Hyperharmonic Euler sums
DefinitionGeneralized hyperharmonic numbers#✓fully proved
For natural numbers \(n\) and \(k\) and a complex number \(r\), the generalized hyperharmonic number \(H_n^{[r]}(k)\) is \(0\) when \(k > n\); when \(k \le n\) it is the elementary symmetric polynomial of degree \(n - k\) in \(r, r + 1, \ldots, r + n - 1\) divided by \(n!\), that is,
The source defines \(H_n^{[r]}(k)\) in (2.4) by \((-1)^k k! \, H_n^{[r]}(k) = D_x^k \binom{n+r-x-1}{n}\) at \(x = 0\), where \(D_x^k\) is the \(k\)-th derivative in \(x\); the explicit form above follows from it. For \(r = 1\) it is the depth-\(k\) multiple harmonic number, and for \(r \notin \{0, -1, \ldots, 1 - n\}\) it equals \(\binom{n+r-1}{n} \sum_{1 \le i_1 < \cdots < i_k \le n} \prod_{j=1}^{k} 1/(i_j + r - 1)\), see (2.10).
noncomputable def generalizedHyperharmonic (n k : ℕ) (r : ℂ) : ℂ :=
if k ≤ n then
(∑ s ∈ (Finset.range n).powersetCard (n - k), ∏ j ∈ s, (r + j)) / n.factorial
else 0
Sources
- Paul Thomas Young, Parametric Euler Sums of Generalized Hyperharmonic Numbers, INTEGERS 26 (2026), Article A81, Section 2, (2.4) and (2.10).
Used by 14
TheoremGeneralized hyperharmonic numbers vanish for \(k > n\)#✓fully proved
For all natural numbers \(n\) and \(k\) with \(n < k\) and every complex number \(r\), the generalized hyperharmonic number \(H_n^{[r]}(k)\) equals \(0\).
theorem generalizedHyperharmonic_of_lt {n k : ℕ} (h : n < k) (r : ℂ) :
generalizedHyperharmonic n k r = 0
noncomputable def generalizedHyperharmonic (n k : ℕ) (r : ℂ) : ℂ :=
if k ≤ n then
(∑ s ∈ (Finset.range n).powersetCard (n - k), ∏ j ∈ s, (r + j)) / n.factorial
else 0
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:34
Sources
- Paul Thomas Young, Parametric Euler Sums of Generalized Hyperharmonic Numbers, INTEGERS 26 (2026), Article A81, Section 2, (2.4).
Uses 1 · Used by 1
TheoremGeneralized hyperharmonic numbers with both indices zero#✓fully proved
For every complex number \(r\), the generalized hyperharmonic number \(H_0^{[r]}(0)\) equals \(1\).
@[simp]
theorem generalizedHyperharmonic_zero_zero (r : ℂ) : generalizedHyperharmonic 0 0 r = 1
noncomputable def generalizedHyperharmonic (n k : ℕ) (r : ℂ) : ℂ :=
if k ≤ n then
(∑ s ∈ (Finset.range n).powersetCard (n - k), ∏ j ∈ s, (r + j)) / n.factorial
else 0
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:34
Sources
- Paul Thomas Young, Parametric Euler Sums of Generalized Hyperharmonic Numbers, INTEGERS 26 (2026), Article A81, Section 2, (2.4).
Uses 1
TheoremThe value \(H_1^{(r)}(0) = r\)#✓fully proved
For every complex number \(r\), the generalized hyperharmonic number \(H_1^{[r]}(0)\) equals \(r\).
theorem generalizedHyperharmonic_one_zero (r : ℂ) : generalizedHyperharmonic 1 0 r = r
noncomputable def generalizedHyperharmonic (n k : ℕ) (r : ℂ) : ℂ :=
if k ≤ n then
(∑ s ∈ (Finset.range n).powersetCard (n - k), ∏ j ∈ s, (r + j)) / n.factorial
else 0
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:34
Sources
- Paul Thomas Young, Parametric Euler Sums of Generalized Hyperharmonic Numbers, INTEGERS 26 (2026), Article A81, Section 2, (2.4).
