AFTD Auto-Formalizing Theoretical Domains

four_point_positive_kernel_boundary_decomposition

theoremverified

For two integer rows on four coordinates with independent first two columns and a strictly positive integer common kernel vector x, there are two nonzero nonnegative integer common kernel vectors u,v, each having a zero coordinate, and positive integer D,c,d (represented in Z) such that D*x=c*u+d*v. The two vectors are not asserted primitive in this lemma.

Statement

theorem four_point_positive_kernel_boundary_decomposition
    (a b x : Fin 4 → ℤ) (hx : ∀ i, 0 < x i)
    (hxa : (∑ i, x i*a i) = 0) (hxb : (∑ i, x i*b i) = 0)
    (hind : a 0*b 1 - a 1*b 0 ≠ 0) :
    ∃ (u v : Fin 4 → ℤ) (i j : Fin 4) (D c d : ℤ),
      (∀ k, 0 ≤ u k) ∧ (∀ k, 0 ≤ v k) ∧ u ≠ 0 ∧ v ≠ 0 ∧
      u i = 0 ∧ v j = 0 ∧ 0 < D ∧ 0 < c ∧ 0 < d ∧
      (∀ k, D*x k = c*u k + d*v k) ∧
      (∑ k, u k*a k) = 0 ∧ (∑ k, u k*b k) = 0 ∧
      (∑ k, v k*a k) = 0 ∧ (∑ k, v k*b k) = 0
Source
Supporting lemma or explicit route obstruction for OP-197 / arXiv:2610.16562 Conjecture1, deep session2. The abstract two-ray fact and finite obstruction are not claimed novel or research-level; no part is the full box conjecture. The five-point kernel has columns (1,1),(1,0),(0,1),(-2,0),(0,-2).
Verified
28 Oct 2026
Axioms
Classical.choiceQuot.soundpropext
Built on 2
Used by 1

The proof, explained

The proof decomposes a strictly positive common kernel vector x into boundary kernel vectors by first constructing a direction vector w in the kernel of a and b that has both positive and negative entries (via `four_point_positive_kernel_mixed_direction`). Intersecting the line through x along direction w with the boundary of the nonnegative orthant (using `integer_positive_line_boundary_decomposition`) yields nonnegative vectors u and v, each possessing a zero coordinate, along with positive integer scalars such that D * x = c * u + d * v. Finally, because u and v are integer linear combinations of x and w—both of which lie in the kernel of a and b—linearity implies that u and v also lie in the common kernel.

  1. There exists a nontrivial integer vector w orthogonal to a and b with coordinate sum zero that has both strictly positive and strictly negative entries.

    Applying the lemma `four_point_positive_kernel_mixed_direction` to the rows a and b and the strictly positive kernel vector x produces an integer vector w satisfying ∑ i, w i = 0, ∑ i, w i * a i = 0, ∑ i, w i * b i = 0, with at least one coordinate w i > 0 and at least one coordinate w j < 0.

    line 15

    uses four_point_positive_kernel_mixed_direction

      obtain ⟨w,hsw,hwa,hwb,hp,hn⟩ := four_point_positive_kernel_mixed_direction a b x hx hxa hxb hind
  2. There exist nonnegative nonzero vectors u, v, indices i, j, and positive integers D, c, d expressing D * x = c * u + d * v, such that u i = 0 and v j = 0.

    Applying `integer_positive_line_boundary_decomposition` with the positive vector x and the mixed-direction vector w yields boundary vectors u and v defined coordinatewise by u k = w i * x k - x i * w k and v k = x j * w k - w j * x k. They are nonnegative, nonzero, satisfy u i = 0 and v j = 0, and express D * x = c * u + d * v for positive integers D, c, d.

    lines 16–17

    uses integer_positive_line_boundary_decomposition

      obtain ⟨u,v,i,j,D,c,d,hu,hv,hun,hvn,hui,hvj,hD,hc,hd,hrel,hue,hve⟩ :=
        integer_positive_line_boundary_decomposition x w hx hp hn
  3. Any linear combination p * x - q * w is orthogonal to any vector r that is orthogonal to both x and w.

