
Preview
Lemma 6.1
Loading preview
Group member preview content is loaded from the rendered-fragment cache.
uses 0used by 1L∃∀N

Let A,B,C\in \mathbb{R}^2 be such that \langle R(\pi/2) A,B\rangle, \langle R(\pi/2) B,C\rangle, \langle R(\pi/2) C,A\rangle > 0. Then the origin lies strictly in triangle ABC.

Lean code for Lemma6.61 theorem
  • complete
    theorem Local.origin_in_triangle {A B C : Euc(2)}
      (hA : 0 < inner  ((rotR (Real.pi / 2)) A) B)
      (hB : 0 < inner  ((rotR (Real.pi / 2)) B) C)
      (hC : 0 < inner  ((rotR (Real.pi / 2)) C) A) :
       a b c, 0 < a  0 < b  0 < c  a  A + b  B + c  C = 0
    theorem Local.origin_in_triangle {A B C : Euc(2)}
      (hA :
        0 <
          inner  ((rotR (Real.pi / 2)) A) B)
      (hB :
        0 <
          inner  ((rotR (Real.pi / 2)) B) C)
      (hC :
        0 <
          inner  ((rotR (Real.pi / 2)) C)
            A) :
       a b c,
        0 < a 
          0 < b 
            0 < c  a  A + b  B + c  C = 0
    [SY25] Lemma 26 
Proof for Lemma 6.6
uses 0

See polyhedron.without.rupert, Lemma 26.

Definition6.7
groupuses 0used by 1L∃∀N

Let \theta, \varphi \in \mathbb{R}, \varepsilon > 0, and set M := M(\theta, \varphi). Three points P_1, P_2, P_3 \in \mathbb{R}^3 with \|P_1\|, \|P_2\|, \|P_3\| \leq 1 are called \varepsilon-spanning for (\theta, \varphi) if:

  • \langle R(\pi/2) M P_1,M P_{2}\rangle > 2 \epsilon(\sqrt{2} + \varepsilon)

  • \langle R(\pi/2) M P_2,M P_{3}\rangle > 2 \epsilon(\sqrt{2} + \varepsilon)

  • \langle R(\pi/2) M P_3,M P_{1}\rangle > 2 \epsilon(\sqrt{2} + \varepsilon)

Lean code for Definition6.71 definition
  • structure(2 fields)defined in Noperthedron/Local/EpsSpanning.lean
    complete
    structure Local.Triangle.Spanning (P : Local.Triangle) (θ φ ε : ) : Prop
    structure Local.Triangle.Spanning
      (P : Local.Triangle) (θ φ ε : ) : Prop
    [SY25] Definition 27. Note that the "+ 1" at the type Fin 3 wraps. 
    pos : 0 < ε
    lt :  (i : Fin 3), 2 * ε * (2 + ε) < inner  ((rotR (Real.pi / 2)) ((rotM θ φ) (P i))) ((rotM θ φ) (P (i + 1)))
Lemma6.8
group
Statement uses 2
Statement dependency previews
Preview
Definition 6.2
Loading preview
Statement dependency preview content is loaded from the rendered-fragment cache.
Used by 2
Reverse dependency previews
Preview
Theorem 6.14
Loading preview
Reverse dependency preview content is loaded from the rendered-fragment cache.
L∃∀N

Let P_1, P_2, P_3 \in \mathbb{R}^3 with \|P_1\|,\|P_2\|,\|P_3\| \leq 1 be \epsilon-spanning for (\bar\theta, \bar\phi) and let \theta, \phi \in \mathbb{R} satisfy |\theta - \bar{\theta}|, |\phi - \bar{\phi}| \leq \epsilon. If \langle X(\theta, \phi), P_i \rangle > 0 for i=1,2,3, then X(\theta, \phi) \in \spanp(P_1, P_2, P_3).

