AFTD Auto-Formalizing Theoretical Domains

op_258_exists_tail_le

theoremverified

For any positive profile with finite ENNReal sum and any ε > 0, there is a cutoff C such that any finite set of colours strictly above C has sum less than ε.

Statement

theorem op_258_exists_tail_le (a : ℕ+ → ℝ) (hpos : ∀ i, 0 < a i)
    (hsum_fin : (∑' i : ℕ+, ENNReal.ofReal (a i)) ≠ ⊤) (ε : ℝ) (hε : 0 < ε) :
    ∃ C : ℕ, ∀ I : Finset ℕ+, (∀ i ∈ I, C < (i : ℕ)) → (∑ i ∈ I, a i) < ε
Source
step towards op_258_forces_step
Verified
22 Oct 2026
Axioms
Classical.choiceQuot.soundpropext
Built on 1
  • op_258_boundedA colouring c of the edges of the complete graph K_n is a-bounded if for every colour i the number of edges of colour i is at most a i *…

The proof, explained

The theorem is proved by first establishing that the positive sequence a is summable in ℝ, translating the finiteness of its infinite sum in [0, ∞] through non-negative reals. By the vanishing criterion for summable series, there is a finite set of indices s outside of which all finite sums have norm strictly bounded by ε. Setting the cutoff C to be the maximum index in s ensures that any finite index set I with elements strictly above C is disjoint from s. Because every term of a is positive, the sum over I is non-negative, and thus bounded by ε without the norm.

  1. The profile sequence a is summable in ℝ.

    Since the sum of a in [0, ∞] is finite (hsum_fin), the sequence of non-negative reals i ↦ (a i).toNNReal is summable by ENNReal.tsum_coe_ne_top_iff_summable. By NNReal.summable_coe.2, coercing this sequence to ℝ preserves summability. Since each a i > 0 (hpos), converting a i to a non-negative real and coercing back to ℝ yields a i identically (Real.coe_toNNReal and le_of_lt). Replacing the coerced sequence with a shows that a is summable in ℝ.

    lines 5–13

    uses ENNReal.tsum_coe_ne_top_iff_summable, NNReal.summable_coe.2, Real.coe_toNNReal, le_of_lt

      have h_summable : Summable a := by
        have h_summable_nn : Summable (fun i => (a i).toNNReal) := by
          rwa [← ENNReal.tsum_coe_ne_top_iff_summable]
        have h_summable_real : Summable (fun i => ((a i).toNNReal : ℝ)) :=
          NNReal.summable_coe.2 h_summable_nn
        have heq : (fun i => ((a i).toNNReal : ℝ)) = a := by
          ext i
          exact Real.coe_toNNReal _ (le_of_lt (hpos i))
        rwa [heq] at h_summable_real
  2. There exists a finite index set s outside of which any finite sum has norm less than ε.

    By the vanishing criterion for summability in complete normed spaces (summable_iff_vanishing_norm.mp), for any ε > 0, there exists a finite set s ⊂ ℕ+ such that any finite set t disjoint from s satisfies ‖∑ i ∈ t, a i‖ < ε.

    line 14

    uses summable_iff_vanishing_norm.mp

      obtain ⟨s, hs⟩ := summable_iff_vanishing_norm.mp h_summable ε hε
  3. Setting the cutoff C to be the maximum index in s ensures that any finite set of indices strictly greater than C is disjoint from s.

    We choose the cutoff C = sup_{j ∈ s} j. Let I ⊂ ℕ+ be any finite set whose elements all satisfy i > C. To show I and s are disjoint (Finset.disjoint_left), suppose for contradiction that some i ∈ I also belongs to s. Then i ≤ C by the definition of the supremum (Finset.le_sup), contradicting the assumption C < i. Linear integer arithmetic (omega) finishes the contradiction.

    lines 15–22

    uses Finset.disjoint_left, Finset.le_sup

      use s.sup (fun j => (j : ℕ))
      intro I hI
      have hdisj : Disjoint I s := by
        rw [Finset.disjoint_left]
        intro i hi hs_in
        have h1 := Finset.le_sup (s := s) (f := fun j => (j : ℕ)) hs_in
        have h2 := hI i hi
        omega
  4. For any finite index set I with elements strictly greater than C, the sum ∑ i ∈ I, a i is strictly less than ε.