    We establish a local helper lemma `hsum`: for any vector r orthogonal to x and w, the sum ∑ k, (p * x k - q * w k) * r k distributes as p * (∑ k, x k * r k) - q * (∑ k, w k * r k). Since both sums are zero by hypothesis, the whole combination simplifies to p * 0 - q * 0 = 0.

    lines 18–25

    uses Finset.mul_sum, Finset.sum_sub_distrib, Finset.sum_congr

      have hsum (r : Fin 4 → ℤ) (hxr : (∑ k, x k*r k)=0) (hwr : (∑ k, w k*r k)=0)
          (p q : ℤ) : (∑ k, (p*x k-q*w k)*r k)=0 := by
        calc
          _ = p*(∑ k, x k*r k)-q*(∑ k, w k*r k) := by
            simp only [Finset.mul_sum, ← Finset.sum_sub_distrib]
            apply Finset.sum_congr rfl
            intro k hk; ring
          _ = 0 := by rw [hxr,hwr]; ring
  4. The witnesses u, v, i, j, D, c, d satisfy all conditions except orthogonality to a and b, which reduces to four remaining subgoals.

    We provide u, v, i, j, D, c, d as existential witnesses. Their nonnegativity, nonzeroness, boundary zeroes at indices i and j, positivity of D, c, d, and the conic combination relation D * x = c * u + d * v are already established by the line boundary decomposition. This leaves four goals: showing that u and v are orthogonal to a and b.

    line 26

      refine ⟨u,v,i,j,D,c,d,hu,hv,hun,hvn,hui,hvj,hD,hc,hd,hrel,?_,?_,?_,?_⟩
  5. u is orthogonal to both a and b.

    By definition, u k = w i * x k - x i * w k, which is the linear combination (w i) * x - (x i) * w. Because both x and w are orthogonal to a and to b, the helper lemma `hsum` directly proves ∑ k, u k * a k = 0 and ∑ k, u k * b k = 0.

    lines 27–28

      · rw [hue]; exact hsum a hxa hwa (w i) (x i)
      · rw [hue]; exact hsum b hxb hwb (w i) (x i)
  6. v is orthogonal to a.

    Rewriting v k as x j * w k - w j * x k, we note that x j * w k - w j * x k = -(w j * x k - x j * w k). Applying `hsum` shows that ∑ k, (w j * x k - x j * w k) * a k = 0. Factoring out the negation using `Finset.sum_neg_distrib` then yields ∑ k, v k * a k = -0 = 0.

    lines 29–35

    uses Finset.sum_neg_distrib, Finset.sum_congr

      · rw [hve]
        have he := hsum a hxa hwa (w j) (x j)
        have hn : (∑ k, (x j*w k-w j*x k)*a k) = -(∑ k, (w j*x k-x j*w k)*a k) := by
          rw [← Finset.sum_neg_distrib]
          apply Finset.sum_congr rfl
          intro k hk; ring
        rw [hn,he]; ring
  7. v is orthogonal to b.

    In the same manner, substituting the definition of v and factoring out the negation, the sum ∑ k, v k * b k equals -∑ k, (w j * x k - x j * w k) * b k. By `hsum`, this sum is 0, giving -0 = 0.

