AFTD Auto-Formalizing Theoretical Domains

cycle_power_dyadic_endpoint_overlap_of_no_island

theoremverified

Conjectured local uniform geometry: if all tight dyadic classes are unions of an initial and terminal interval, with arbitrary unequal widths, exact dyadic edges and full ordinary coverage force endpoint overlap. No periodicity or optimal scalar-count assumption is required.

Statement

theorem cycle_power_dyadic_endpoint_overlap_of_no_island
    (m : ℕ) (hm : 1 ≤ m) (x : ℤ → ℕ → Bool)
    (htight : ∀ p, cycle_power_dyadic_class_weight m (x p) =
      (Finset.range (2 ^ m)).sum (fun a => (x p a).toNat))
    (hexact : ∀ j < m, ∀ v : ℤ, (Finset.range (2 ^ j)).sum (fun a =>
      (x (v - (a : ℤ)) a && x (v - (a : ℤ)) (a + 2 ^ m - 2 ^ j)).toNat) = 1)
    (hcover : ∀ d : ℕ, 1 ≤ d → d < 2 ^ m → ∀ v : ℤ,
      ∃ a : ℕ, a < 2 ^ m - d ∧ x (v - (a : ℤ)) a = true ∧
        x (v - (a : ℤ)) (a + d) = true)
    (hshape : ∀ (p : ℤ) (a b c : ℕ), a < b → b < c → c < 2 ^ m →
      x p a = false → x p b = true → x p c = true) :
    ∀ (p : ℤ) (q : ℕ), 1 ≤ q → q < 2 ^ m →
      x (p + (q : ℤ)) (2 ^ m - 1 - q) = true →
      ∃ a : ℕ, a < q ∧ x (p + (q : ℤ)) (2 ^ m - q + a) = true ∧
        x (p + ((2 ^ m : ℕ) : ℤ)) a = true
Source
Conjectured uniform helper for OP-205 / RT-1, following arXiv:2610.15635. Experiments6472 and6481 suggest a split into global periodic two-interval classification and local endpoint geometry, allowing arbitrary unequal interval widths. The statements are unproved; finite SAT evidence is not a uniform theorem. No advance is claimed.
Verified
27 Oct 2026
Axioms
Classical.choiceQuot.soundpropext
Built on 3
Used by 1

The proof, explained

The proof first shows that every configuration has both boundary endpoints active ($x(p, 0) = \text{true}$ and $x(p, 2^m - 1) = \text{true}$) by specializing the exact dyadic edge matching hypothesis to scale $j = 0$. Given an active bit $x(p + q, 2^m - 1 - q) = \text{true}$, we apply the no-island corner lemma at step size $h = 2^m - q$, yielding two alternatives. If the next bit $x(p + q, 2^m - q)$ is active, setting $a = 0$ matches it against the active left endpoint of configuration $p + 2^m$; if instead the shifted corner bit is active, setting $a = q - 1$ matches the active right endpoint of configuration $p + q$ against the shifted configuration at position $q - 1$.

  1. For every shift $p \in \mathbb{Z}$, both boundary endpoints are active: $x(p, 0) = \text{true}$ and $x(p, 2^m - 1) = \text{true}$.

    We instantiate the exact dyadic edge matching hypothesis hexact at dyadic scale $j = 0$, which is valid since $m \ge 1$. Since $2^0 = 1$, the sum reduces to the single term at index $a = 0$: $(x(p, 0) \land x(p, 2^m - 1)).\text{toNat} = 1$. By case analysis on the boolean values, this equality forces both $x(p, 0) = \text{true}$ and $x(p, 2^m - 1) = \text{true}$.

    lines 18–21

    uses pow_zero, sum_range_one, Nat.cast_zero, sub_zero, Nat.zero_add

      have hroot (p : ℤ) : x p 0 = true ∧ x p (2^m-1) = true := by
        have hh := hexact 0 (by omega) p
        norm_num only [pow_zero,sum_range_one,Nat.cast_zero,sub_zero,Nat.zero_add] at hh
        cases h0 : x p 0 <;> cases h1 : x p (2^m-1) <;> simp_all
  2. Introduce $p$, $q$ with $1 \le q < 2^m$, and the hypothesis, rewriting the index to $x(p + q, 2^m - q - 1) = \text{true}$.

    We fix $p \in \mathbb{Z}$, $q \in \mathbb{N}$ with $1 \le q < 2^m$, and the assumption $x(p + q, 2^m - 1 - q) = \text{true}$. Because $q < 2^m$ and $1 \le q$, natural number arithmetic rewrites $2^m - 1 - q$ into $(2^m - q) - 1$, matching the form required to apply the corner lemma.

    lines 22–24

      intro p q hq0 hq1 hmem
      have hprev : x (p+(q : ℤ)) (2^m-q-1) = true := by
        convert hmem using 1 <;> congr 1 <;> omega
  3. By the no-island corner lemma with step $h = 2^m - q$, either $x(p + q, 2^m - q) = \text{true}$ or $x(p + q + (2^m - q), 2^m - (2^m - q) - 1) = \text{true}$.

    We apply cycle_power_dyadic_no_island_corner with shift $p + q$ and step $h = 2^m - q$. The bounds $1 \le 2^m - q < 2^m$ hold by arithmetic from $1 \le q < 2^m$, and the premise $x(p + q, h - 1) = \text{true}$ is our rewritten hypothesis. The lemma produces a disjunction between the two stated cases.

    lines 25–27

    uses cycle_power_dyadic_no_island_corner

      have hc := cycle_power_dyadic_no_island_corner m x htight hexact hshape hcover
        (p+(q : ℤ)) (2^m-q) (by omega) (by omega) hprev
      rcases hc with hc | hc
  4. In the first case, $a = 0$ is a valid witness.