Uses 1
TheoremThe value \(H_1^{(r)}(1) = 1\)#✓fully proved
For every complex number \(r\), the generalized hyperharmonic number \(H_1^{[r]}(1)\) equals \(1\).
theorem generalizedHyperharmonic_one_one (r : ℂ) : generalizedHyperharmonic 1 1 r = 1
noncomputable def generalizedHyperharmonic (n k : ℕ) (r : ℂ) : ℂ :=
if k ≤ n then
(∑ s ∈ (Finset.range n).powersetCard (n - k), ∏ j ∈ s, (r + j)) / n.factorial
else 0
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:34
Sources
- Paul Thomas Young, Parametric Euler Sums of Generalized Hyperharmonic Numbers, INTEGERS 26 (2026), Article A81, Section 2, (2.4).
Uses 1
DefinitionGeneralized hyperharmonic Euler sum#✓fully proved
For an integer \(k \ge 0\), complex numbers \(r\) and \(a\) and an integer \(p\), the generalized hyperharmonic Euler sum is \(S(k, r; p; a) = \sum_{n \ge k} H_n^{[r]}(k) / (n + a)^p\), where \(H_n^{[r]}(k)\) is the generalized hyperharmonic number. A term whose denominator \((n + a)^p\) vanishes counts as \(0\), and \(S(k, r; p; a) = 0\) when the series is not summable.
noncomputable def hyperharmonicEulerSum (k : ℕ) (r : ℂ) (p : ℤ) (a : ℂ) : ℂ :=
∑' n : ℕ, generalizedHyperharmonic (n + k) k r / (((n + k : ℕ) : ℂ) + a) ^ p
noncomputable def generalizedHyperharmonic (n k : ℕ) (r : ℂ) : ℂ :=
if k ≤ n then
(∑ s ∈ (Finset.range n).powersetCard (n - k), ∏ j ∈ s, (r + j)) / n.factorial
else 0
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:34
Sources
- Paul Thomas Young, Parametric Euler Sums of Generalized Hyperharmonic Numbers, INTEGERS 26 (2026), Article A81, Corollary 1 display.
Uses 1 · Used by 9
TheoremHyperharmonic Euler sums as sums over all indices#✓fully proved
For every integer \(k \ge 0\), complex numbers \(r\) and \(a\), and integer \(p\), write \(H_n^{[r]}(k)\) for the generalized hyperharmonic number. Then the hyperharmonic Euler sum \(S(k, r; p; a) = \sum_{n \ge k} H_n^{[r]}(k) / (n + a)^p\) equals \(\sum_{n \ge 0} H_n^{[r]}(k) / (n + a)^p\), because \(H_n^{[r]}(k) = 0\) for \(n < k\).
Both sides are the sums of the series in the sense of unconditional summation, with the value \(0\) assigned to a series that does not converge, and a term with \(n + a = 0\) and \(p \ne 0\) is read as \(0\), so the identity holds without any hypothesis.
theorem hyperharmonicEulerSum_eq_tsum (k : ℕ) (r : ℂ) (p : ℤ) (a : ℂ) :
hyperharmonicEulerSum k r p a =
∑' n : ℕ, generalizedHyperharmonic n k r / ((n : ℂ) + a) ^ p
noncomputable def hyperharmonicEulerSum (k : ℕ) (r : ℂ) (p : ℤ) (a : ℂ) : ℂ :=
∑' n : ℕ, generalizedHyperharmonic (n + k) k r / (((n + k : ℕ) : ℂ) + a) ^ p
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:74
noncomputable def generalizedHyperharmonic (n k : ℕ) (r : ℂ) : ℂ :=
if k ≤ n then
(∑ s ∈ (Finset.range n).powersetCard (n - k), ∏ j ∈ s, (r + j)) / n.factorial
else 0
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:34
Sources
- Paul Thomas Young, Parametric Euler Sums of Generalized Hyperharmonic Numbers, INTEGERS 26 (2026), Article A81, Corollary 1 display.