    Since I is disjoint from s, the vanishing property implies ‖∑ i ∈ I, a i‖ < ε. Because each term a i is strictly positive, the finite sum is non-negative by Finset.sum_nonneg (using le_of_lt). Hence its absolute value equals the sum itself by Real.norm_of_nonneg, yielding ∑ i ∈ I, a i < ε.

    lines 23–25

    uses Finset.sum_nonneg, le_of_lt, Real.norm_of_nonneg

      specialize hs I hdisj
      have hnonneg : 0 ≤ ∑ i ∈ I, a i := Finset.sum_nonneg fun i _ => le_of_lt (hpos i)
      rwa [Real.norm_of_nonneg hnonneg] at hs
ENNReal.tsum_coe_ne_top_iff_summable
A sequence of non-negative real numbers has a finite sum in [0, ∞] if and only if it is summable.
NNReal.summable_coe.2
If a sequence of non-negative real numbers is summable, then its coercion to the real numbers is also summable.
Real.coe_toNNReal
For any real number r ≥ 0, converting r to a non-negative real and coercing back to ℝ yields r.
le_of_lt
A strict inequality a < b implies the weak inequality a ≤ b.
summable_iff_vanishing_norm.mp
If a series is summable in a complete normed space, then for every ε > 0 there exists a finite index set s such that the norm of the sum over any finite set disjoint from s is strictly less than ε.
Finset.disjoint_left
Two finite sets s and t are disjoint if and only if no element of s belongs to t.
Finset.le_sup
Any element of a finite set evaluates under a function to at most the supremum of that function over the set.
Finset.sum_nonneg
A finite sum of non-negative terms is non-negative.
Real.norm_of_nonneg
The norm (absolute value) of a non-negative real number equals the number itself.

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

/-- For any positive profile with finite ENNReal sum and any ε > 0, there is a cutoff C such that any finite set of colours strictly above C has sum less than ε. -/theorem op_258_exists_tail_le (a : ℕ+ → ℝ) (hpos : ∀ i, 0 < a i)    (hsum_fin : (∑' i : ℕ+, ENNReal.ofReal (a i)) ≠ ⊤) (ε : ℝ) (hε : 0 < ε) :    ∃ C : ℕ, ∀ I : Finset ℕ+, (∀ i ∈ I, C < (i : ℕ)) → (∑ i ∈ I, a i) < ε := by  have h_summable : Summable a := by    have h_summable_nn : Summable (fun i => (a i).toNNReal) := by      rwa [← ENNReal.tsum_coe_ne_top_iff_summable]    have h_summable_real : Summable (fun i => ((a i).toNNReal : ℝ)) :=      NNReal.summable_coe.2 h_summable_nn    have heq : (fun i => ((a i).toNNReal : ℝ)) = a := by      ext i      exact Real.coe_toNNReal _ (le_of_lt (hpos i))    rwa [heq] at h_summable_real  obtain ⟨s, hs⟩ := summable_iff_vanishing_norm.mp h_summable ε hε  use s.sup (fun j => (j : ℕ))  intro I hI  have hdisj : Disjoint I s := by    rw [Finset.disjoint_left]    intro i hi hs_in    have h1 := Finset.le_sup (s := s) (f := fun j => (j : ℕ)) hs_in    have h2 := hI i hi    omega  specialize hs I hdisj  have hnonneg : 0 ≤ ∑ i ∈ I, a i := Finset.sum_nonneg fun i _ => le_of_lt (hpos i)  rwa [Real.norm_of_nonneg hnonneg] at hs
See it in the graphDiscuss this resultChallenge it