Lean code for Lemma6.81 theorem
  • complete
    theorem Local.vecX_spanning {ε θ θ_ φ φ_ : } (P : Local.Triangle)
      ( : |θ - θ_|  ε) ( : |φ - φ_|  ε)
      (hSpanning : P.Spanning θ_ φ_ ε) (hP :  (i : Fin 3), P i  1)
      (hX :  (i : Fin 3), 0 < inner  (vecX θ φ) (P i)) :
      vecX θ φ  Local.Spanp P
    theorem Local.vecX_spanning {ε θ θ_ φ φ_ : }
      (P : Local.Triangle) ( : |θ - θ_|  ε)
      ( : |φ - φ_|  ε)
      (hSpanning : P.Spanning θ_ φ_ ε)
      (hP :  (i : Fin 3), P i  1)
      (hX :
         (i : Fin 3),
          0 < inner  (vecX θ φ) (P i)) :
      vecX θ φ  Local.Spanp P
    [SY25] Lemma 28 
Proof for Lemma 6.8
Proof uses 3
Proof dependency previews
Preview
Lemma 3.4
Loading preview
Proof dependency preview content is loaded from the rendered-fragment cache.

See polyhedron.without.rupert, Lemma 28.

Lemma6.9
Group: Distance and local-maximality sector estimates. (3)
Group member previews
Preview
Definition 6.10
Loading preview
Group member preview content is loaded from the rendered-fragment cache.
uses 0used by 1L∃∀N

Let P, Q \in \mathbb{R}^3 with \|P\|, \|Q\| \leq 1, and let \epsilon>0, \bar\theta_1,\bar\phi_1,\bar\theta_2,\bar\phi_2,\bar\alpha \in \mathbb{R}. Set

T := \left(R(\bar\alpha) M(\bar\theta_1, \bar\phi_1) P + M(\bar\theta_2, \bar\phi_2) Q\right)/2 \in \mathbb{R}^2,

and assume \delta \geq \|T - M(\bar\theta_2, \bar\phi_2) Q\|. If |\bar\theta_1-\theta_1|, |\bar\phi_1-\phi_1|, |\bar\theta_2-\theta_2|, |\bar\phi_2-\phi_2|, |\bar\alpha - \alpha| \leq \epsilon, then R(\alpha)M(\theta_1, \phi_1) P and M(\theta_2, \phi_2) Q lie in \mathrm{Circ}_{\delta + \sqrt{5} \epsilon}(T).

Lean code for Lemma6.91 theorem
  • theoremdefined in Noperthedron/Local.lean
    complete
    theorem Local.inCirc {δ ε θ₁ θ₁_ θ₂ θ₂_ φ₁ φ₁_ φ₂ φ₂_ α α_ : } {P Q : Euc(3)}
      (hP : P  1) (hQ : Q  1) ( : 0 < ε) (hθ₁ : |θ₁ - θ₁_|  ε)
      (hφ₁ : |φ₁ - φ₁_|  ε) (hθ₂ : |θ₂ - θ₂_|  ε) (hφ₂ : |φ₂ - φ₂_|  ε)
      ( : |α - α_|  ε) :
      have T :=
        midpoint  ((rotR α_) ((rotM θ₁_ φ₁_) P)) ((rotM θ₂_ φ₂_) Q);
      T - (rotM θ₂_ φ₂_) Q  δ 
        (rotR α) ((rotM θ₁ φ₁) P)  Metric.ball T (5 * ε + δ) 
          (rotM θ₂ φ₂) Q  Metric.ball T (5 * ε + δ)
    theorem Local.inCirc
      {δ ε θ₁ θ₁_ θ₂ θ₂_ φ₁ φ₁_ φ₂ φ₂_ α α_ :
        }
      {P Q : Euc(3)} (hP : P  1)
      (hQ : Q  1) ( : 0 < ε)
      (hθ₁ : |θ₁ - θ₁_|  ε)
      (hφ₁ : |φ₁ - φ₁_|  ε)
      (hθ₂ : |θ₂ - θ₂_|  ε)
      (hφ₂ : |φ₂ - φ₂_|  ε)
      ( : |α - α_|  ε) :
      have T :=
        midpoint 
          ((rotR α_) ((rotM θ₁_ φ₁_) P))
          ((rotM θ₂_ φ₂_) Q);
      T - (rotM θ₂_ φ₂_) Q  δ 
        (rotR α) ((rotM θ₁ φ₁) P) 
            Metric.ball T (5 * ε + δ) 
          (rotM θ₂ φ₂) Q 
            Metric.ball T (5 * ε + δ)
    [SY25] Lemma 30 
Proof for Lemma 6.9
Proof uses 2
Proof dependency previews
Preview
Lemma 3.4
Loading preview
Proof dependency preview content is loaded from the rendered-fragment cache.