Uses 3
DefinitionOdd multiple harmonic numbers#✓fully proved
For natural numbers \(k\) and \(n\), the depth-\(k\) odd harmonic number is the real number
In particular \(O_n(0) = 1\) (the empty product) and \(O_n(k) = 0\) for \(k > n\) (the empty sum).
noncomputable def oddMultipleHarmonic (k n : ℕ) : ℝ :=
∑ s ∈ (Finset.Icc 1 n).powersetCard k, ∏ i ∈ s, ((i : ℝ) - 1 / 2)⁻¹
Sources
- Paul Thomas Young, Parametric Euler Sums of Generalized Hyperharmonic Numbers, INTEGERS 26 (2026), Article A81, (1.8) and (2.11).
Used by 18
TheoremOdd multiple harmonic numbers of depth zero#✓fully proved
For every natural number \(n\), the depth-zero odd harmonic number \(O_n(0)\) equals \(1\).
@[simp]
theorem oddMultipleHarmonic_zero (n : ℕ) : oddMultipleHarmonic 0 n = 1
noncomputable def oddMultipleHarmonic (k n : ℕ) : ℝ :=
∑ s ∈ (Finset.Icc 1 n).powersetCard k, ∏ i ∈ s, ((i : ℝ) - 1 / 2)⁻¹
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:98
Sources
- Paul Thomas Young, Parametric Euler Sums of Generalized Hyperharmonic Numbers, INTEGERS 26 (2026), Article A81, (2.11).
Uses 1
TheoremVanishing of odd multiple harmonic numbers of large depth#✓fully proved
For natural numbers \(k\) and \(n\) with \(k > n\), the depth-\(k\) odd harmonic number \(O_n(k)\) equals \(0\).
theorem oddMultipleHarmonic_of_lt {k n : ℕ} (h : n < k) : oddMultipleHarmonic k n = 0
noncomputable def oddMultipleHarmonic (k n : ℕ) : ℝ :=
∑ s ∈ (Finset.Icc 1 n).powersetCard k, ∏ i ∈ s, ((i : ℝ) - 1 / 2)⁻¹
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:98
Sources
- Paul Thomas Young, Parametric Euler Sums of Generalized Hyperharmonic Numbers, INTEGERS 26 (2026), Article A81, (2.11).
Uses 1
TheoremThe first odd harmonic number#✓fully proved
The depth-one odd harmonic number \(O_1(1)\) equals \(2\).
theorem oddMultipleHarmonic_one_one : oddMultipleHarmonic 1 1 = 2
noncomputable def oddMultipleHarmonic (k n : ℕ) : ℝ :=
∑ s ∈ (Finset.Icc 1 n).powersetCard k, ∏ i ∈ s, ((i : ℝ) - 1 / 2)⁻¹
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:98
Sources
- Paul Thomas Young, Parametric Euler Sums of Generalized Hyperharmonic Numbers, INTEGERS 26 (2026), Article A81, (2.11).
Uses 1
DefinitionThe algebra generated by \(\log 2\), \(\pi^2\) and odd zeta values#✓fully proved
This is the \(\mathbb{Q}\)-subalgebra of \(\mathbb{C}\) generated by \(\log 2\), \(\pi^2\) and the zeta values \(\zeta(3), \zeta(5), \zeta(7), \dots\) (that is, \(\zeta(2j+1)\) for all integers \(j \ge 1\)). It is written \(\mathbb{Q}[\log 2, \pi^2, \zeta(3), \zeta(5), \dots]\).
noncomputable def eulerSumPeriodAlgebra : Subalgebra ℚ ℂ :=
Algebra.adjoin ℚ
({(Real.log 2 : ℂ), (π : ℂ) ^ 2} ∪ Set.range fun j : ℕ => riemannZeta (2 * j + 3))
Sources
- Paul Thomas Young, Parametric Euler Sums of Generalized Hyperharmonic Numbers, INTEGERS 26 (2026), Article A81, Corollary 2.