    lines 36–42

    uses Finset.sum_neg_distrib, Finset.sum_congr

      · rw [hve]
        have he := hsum b hxb hwb (w j) (x j)
        have hn : (∑ k, (x j*w k-w j*x k)*b k) = -(∑ k, (w j*x k-x j*w k)*b k) := by
          rw [← Finset.sum_neg_distrib]
          apply Finset.sum_congr rfl
          intro k hk; ring
        rw [hn,he]; ring
Finset.mul_sum
Multiplication distributes over finite sums: a * ∑ i, f i = ∑ i, a * f i.
Finset.sum_congr
Two finite sums over identical sets are equal if their summands agree at every element.
Finset.sum_neg_distrib
Negation distributes over finite sums: ∑ i, -f i = -∑ i, f i.
Finset.sum_sub_distrib
Subtraction distributes over finite sums: ∑ i, (f i - g i) = ∑ i, f i - ∑ i, g i.
four_point_positive_kernel_mixed_direction
Given two 4-vectors a, b with independent first two entries and a strictly positive common kernel vector x, there exists a vector w orthogonal to a and b whose coordinates sum to zero and which has both strictly positive and strictly negative entries.
integer_positive_line_boundary_decomposition
Given a strictly positive vector x and a vector w with both positive and negative entries, moving along the line x + t w hits the boundary of the positive orthant in two nonzero nonnegative vectors u, v with coordinate zeroes at i, j, expressing a positive integer multiple D * x as a conical combination c * u + d * v with c, d > 0.

Written by a model (gemini-3.8-flash) from the Lean proof and the goal states Lean reported at each step. Lean verified the proof; this explanation is a reading aid and is not itself machine-checked.

Lean source view module on GitHub

open Finsetset_option maxHeartbeats 1000000set_option maxRecDepth 4000theorem four_point_positive_kernel_boundary_decomposition    (a b x : Fin 4 → ℤ) (hx : ∀ i, 0 < x i)    (hxa : (∑ i, x i*a i) = 0) (hxb : (∑ i, x i*b i) = 0)    (hind : a 0*b 1 - a 1*b 0 ≠ 0) :    ∃ (u v : Fin 4 → ℤ) (i j : Fin 4) (D c d : ℤ),      (∀ k, 0 ≤ u k) ∧ (∀ k, 0 ≤ v k) ∧ u ≠ 0 ∧ v ≠ 0 ∧      u i = 0 ∧ v j = 0 ∧ 0 < D ∧ 0 < c ∧ 0 < d ∧      (∀ k, D*x k = c*u k + d*v k) ∧      (∑ k, u k*a k) = 0 ∧ (∑ k, u k*b k) = 0 ∧      (∑ k, v k*a k) = 0 ∧ (∑ k, v k*b k) = 0 := by  obtain ⟨w,hsw,hwa,hwb,hp,hn⟩ := four_point_positive_kernel_mixed_direction a b x hx hxa hxb hind  obtain ⟨u,v,i,j,D,c,d,hu,hv,hun,hvn,hui,hvj,hD,hc,hd,hrel,hue,hve⟩ :=    integer_positive_line_boundary_decomposition x w hx hp hn  have hsum (r : Fin 4 → ℤ) (hxr : (∑ k, x k*r k)=0) (hwr : (∑ k, w k*r k)=0)      (p q : ℤ) : (∑ k, (p*x k-q*w k)*r k)=0 := by    calc      _ = p*(∑ k, x k*r k)-q*(∑ k, w k*r k) := by        simp only [Finset.mul_sum, ← Finset.sum_sub_distrib]        apply Finset.sum_congr rfl        intro k hk; ring      _ = 0 := by rw [hxr,hwr]; ring  refine ⟨u,v,i,j,D,c,d,hu,hv,hun,hvn,hui,hvj,hD,hc,hd,hrel,?_,?_,?_,?_⟩  · rw [hue]; exact hsum a hxa hwa (w i) (x i)  · rw [hue]; exact hsum b hxb hwb (w i) (x i)  · rw [hve]    have he := hsum a hxa hwa (w j) (x j)    have hn : (∑ k, (x j*w k-w j*x k)*a k) = -(∑ k, (w j*x k-x j*w k)*a k) := by      rw [← Finset.sum_neg_distrib]      apply Finset.sum_congr rfl      intro k hk; ring    rw [hn,he]; ring  · rw [hve]    have he := hsum b hxb hwb (w j) (x j)    have hn : (∑ k, (x j*w k-w j*x k)*b k) = -(∑ k, (w j*x k-x j*w k)*b k) := by      rw [← Finset.sum_neg_distrib]      apply Finset.sum_congr rfl      intro k hk; ring    rw [hn,he]; ring
See it in the graphDiscuss this resultChallenge it