See polyhedron.without.rupert, Lemma 30.

Definition6.10
Group: Distance and local-maximality sector estimates. (3)
Group member previews
Preview
Lemma 6.9
Loading preview
Group member preview content is loaded from the rendered-fragment cache.
uses 0used by 1L∃∀N

Let \PP \subset \R^2 be a convex polygon and Q \in \PP one of its vertices. Assume that for some \overline{Q} \in \R^2 it holds that Q \in \Circ_{\delta}(\overline{Q}), i.e. \|Q - \overline{Q}\| < \delta. Define \Sect_\delta(\overline{Q}) \coloneqq \Circ_{\delta}(\overline{Q}) \cap \PP^\circ as the intersection between \Circ_{\delta}(\overline{Q}) and the interior of the convex hull of \PP.

Moreover, Q \in \PP is called \delta-locally maximally distant with respect to \overline{Q} (\delta-LMD(\overline{Q})) if for all A \in \Sect_\delta(\overline{Q}) it holds that \|Q\| > \|A\|.

Lean code for Definition6.101 definition
  • def Local.LocallyMaximallyDistant (δ : ) (Q Q_ : Euc(2))
      (P : Finset Euc(2)) : Prop
    def Local.LocallyMaximallyDistant (δ : )
      (Q Q_ : Euc(2)) (P : Finset Euc(2)) :
      Prop
    [SY25] Definition 31
    "Q is δ-locally maximally distant with respect to Q_" or "Q is δ-LMD(Q_)".
    
Lemma6.11
Group: Distance and local-maximality sector estimates. (3)
Group member previews
Preview
Lemma 6.9
Loading preview
Group member preview content is loaded from the rendered-fragment cache.
uses 1used by 1L∃∀N

Let \mathbf{P} be a convex polygon and Q \in \mathbf{P} a vertex. Let \overline{Q} \in \mathbb{R}^2 with \|Q - \overline{Q}\| < \delta for some \delta>0. Assume there exists r > 0 with \|Q\| > r such that

\frac{\langle Q, Q - P_j \rangle}{\|Q\|\|Q - P_j\|} \geq \delta/r

for all other vertices P_j \in \mathbf{P} \setminus Q. Then Q is \delta-locally maximally distant with respect to \overline{Q}.

Lean code for Lemma6.111 theorem
  • theorem Local.inner_ge_implies_LMD {P : Finset Euc(2)} {Q Q_ : Euc(2)} {δ r : }
      (hQ : Q  P) (hQ_ : Q - Q_ < δ) (hr : 0 < r) (hrQ : r < Q)
      (hle :
         Pᵢ  P, Pᵢ  Q  δ / r  inner  Q (Q - Pᵢ) / (Q * Q - Pᵢ)) :
      Local.LocallyMaximallyDistant δ Q Q_ P
    theorem Local.inner_ge_implies_LMD
      {P : Finset Euc(2)} {Q Q_ : Euc(2)}
      {δ r : } (hQ : Q  P)
      (hQ_ : Q - Q_ < δ) (hr : 0 < r)
      (hrQ : r < Q)
      (hle :
         Pᵢ  P,
          Pᵢ  Q 
            δ / r 
              inner  Q (Q - Pᵢ) /
                (Q * Q - Pᵢ)) :
      Local.LocallyMaximallyDistant δ Q Q_ P
    [SY25] Lemma 32
    
Proof for Lemma 6.11
uses 0

See polyhedron.without.rupert, Lemma 32.

Lemma6.12
Group: Distance and local-maximality sector estimates. (3)
Group member previews
Preview
Lemma 6.9
Loading preview
Group member preview content is loaded from the rendered-fragment cache.
uses 0used by 1L∃∀N

Let \epsilon>0 and \theta,\bar\theta, \phi, \bar\phi \in \mathbb{R} with |\theta - \bar{\theta}|, |\phi - \bar{\phi}| \leq \epsilon. Define M = M(\theta, \phi) and \overline{M} = M(\bar\theta, \bar\phi), and let P, Q \in \mathbb{R}^3 with \|P\|, \|Q\| \leq 1. Assume that