Used by 9
TheoremGenerators of the Euler sum period algebra#✓fully proved
The numbers \(\log 2\) and \(\pi^2\) and the odd zeta values \(\zeta(2j + 3)\) for all integers \(j \ge 0\) lie in the \(\mathbb{Q}\)-subalgebra \(\mathbb{Q}[\log 2, \pi^2, \zeta(2j+1)_{j \ge 1}]\) of \(\mathbb{C}\) that they generate.
theorem generators_mem_eulerSumPeriodAlgebra :
(Real.log 2 : ℂ) ∈ eulerSumPeriodAlgebra ∧ (π : ℂ) ^ 2 ∈ eulerSumPeriodAlgebra ∧
∀ j : ℕ, riemannZeta (2 * j + 3) ∈ eulerSumPeriodAlgebra
noncomputable def eulerSumPeriodAlgebra : Subalgebra ℚ ℂ :=
Algebra.adjoin ℚ
({(Real.log 2 : ℂ), (π : ℂ) ^ 2} ∪ Set.range fun j : ℕ => riemannZeta (2 * j + 3))
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:129
Sources
- Paul Thomas Young, Parametric Euler Sums of Generalized Hyperharmonic Numbers, INTEGERS 26 (2026), Article A81, Corollary 2.
Uses 1
DefinitionThe algebra generated by \(\log 2\), \(\pi\), odd zeta values and even beta values#✓fully proved
For integers \(s \ge 2\) let \(\beta(s) = \sum_{n \ge 0} (-1)^n / (2n+1)^s\) be the value of the Dirichlet beta function. This is the \(\mathbb{Q}\)-subalgebra of \(\mathbb{C}\) generated by \(\log 2\), \(\pi\), the zeta values \(\zeta(2j+1)\) for all integers \(j \ge 1\) and the beta values \(\beta(2j)\) for all integers \(j \ge 1\). It is written \(\mathbb{Q}[\log 2, \pi, \zeta(2j+1)_{j \ge 1}, \beta(2j)_{j \ge 1}]\).
noncomputable def eulerSumBetaPeriodAlgebra : Subalgebra ℚ ℂ :=
Algebra.adjoin ℚ
({(Real.log 2 : ℂ), (π : ℂ)} ∪ Set.range (fun j : ℕ => riemannZeta (2 * j + 3)) ∪
Set.range fun j : ℕ =>
((∑' n : ℕ, (-1 : ℝ) ^ n / (2 * n + 1 : ℝ) ^ (2 * j + 2) : ℝ) : ℂ))
Sources
- Paul Thomas Young, Parametric Euler Sums of Generalized Hyperharmonic Numbers, INTEGERS 26 (2026), Article A81, Corollary 6.
Used by 8
TheoremGenerators of the Euler sum beta period algebra#✓fully proved
The numbers \(\log 2\) and \(\pi\), the zeta values \(\zeta(2j + 3)\) for all integers \(j \ge 0\), and the Dirichlet beta values \(\beta(2j + 2) = \sum_{n \ge 0} (-1)^n / (2n + 1)^{2j+2}\) for all integers \(j \ge 0\) lie in the \(\mathbb{Q}\)-subalgebra \(\mathbb{Q}[\log 2, \pi, \zeta(2j+1)_{j \ge 1}, \beta(2j)_{j \ge 1}]\) of \(\mathbb{C}\) that they generate.
theorem generators_mem_eulerSumBetaPeriodAlgebra :
(Real.log 2 : ℂ) ∈ eulerSumBetaPeriodAlgebra ∧ (π : ℂ) ∈ eulerSumBetaPeriodAlgebra ∧
(∀ j : ℕ, riemannZeta (2 * j + 3) ∈ eulerSumBetaPeriodAlgebra) ∧
∀ j : ℕ, ((∑' n : ℕ, (-1 : ℝ) ^ n / (2 * n + 1 : ℝ) ^ (2 * j + 2) : ℝ) : ℂ) ∈
eulerSumBetaPeriodAlgebra
noncomputable def eulerSumBetaPeriodAlgebra : Subalgebra ℚ ℂ :=
Algebra.adjoin ℚ
({(Real.log 2 : ℂ), (π : ℂ)} ∪ Set.range (fun j : ℕ => riemannZeta (2 * j + 3)) ∪
Set.range fun j : ℕ =>
((∑' n : ℕ, (-1 : ℝ) ^ n / (2 * n + 1 : ℝ) ^ (2 * j + 2) : ℝ) : ℂ))
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:148
Sources
- Paul Thomas Young, Parametric Euler Sums of Generalized Hyperharmonic Numbers, INTEGERS 26 (2026), Article A81, Corollary 6.