    Assume $x(p + q, 2^m - q) = \text{true}$. We choose $a = 0$, which satisfies $a < q$ since $1 \le q$. The first required condition $x(p + q, 2^m - q + 0) = \text{true}$ is the case assumption itself, and the second condition $x(p + 2^m, 0) = \text{true}$ is the left endpoint property from hroot at shift $p + 2^m$.

    lines 28–29

    uses Nat.add_zero

      · refine ⟨0,by omega,?_,(hroot _).1⟩
        simpa only [Nat.add_zero] using hc
  5. In the second case, $a = q - 1$ is a valid witness.

    Assume $x(p + q + (2^m - q), 2^m - (2^m - q) - 1) = \text{true}$. We choose $a = q - 1$, which satisfies $q - 1 < q$ since $1 \le q$. The first condition $x(p + q, 2^m - q + (q - 1)) = \text{true}$ simplifies to the right endpoint $x(p + q, 2^m - 1) = \text{true}$, which holds by hroot. For the second condition, simplifying the shift $p + q + (2^m - q) = p + 2^m$ and the index $2^m - (2^m - q) - 1 = q - 1$ shows that the case assumption is precisely $x(p + 2^m, q - 1) = \text{true}$.

    lines 30–36

      · refine ⟨q-1,by omega,?_,?_⟩
        · have he : 2^m-q+(q-1)=2^m-1 := by omega
          rw [he]
          exact (hroot _).2
        · have hr : p+(q : ℤ)+((2^m-q : ℕ) : ℤ)=p+((2^m : ℕ) : ℤ) := by omega
          have hi : 2^m-(2^m-q)-1=q-1 := by omega
          rwa [hr,hi] at hc
pow_zero
Any element raised to the zeroth power equals 1 ($a^0 = 1$).
sum_range_one
A sum over the range of length 1 contains only the index 0, so $\sum_{k \in [0, 1)} f(k) = f(0)$.
Nat.cast_zero
The coercion of natural number 0 to a ring is 0.
sub_zero
Subtracting zero leaves an element unchanged ($a - 0 = a$).
Nat.zero_add
Zero is the left identity for natural addition ($0 + n = n$).
cycle_power_dyadic_no_island_corner
Under the tight dyadic weight, exact matching, no-island shape, and coverage assumptions, if the $(h-1)$-th bit of configuration $p$ is true for $1 \le h < 2^m$, then either the $h$-th bit $x(p, h)$ is true or the complementary boundary bit $x(p+h, 2^m - h - 1)$ of the shifted configuration is true.
Nat.add_zero
Zero is the right identity for natural addition ($n + 0 = n$).

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 Finsettheorem cycle_power_dyadic_endpoint_overlap_of_no_island    (m : ℕ) (hm : 1 ≤ m) (x : ℤ → ℕ → Bool)    (htight : ∀ p, cycle_power_dyadic_class_weight m (x p) =      (Finset.range (2 ^ m)).sum (fun a => (x p a).toNat))    (hexact : ∀ j < m, ∀ v : ℤ, (Finset.range (2 ^ j)).sum (fun a =>      (x (v - (a : ℤ)) a && x (v - (a : ℤ)) (a + 2 ^ m - 2 ^ j)).toNat) = 1)    (hcover : ∀ d : ℕ, 1 ≤ d → d < 2 ^ m → ∀ v : ℤ,      ∃ a : ℕ, a < 2 ^ m - d ∧ x (v - (a : ℤ)) a = true ∧        x (v - (a : ℤ)) (a + d) = true)    (hshape : ∀ (p : ℤ) (a b c : ℕ), a < b → b < c → c < 2 ^ m →      x p a = false → x p b = true → x p c = true) :    ∀ (p : ℤ) (q : ℕ), 1 ≤ q → q < 2 ^ m →      x (p + (q : ℤ)) (2 ^ m - 1 - q) = true →      ∃ a : ℕ, a < q ∧ x (p + (q : ℤ)) (2 ^ m - q + a) = true ∧        x (p + ((2 ^ m : ℕ) : ℤ)) a = true := by  have hroot (p : ℤ) : x p 0 = true ∧ x p (2^m-1) = true := by    have hh := hexact 0 (by omega) p    norm_num only [pow_zero,sum_range_one,Nat.cast_zero,sub_zero,Nat.zero_add] at hh    cases h0 : x p 0 <;> cases h1 : x p (2^m-1) <;> simp_all  intro p q hq0 hq1 hmem  have hprev : x (p+(q : ℤ)) (2^m-q-1) = true := by    convert hmem using 1 <;> congr 1 <;> omega  have hc := cycle_power_dyadic_no_island_corner m x htight hexact hshape hcover    (p+(q : ℤ)) (2^m-q) (by omega) (by omega) hprev  rcases hc with hc | hc  · refine ⟨0,by omega,?_,(hroot _).1⟩    simpa only [Nat.add_zero] using hc  · refine ⟨q-1,by omega,?_,?_⟩    · have he : 2^m-q+(q-1)=2^m-1 := by omega      rw [he]      exact (hroot _).2    · have hr : p+(q : ℤ)+((2^m-q : ℕ) : ℤ)=p+((2^m : ℕ) : ℤ) := by omega      have hi : 2^m-(2^m-q)-1=q-1 := by omega      rwa [hr,hi] at hc
See it in the graphDiscuss this resultChallenge it