AFTD Auto-Formalizing Theoretical Domains

recurrence_potential_weights_pos

theoremverified

For any recurrence satisfying (2.1), the potential weights π_i are strictly positive for all i : Fin (N + 1).

Statement

theorem recurrence_potential_weights_pos (N : ℕ) (a b c : Fin (N + 1) → ℝ)
    (hrec : satisfies_recurrence_2_1 N a b c) (i : Fin (N + 1)) :
    0 < recurrence_potential_weights N a c i
Source
arXiv:2611.01561, Proof of Lemma 2.1
Verified
07 Nov 2026
Axioms
Classical.choiceQuot.soundpropext
Built on 2
  • satisfies_recurrence_2_1Real coefficients a, b, c : Fin (N + 1) → ℝ satisfy condition (2.1) if a 0 = 1, b 0 = 0, c 0 = 0, a N = 0, a i + b i + c i = 1 for all i,…
  • recurrence_potential_weightsFor N : ℕ, a, c : Fin (N + 1) → ℝ, and i : Fin (N + 1), the potential weight π_i is the product over j < i.val of a(j) / c(j + 1).

Read back from the Lean

For any natural number N, any functions a, b, c : Fin (N + 1) → ℝ, if (a, b, c) satisfies satisfies_recurrence_2_1 N a b c, then for every index i : Fin (N + 1), the potential weight recurrence_potential_weights N a c i is strictly positive (> 0).

Written by a model that saw only the Lean, never the English above. If the two disagree, that is worth a challenge.

The proof, explained

By definition, the potential weight π_i is the finite product of ratios a_j / c_{j+1} for 0 ≤ j < i. To show that this product is strictly positive, it suffices to show that every factor is strictly positive. Because j < i ≤ N, we have j < N and j + 1 > 0, so the recurrence condition (2.1) guarantees that both a_j and c_{j+1} are strictly positive, making their quotient positive.

  1. It suffices to show that each factor in the product defining the potential weight is strictly positive.

    Unfolding the definition of recurrence_potential_weights, the weight π_i is given by the finite product ∏_{j < i} (a_j / c_{j+1}). By Finset.prod_pos, a product of strictly positive numbers is strictly positive, so it suffices to fix an arbitrary index j < i and show that the ratio a_j / c_{j+1} is positive.

    lines 5–7

    uses Finset.prod_pos

      unfold recurrence_potential_weights
      apply Finset.prod_pos
      intro j _
  2. Extract the positivity hypotheses from recurrence condition (2.1) and record the bounds on j.

    From the recurrence hypothesis hrec, we unpack the conditions that a_k > 0 for all k < N (hypothesis ha) and c_k > 0 for all k > 0 (hypothesis hc). We also record the inequalities j < i and i < N + 1 from the definitions of the finite index types.

    lines 8–10

      obtain ⟨-, -, -, -, -, ha, hc⟩ := hrec
      have hj : j.val < i.val := j.isLt
      have hi : i.val < N + 1 := i.isLt
  3. Each factor a_j / c_{j+1} is strictly positive.

    By div_pos, the quotient is positive if both numerator and denominator are positive. For the numerator a_j, since j < i < N + 1, linear arithmetic gives j < N, so ha yields a_j > 0. For the denominator c_{j+1}, since j ≥ 0, linear arithmetic gives j + 1 > 0, so hc yields c_{j+1} > 0.

    lines 11–13

    uses div_pos

      apply div_pos
      · exact ha _ (by show j.val < N; omega)
      · exact hc _ (by show 0 < j.val + 1; omega)
Finset.prod_pos
A finite product of strictly positive elements is strictly positive.
div_pos
The quotient of two strictly positive elements is strictly positive.

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

/-- The potential weights are strictly positive under condition (2.1). -/theorem recurrence_potential_weights_pos (N : ℕ) (a b c : Fin (N + 1) → ℝ)    (hrec : satisfies_recurrence_2_1 N a b c) (i : Fin (N + 1)) :    0 < recurrence_potential_weights N a c i := by  unfold recurrence_potential_weights  apply Finset.prod_pos  intro j _  obtain ⟨-, -, -, -, -, ha, hc⟩ := hrec  have hj : j.val < i.val := j.isLt  have hi : i.val < N + 1 := i.isLt  apply div_pos  · exact ha _ (by show j.val < N; omega)  · exact hc _ (by show 0 < j.val + 1; omega)
See it in the graphDiscuss this resultChallenge it