Uses 1
TheoremBeta-function evaluation of hyperharmonic Euler sums#●ready to prove
Let \(k, p \ge 0\) be integers and let \(r, a\) be complex numbers with \(\operatorname{Re}(r) < p + 1\) and \(\operatorname{Re}(a) > -k\), such that \(r\) is not a positive integer and \(a\) is not a nonpositive integer. Write \(H_n^{[r]}(k)\) for the generalized hyperharmonic number and let \(S(k, r; p + 1; a) = \sum_{n \ge k} H_n^{[r]}(k) / (n + a)^{p+1}\). Then
where \(B(a, 1 - r) = \Gamma(a) \Gamma(1 - r) \cdot (1/\Gamma)(a + 1 - r)\), with \(1/\Gamma\) the entire reciprocal of the Gamma function (zero at the poles of \(\Gamma\)), and \(D_a^p D_r^k\) means the \(k\)-th complex derivative in \(r\) followed by the \(p\)-th complex derivative in \(a\).
The source also states the identity at positive integers \(r\), after removing the singularity of \(\Gamma(1 - r)\), and at nonpositive integers \(a\), and extends it by analytic continuation; those cases are not included here.
public theorem_wanted hyperharmonicEulerSum_eq_iteratedDeriv_beta (k p : ℕ) (r a : ℂ)
(hr : r.re < p + 1) (ha : -(k : ℝ) < a.re) (hr' : ∀ m : ℕ, r ≠ m + 1)
(ha' : ∀ m : ℕ, a ≠ -m) :
(p.factorial * k.factorial : ℂ) * hyperharmonicEulerSum k r (p + 1) a =
(-1) ^ p * iteratedDeriv p (fun b => iteratedDeriv k
(fun s => Complex.Gamma b * Complex.Gamma (1 - s) * (Complex.Gamma (b + 1 - s))⁻¹) r) a
noncomputable def hyperharmonicEulerSum (k : ℕ) (r : ℂ) (p : ℤ) (a : ℂ) : ℂ :=
∑' n : ℕ, generalizedHyperharmonic (n + k) k r / (((n + k : ℕ) : ℂ) + a) ^ p
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:74
Sources
- Paul Thomas Young, Parametric Euler Sums of Generalized Hyperharmonic Numbers, INTEGERS 26 (2026), Article A81, Theorem 1.
Uses 1
TheoremPolygamma structure of hyperharmonic Euler sums#●ready to prove
Let \(k \ge 0\), \(r\) and \(p\) be integers with \(1 \le r \le p\), and let \(a \in \mathbb{C} \setminus \mathbb{Z}\). Write \(H_n^{[r]}(k)\) for the generalized hyperharmonic number. Then the series \(S(k, r; p + 1; a) = \sum_{n \ge k} H_n^{[r]}(k) / (n + a)^{p+1}\) converges absolutely, and its sum lies in the algebra generated over the field \(\mathbb{Q}(a)\) by \(\psi(a) + \gamma\), the derivatives \(\psi^{(j)}(a)\) for \(j \ge 1\) and the zeta values \(\zeta(j)\) for \(j \ge 2\), where \(\psi\) is the digamma function and \(\gamma\) is the Euler–Mascheroni constant.
