AFTD Auto-Formalizing Theoretical Domains

energy_filter_quartic_tannery

theoremverified

For an absolutely summable quartic cycle, a uniformly bounded shifted kernel with a common pointwise limit passes through the cycle sum by dominated convergence.

Statement

theorem energy_filter_quartic_tannery
    (a : ℤ → ℝ) (K : ℕ → ℤ → ℤ → ℤ → ℝ) (c C : ℝ) (hC : 0 ≤ C)
    (hab : Summable (fun p : ℤ × ℤ × ℤ =>
      |a p.1 * a p.2.1 * a p.2.2 * a (p.1 - p.2.1 + p.2.2)|))
    (hlim : ∀ u v w, Filter.Tendsto (fun k => K k u v w) Filter.atTop (nhds c))
    (hbound : ∀ᶠ k in Filter.atTop, ∀ u v w, |K k u v w| ≤ C) :
    Filter.Tendsto
      (fun k : ℕ => ∑' p : ℤ × ℤ × ℤ,
        (a p.1 * a p.2.1 * a p.2.2 * a (p.1 - p.2.1 + p.2.2)) *
          K k p.1 p.2.1 (p.2.1 - p.2.2))
      Filter.atTop
      (nhds ((∑' p : ℤ × ℤ × ℤ,
        a p.1 * a p.2.1 * a p.2.2 * a (p.1 - p.2.1 + p.2.2)) * c))
Source
OP-892: shifted quartic reindexing and dominated convergence.
Verified
14 Nov 2026
Axioms
Classical.choiceQuot.soundpropext
Used by 1

The proof, explained

The proof applies Tannery's theorem—the dominated convergence theorem for infinite series (tendsto_tsum_of_dominated_convergence). The quartic summand Q(p) is absolutely summable by assumption, which yields a summable majorant p ↦ C * |Q(p)|. Because the shifted kernel converges pointwise to c and is eventually uniformly bounded by C, the dominated convergence theorem allows the limit as k → ∞ to pass inside the summation. Finally, the scalar c is factored out of the resulting series by linearity of summation.

  1. The quartic term Q(p) is absolutely summable (hence summable), and p ↦ C * |Q(p)| is a summable dominating function.

    Define Q(p) = a(p₁) a(p₂) a(p₃) a(p₁ - p₂ + p₃) for p = (p₁, p₂, p₃) ∈ ℤ³. The hypothesis hab asserts that p ↦ |Q(p)| is summable. Absolute summability in ℝ implies unconditional summability, so Q itself is summable. Scaling this summable series by the non-negative constant C gives that p ↦ C * |Q(p)| is also summable, providing the dominating majorant.

    lines 14–18

      let Q : ℤ × ℤ × ℤ → ℝ := fun p =>
        a p.1 * a p.2.1 * a p.2.2 * a (p.1 - p.2.1 + p.2.2)
      have hQabs : Summable (fun p => |Q p|) := by simpa [Q] using hab
      have hQ : Summable Q := hQabs.of_abs
      have hsum : Summable (fun p => C * |Q p|) := hQabs.mul_left C
  2. For each fixed p, the term Q(p) * K(k, p₁, p₂, p₂ - p₃) converges to Q(p) * c as k → ∞.

    For any fixed triple p = (p₁, p₂, p₃), the shifted kernel K(k, p₁, p₂, p₂ - p₃) tends to c as k → ∞ by hypothesis hlim. Multiplying this limit by the constant Q(p) using tendsto_const_nhds and the continuity of multiplication shows that the product converges to Q(p) * c.

    lines 19–23

    uses tendsto_const_nhds

      have hpoint : ∀ p : ℤ × ℤ × ℤ,
          Filter.Tendsto (fun k : ℕ => Q p * K k p.1 p.2.1 (p.2.1 - p.2.2))
            Filter.atTop (nhds (Q p * c)) := by
        intro p
        exact tendsto_const_nhds.mul (hlim p.1 p.2.1 (p.2.1 - p.2.2))
  3. Eventually for large k, the norm ‖Q(p) * K(k, p₁, p₂, p₂ - p₃)‖ is bounded by C * |Q(p)| for all p.

    By hypothesis hbound, the kernel values are eventually uniformly bounded by C for all inputs. Using Real.norm_eq_abs and abs_mul, the norm of the summand is |Q(p)| * |K(k, …)|. Since |Q(p)| ≥ 0 by abs_nonneg, multiplying the kernel bound by |Q(p)| via mul_le_mul_of_nonneg_left gives |Q(p)| * |K(k, …)| ≤ |Q(p)| * C = C * |Q(p)| by commutativity of multiplication (mul_comm).

    lines 24–31

    uses Real.norm_eq_abs, abs_mul, mul_le_mul_of_nonneg_left, abs_nonneg, mul_comm

      have hbd : ∀ᶠ k : ℕ in Filter.atTop, ∀ p : ℤ × ℤ × ℤ,
          ‖Q p * K k p.1 p.2.1 (p.2.1 - p.2.2)‖ ≤ C * |Q p| := by
        filter_upwards [hbound] with k hk p
        rw [Real.norm_eq_abs, abs_mul]
        calc
          |Q p| * |K k p.1 p.2.1 (p.2.1 - p.2.2)| ≤ |Q p| * C :=
            mul_le_mul_of_nonneg_left (hk p.1 p.2.1 (p.2.1 - p.2.2)) (abs_nonneg _)
          _ = C * |Q p| := mul_comm _ _
  4. The infinite series converges to (∑' p, Q(p)) * c as k → ∞.