\frac{\langle \overline{M} P,\overline{M} (P-Q)\rangle - 2 \epsilon \|P-Q\| \cdot (\sqrt{2}+\varepsilon)}{ \big(\|\overline{M} P\|+\sqrt{2} \varepsilon \big) \cdot \big(\|\overline{M}(P-Q)\|+2 \sqrt{2} \varepsilon\big)} > 0. $$

Then:

\frac{\langle MP,M(P-Q)\rangle}{\|MP\|\,\|M(P-Q)\|} \geq \frac{\langle \overline{M} P,\overline{M} (P-Q)\rangle - 2 \epsilon \|P-Q\| \cdot (\sqrt{2}+\varepsilon)}{ (\|\overline{M} P\|+\sqrt{2} \varepsilon ) \cdot (\|\overline{M}(P-Q)\|+2 \sqrt{2} \varepsilon)}.

Lean code for Lemma6.121 theorem
  • theoremdefined in Noperthedron/Local/Coss.lean
    complete
    theorem Local.coss {ε θ θ_ φ φ_ : } {P Q : Euc(3)} (hP : P  1)
      (hQ : Q  1) ( : 0 < ε) ( : |θ - θ_|  ε) ( : |φ - φ_|  ε) :
      have M := rotM θ φ;
      have M_ := rotM θ_ φ_;
      0 <
          (inner  (M_ P) (M_ (P - Q)) - 2 * ε * P - Q * (2 + ε)) /
            ((M_ P + 2 * ε) * (M_ (P - Q) + 2 * 2 * ε)) 
        (inner  (M_ P) (M_ (P - Q)) - 2 * ε * P - Q * (2 + ε)) /
            ((M_ P + 2 * ε) * (M_ (P - Q) + 2 * 2 * ε)) 
          inner  (M P) (M (P - Q)) / (M P * M (P - Q))
    theorem Local.coss {ε θ θ_ φ φ_ : }
      {P Q : Euc(3)} (hP : P  1)
      (hQ : Q  1) ( : 0 < ε)
      ( : |θ - θ_|  ε)
      ( : |φ - φ_|  ε) :
      have M := rotM θ φ;
      have M_ := rotM θ_ φ_;
      0 <
          (inner  (M_ P) (M_ (P - Q)) -
              2 * ε * P - Q * (2 + ε)) /
            ((M_ P + 2 * ε) *
              (M_ (P - Q) + 2 * 2 * ε)) 
        (inner  (M_ P) (M_ (P - Q)) -
              2 * ε * P - Q * (2 + ε)) /
            ((M_ P + 2 * ε) *
              (M_ (P - Q) + 2 * 2 * ε)) 
          inner  (M P) (M (P - Q)) /
            (M P * M (P - Q))
    [SY25] Lemma 33 
Proof for Lemma 6.12
Proof uses 2
Proof dependency previews
Preview
Lemma 3.4
Loading preview
Proof dependency preview content is loaded from the rendered-fragment cache.

See polyhedron.without.rupert, Lemma 33.

Lemma6.13
uses 0used by 1L∃∀N

Let P_1,P_2,P_3, Q_1,Q_2,Q_3 \in \mathbb{R}^3. Define P := (P_1|P_2|P_3) and Q := (Q_1|Q_2|Q_3) and assume Q is invertible. Then P_1, P_2, P_3 and Q_1, Q_2, Q_3 are congruent iff P^t P = Q^t Q.

Lean code for Lemma6.131 theorem
  • complete
    theorem Local.congruent_iff_sym_matrix_eq (P Q : Local.Triangle)
      (hQ : Invertible Q.toMatrix) :
      P.Congruent Q  P.toSymMatrix = Q.toSymMatrix
    theorem Local.congruent_iff_sym_matrix_eq
      (P Q : Local.Triangle)
      (hQ : Invertible Q.toMatrix) :
      P.Congruent Q 
        P.toSymMatrix = Q.toSymMatrix
    [SY25] Lemma 35
    
Proof for Lemma 6.13
uses 0

From polyhedron.without.rupert, Lemma 35. Note that P^t P = Q^t Q is equivalent to saying that \langle P_i,P_j\rangle = \langle Q_i,Q_j\rangle for all 1 \leq i,j \leq 3. Moreover, the condition on invertibility of Q can be dropped, however then the proof becomes somewhat less straightforward.