The source also extends the result by continuity to integers \(a > -k\); that case is not included here.
public theorem_wanted hyperharmonicEulerSum_mem_adjoin_digamma (k p r : ℕ) (hr : 1 ≤ r)
(hrp : r ≤ p) (a : ℂ) (ha : ∀ m : ℤ, a ≠ m) :
Summable (fun n : ℕ =>
generalizedHyperharmonic (n + k) k r / (((n + k : ℕ) : ℂ) + a) ^ ((p : ℤ) + 1)) ∧
hyperharmonicEulerSum k r (p + 1) a ∈
Algebra.adjoin (IntermediateField.adjoin ℚ {a})
({Complex.digamma a + Real.eulerMascheroniConstant} ∪
Set.range (fun j : ℕ => iteratedDeriv (j + 1) Complex.digamma a) ∪
Set.range fun j : ℕ => riemannZeta (j + 2))
noncomputable def generalizedHyperharmonic (n k : ℕ) (r : ℂ) : ℂ :=
if k ≤ n then
(∑ s ∈ (Finset.range n).powersetCard (n - k), ∏ j ∈ s, (r + j)) / n.factorial
else 0
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:34
noncomputable def hyperharmonicEulerSum (k : ℕ) (r : ℂ) (p : ℤ) (a : ℂ) : ℂ :=
∑' n : ℕ, generalizedHyperharmonic (n + k) k r / (((n + k : ℕ) : ℂ) + a) ^ p
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:74
Sources
- Paul Thomas Young, Parametric Euler Sums of Generalized Hyperharmonic Numbers, INTEGERS 26 (2026), Article A81, Theorem 2.
Uses 2
TheoremHyperharmonic Euler sums with rational parameters#●ready to prove
Let \(a\) and \(r\) be rational numbers that are not integers, with \(a - r \in \mathbb{Z}\), let \(N \ge 1\) be an integer with \(N a \in \mathbb{Z}\), and let \(k \ge 0\) and \(p > r\) be integers. Put \(M = \mathrm{lcm}(2N, 4)\) and \(\omega = e^{2 \pi i / M}\), a primitive \(M\)-th root of unity. Write \(H_n^{[r]}(k)\) for the generalized hyperharmonic numbers. Then the generalized hyperharmonic Euler sum \(S(k, r; p; a) = \sum_{n \ge k} H_n^{[r]}(k) / (n + a)^p\) converges, and \(S(k, r; p; a) = \pi x\) for some \(x\) in the algebra generated over the field \(\mathbb{Q}(\omega + \omega^{-1})\) by \(\psi(a) + \gamma\), the derivatives \(\psi^{(j)}(a)\) for \(j \ge 1\), \(\pi^2\) and the odd zeta values \(\zeta(3), \zeta(5), \dots\), where \(\psi\) is the digamma function and \(\gamma\) is the Euler–Mascheroni constant.
The field \(\mathbb{Q}(\omega + \omega^{-1})\) does not depend on which primitive \(M\)-th root of unity \(\omega\) is chosen.
public theorem_wanted hyperharmonicEulerSum_rat_mem_pi_mul_adjoin (a r : ℚ)
(ha : ∀ m : ℤ, a ≠ m) (hr : ∀ m : ℤ, r ≠ m) (N : ℕ) (hN : 0 < N)
(hNa : ∃ m : ℤ, N * a = m) (har : ∃ m : ℤ, a - r = m) (k : ℕ) (p : ℤ) (hp : r < p) :
let ζ : ℂ := Complex.exp (2 * π * Complex.I / (Nat.lcm (2 * N) 4 : ℂ))
Summable (fun n : ℕ =>
generalizedHyperharmonic (n + k) k r / (((n + k : ℕ) : ℂ) + a) ^ p) ∧
∃ x ∈ Algebra.adjoin (IntermediateField.adjoin ℚ {ζ + ζ⁻¹})
({Complex.digamma a + Real.eulerMascheroniConstant} ∪
Set.range (fun j : ℕ => iteratedDeriv (j + 1) Complex.digamma a) ∪
{(π : ℂ) ^ 2} ∪ Set.range fun j : ℕ => riemannZeta (2 * j + 3)),
hyperharmonicEulerSum k r p a = π * x
noncomputable def generalizedHyperharmonic (n k : ℕ) (r : ℂ) : ℂ :=
if k ≤ n then
(∑ s ∈ (Finset.range n).powersetCard (n - k), ∏ j ∈ s, (r + j)) / n.factorial
else 0
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:34
noncomputable def hyperharmonicEulerSum (k : ℕ) (r : ℂ) (p : ℤ) (a : ℂ) : ℂ :=
∑' n : ℕ, generalizedHyperharmonic (n + k) k r / (((n + k : ℕ) : ℂ) + a) ^ p
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:74
Sources
- Paul Thomas Young, Parametric Euler Sums of Generalized Hyperharmonic Numbers, INTEGERS 26 (2026), Article A81, Theorem 3.