    Applying the dominated convergence theorem for infinite series (tendsto_tsum_of_dominated_convergence) with the summable dominating bound C * |Q(p)|, the pointwise limits, and the eventual uniform domination establishes that the sum of the series converges to ∑' p, (Q(p) * c). Factoring out the constant scalar c from the sum rewrites the limit value as (∑' p, Q(p)) * c, which matches the theorem goal.

    lines 32–35

    uses tendsto_tsum_of_dominated_convergence

      have ht := tendsto_tsum_of_dominated_convergence hsum hpoint hbd
      have hs : (∑' p, Q p * c) = (∑' p, Q p) * c := hQ.tsum_mul_right c
      rw [hs] at ht
      exact ht
tendsto_const_nhds
The constant sequence with value x converges to x along any filter.
Real.norm_eq_abs
On the real numbers, the norm of an element is equal to its absolute value.
abs_mul
The absolute value of a product in an ordered ring equals the product of the absolute values.
mul_le_mul_of_nonneg_left
Multiplying an inequality on the left by a non-negative element preserves the inequality.
abs_nonneg
The absolute value of an element is always non-negative.
mul_comm
Multiplication in a commutative magma is commutative: a * b = b * a.
tendsto_tsum_of_dominated_convergence
If a family of functions converges pointwise and is eventually bounded in norm by a summable function, then the sum of the series converges to the sum of the pointwise limits.

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

theorem energy_filter_quartic_tannery    (a : ℤ → ℝ) (K : ℕ → ℤ → ℤ → ℤ → ℝ) (c C : ℝ) (hC : 0 ≤ C)    (hab : Summable (fun p : ℤ × ℤ × ℤ =>      |a p.1 * a p.2.1 * a p.2.2 * a (p.1 - p.2.1 + p.2.2)|))    (hlim : ∀ u v w, Filter.Tendsto (fun k => K k u v w) Filter.atTop (nhds c))    (hbound : ∀ᶠ k in Filter.atTop, ∀ u v w, |K k u v w| ≤ C) :    Filter.Tendsto      (fun k : ℕ => ∑' p : ℤ × ℤ × ℤ,        (a p.1 * a p.2.1 * a p.2.2 * a (p.1 - p.2.1 + p.2.2)) *          K k p.1 p.2.1 (p.2.1 - p.2.2))      Filter.atTop      (nhds ((∑' p : ℤ × ℤ × ℤ,        a p.1 * a p.2.1 * a p.2.2 * a (p.1 - p.2.1 + p.2.2)) * c)) := by  let Q : ℤ × ℤ × ℤ → ℝ := fun p =>    a p.1 * a p.2.1 * a p.2.2 * a (p.1 - p.2.1 + p.2.2)  have hQabs : Summable (fun p => |Q p|) := by simpa [Q] using hab  have hQ : Summable Q := hQabs.of_abs  have hsum : Summable (fun p => C * |Q p|) := hQabs.mul_left C  have hpoint : ∀ p : ℤ × ℤ × ℤ,      Filter.Tendsto (fun k : ℕ => Q p * K k p.1 p.2.1 (p.2.1 - p.2.2))        Filter.atTop (nhds (Q p * c)) := by    intro p    exact tendsto_const_nhds.mul (hlim p.1 p.2.1 (p.2.1 - p.2.2))  have hbd : ∀ᶠ k : ℕ in Filter.atTop, ∀ p : ℤ × ℤ × ℤ,      ‖Q p * K k p.1 p.2.1 (p.2.1 - p.2.2)‖ ≤ C * |Q p| := by    filter_upwards [hbound] with k hk p    rw [Real.norm_eq_abs, abs_mul]    calc      |Q p| * |K k p.1 p.2.1 (p.2.1 - p.2.2)| ≤ |Q p| * C :=        mul_le_mul_of_nonneg_left (hk p.1 p.2.1 (p.2.1 - p.2.2)) (abs_nonneg _)      _ = C * |Q p| := mul_comm _ _  have ht := tendsto_tsum_of_dominated_convergence hsum hpoint hbd  have hs : (∑' p, Q p * c) = (∑' p, Q p) * c := hQ.tsum_mul_right c  rw [hs] at ht  exact ht
See it in the graphDiscuss this resultChallenge it