If P_1, P_2, P_3 and Q_1, Q_2, Q_3 are congruent then \langle P_i,P_j\rangle = \langle LQ_i,LQ_j\rangle = \langle Q_i,Q_j\rangle, thus P^t P = Q^t Q. For the other direction, we claim that L \coloneqq PQ^{-1} is orthonormal and satisfies that P_i = LQ_i for all i=1,2,3. Indeed, L^t L = (PQ^{-1})^t (PQ^{-1}) = (Q^t)^{-1} P^t P Q^{-1} = \mathrm{Id} and it holds that LQ=PQ^{-1}Q=P, thus LQ_i = P_i.

Theorem6.14
uses 0used by 1L∃∀N

Let \PPP be a polyhedron with radius \rho=1 and P_1, P_2, P_3, Q_1, Q_2, Q_3 \in \PPP be not necessarily distinct. Assume that P_1, P_2, P_3 and Q_1, Q_2, Q_3 are congruent.

Let \epsilon>0 and \thetab_1,\phib_1,\thetab_2,\phib_2,\alphab \in \R, then set \Xib \coloneqq X(\thetab_1,\phib_1), \Xiib \coloneqq X(\thetab_2,\phib_2) as well as \Mib \coloneqq M(\thetab_1,\phib_1), \Miib \coloneqq M(\thetab_2,\phib_2). Assume that there exist \sigma_P, \sigma_Q \in \{0,1\} such that

(-1)^{\sigma_P} \langle \Xib,P_i\rangle>\sqrt{2}\varepsilon \quad \text{and} \quad (-1)^{\sigma_Q} \langle \Xiib , Q_i\rangle>\sqrt{2}\varepsilon, $$

for all i=1,2,3. Moreover, assume that P_1,P_2,P_3 are \epsilon-spanning for (\thetab_1,\phib_1) and that Q_1,Q_2,Q_3 are \epsilon-spanning for (\thetab_2,\phib_2). Finally, assume that for all i = 1,2,3 and any Q_j \in \PPP \setminus Q_i it holds that

\frac{\langle \Miib Q_i,\Miib (Q_i-Q_j)\rangle - 2 \epsilon \|Q_i-Q_j\| \cdot (\sqrt{2}+\varepsilon)}{ \big(\|\Miib Q_i\|+\sqrt{2} \varepsilon \big) \cdot \big(\|\Miib(Q_i-Q_j)\|+2 \sqrt{2} \varepsilon\big)} > \frac{\sqrt{5} \epsilon + \delta}{r}, $$

for some r >0 such that \min_{i=1,2,3}\| \Miib Q_i \| > r + \sqrt{2} \epsilon and for some \delta \in \R with

\delta \geq \max_{i=1,2,3}\left\|R(\alphab) \Mib P_i - \Miib Q_i\right\|/2. $$

Then there exists no solution to Rupert's problem R(\alpha) M(\theta_1,\phi_1)\PPP \subset M(\theta_2,\phi_2)\PPP^\circ with

(\theta_1, \phi_1, \theta_2, \phi_2, \alpha) \in [\thetab_1\pm\epsilon,\phib_1\pm\epsilon,\thetab_2\pm\epsilon,\phib_2\pm\epsilon,\alphab\pm\epsilon] \coloneqq U \subseteq \R^5.

Lean code for Theorem6.141 theorem
  • theoremdefined in Noperthedron/Local.lean
    complete
    theorem Local.local_theorem {ι : Type} [Fintype ι] [Nonempty ι]
      (poly : GoodPoly ι) (p_ : Pose ) (ε : )
      (pc : Local.LocalTheoremPrecondition poly p_ ε) :
      ¬ p  Metric.closedBall p_ ε, RupertPose p poly.hull
    theorem Local.local_theorem {ι : Type} [Fintype ι]
      [Nonempty ι] (poly : GoodPoly ι)
      (p_ : Pose ) (ε : )
      (pc :
        Local.LocalTheoremPrecondition poly p_
          ε) :
      ¬ p  Metric.closedBall p_ ε,
          RupertPose p poly.hull
    [SY25] Theorem 36
    
Proof for Theorem 6.14
Proof uses 8
Proof dependency previews

See polyhedron.without.rupert, Theorem 36.