Uses 2
TheoremDuality for hyperharmonic Euler sums#●ready to prove
Let \(k, p \ge 0\) be integers and let \(r, t\) be complex numbers with \(\operatorname{Re}(r) < p + 1\) and \(\operatorname{Re}(t) < k + 1\). Write \(H_n^{[r]}(k)\) for the generalized hyperharmonic number. Then the series \(S(k, r; p + 1; 1 - t) = \sum_{n \ge k} H_n^{[r]}(k) / (n + 1 - t)^{p+1}\) and \(S(p, t; k + 1; 1 - r) = \sum_{n \ge p} H_n^{[t]}(p) / (n + 1 - r)^{k+1}\) both converge absolutely, and \(S(k, r; p + 1; 1 - t) = S(p, t; k + 1; 1 - r)\).
public theorem_wanted hyperharmonicEulerSum_duality (k p : ℕ) (r t : ℂ) (hr : r.re < p + 1)
(ht : t.re < k + 1) :
Summable (fun n : ℕ =>
generalizedHyperharmonic (n + k) k r / (((n + k : ℕ) : ℂ) + (1 - t)) ^ ((p : ℤ) + 1)) ∧
Summable (fun n : ℕ =>
generalizedHyperharmonic (n + p) p t / (((n + p : ℕ) : ℂ) + (1 - r)) ^ ((k : ℤ) + 1)) ∧
hyperharmonicEulerSum k r (p + 1) (1 - t) = hyperharmonicEulerSum p t (k + 1) (1 - r)
noncomputable def generalizedHyperharmonic (n k : ℕ) (r : ℂ) : ℂ :=
if k ≤ n then
(∑ s ∈ (Finset.range n).powersetCard (n - k), ∏ j ∈ s, (r + j)) / n.factorial
else 0
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:34
noncomputable def hyperharmonicEulerSum (k : ℕ) (r : ℂ) (p : ℤ) (a : ℂ) : ℂ :=
∑' n : ℕ, generalizedHyperharmonic (n + k) k r / (((n + k : ℕ) : ℂ) + a) ^ p
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:74
Sources
- Paul Thomas Young, Parametric Euler Sums of Generalized Hyperharmonic Numbers, INTEGERS 26 (2026), Article A81, Corollary 1.
Uses 2
TheoremHyperharmonic Euler sums with integer parameter and half-integer shift#●ready to prove
Let \(k \ge 0\), \(r\) and \(p\) be integers with \(1 \le r < p\), and let \(a \in \frac{1}{2} + \mathbb{Z}\). Write \(H_n^{[r]}(k)\) for the generalized hyperharmonic number. Then the series \(S(k, r; p; a) = \sum_{n \ge k} H_n^{[r]}(k) / (n + a)^p\) converges absolutely, and its sum lies in the \(\mathbb{Q}\)-subalgebra \(\mathbb{Q}[\log 2, \pi^2, \zeta(2j+1)_{j \ge 1}]\) of \(\mathbb{C}\) generated by \(\log 2\), \(\pi^2\) and the odd zeta values.
The source prints \(1 \le r \le p\), but at \(r = p\) the series can diverge (for \(k = 0\), \(r = p = 1\) and \(a = 1/2\) it is \(\sum_{n \ge 0} 1/(n + 1/2)\)), and the source's Theorem 2 has exponent \(p + 1\) for \(r \le p\); so \(r < p\) is assumed.
public theorem_wanted hyperharmonicEulerSum_half_mem_eulerSumPeriodAlgebra (k p r : ℕ)
(hr : 1 ≤ r) (hrp : r < p) (c : ℤ) :
Summable (fun n : ℕ =>
generalizedHyperharmonic (n + k) k r / (((n + k : ℕ) : ℂ) + (c + 1 / 2)) ^ (p : ℤ)) ∧
hyperharmonicEulerSum k r p (c + 1 / 2) ∈ eulerSumPeriodAlgebra
noncomputable def generalizedHyperharmonic (n k : ℕ) (r : ℂ) : ℂ :=
if k ≤ n then
(∑ s ∈ (Finset.range n).powersetCard (n - k), ∏ j ∈ s, (r + j)) / n.factorial
else 0
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:34
noncomputable def hyperharmonicEulerSum (k : ℕ) (r : ℂ) (p : ℤ) (a : ℂ) : ℂ :=
∑' n : ℕ, generalizedHyperharmonic (n + k) k r / (((n + k : ℕ) : ℂ) + a) ^ p
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:74
noncomputable def eulerSumPeriodAlgebra : Subalgebra ℚ ℂ :=
Algebra.adjoin ℚ
({(Real.log 2 : ℂ), (π : ℂ) ^ 2} ∪ Set.range fun j : ℕ => riemannZeta (2 * j + 3))
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:129
Sources
- Paul Thomas Young, Parametric Euler Sums of Generalized Hyperharmonic Numbers, INTEGERS 26 (2026), Article A81, Corollary 2, first part.
TheoremOdd harmonic central binomial sums at integer shifts#●ready to prove
Let \(k \ge 0\) and \(p > 0\) be integers, let \(a\) be an integer with \(a > -k\), and let \(O_n(k)\) denote the depth-\(k\) odd harmonic number. Then the series of real numbers
converges, and its sum lies in the algebra \(\mathbb{Q}[\log 2, \pi^2, \zeta(3), \zeta(5), \dots]\) viewed inside \(\mathbb{C}\). The terms with \(n < k\) are \(0\), since \(O_n(k) = 0\) for \(n < k\), so this is also the sum over \(n \ge k\).
public theorem_wanted centralBinom_oddMultipleHarmonic_sum_mem_eulerSumPeriodAlgebra
(k p : ℕ) (hp : 0 < p) (a : ℤ) (ha : -(k : ℤ) < a) :
Summable (fun n : ℕ =>
((2 * n).choose n : ℝ) * oddMultipleHarmonic k n / (4 ^ n * ((n : ℝ) + a) ^ p)) ∧
((∑' n : ℕ, ((2 * n).choose n : ℝ) * oddMultipleHarmonic k n /
(4 ^ n * ((n : ℝ) + a) ^ p) : ℝ) : ℂ) ∈ eulerSumPeriodAlgebra
noncomputable def oddMultipleHarmonic (k n : ℕ) : ℝ :=
∑ s ∈ (Finset.Icc 1 n).powersetCard k, ∏ i ∈ s, ((i : ℝ) - 1 / 2)⁻¹
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:98
noncomputable def eulerSumPeriodAlgebra : Subalgebra ℚ ℂ :=
Algebra.adjoin ℚ
({(Real.log 2 : ℂ), (π : ℂ) ^ 2} ∪ Set.range fun j : ℕ => riemannZeta (2 * j + 3))
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:129
Sources
- Paul Thomas Young, Parametric Euler Sums of Generalized Hyperharmonic Numbers, INTEGERS 26 (2026), Article A81, Corollary 2, second part.
Uses 2
TheoremCentral binomial odd harmonic sum with denominator \(4^n (n + 1)^2\)#●ready to prove
For every integer \(k \ge 0\), with \(O_n(k)\) the depth-\(k\) odd harmonic number,
The terms with \(n < k\) vanish, since \(O_n(k) = 0\) for \(n < k\).
public theorem_wanted centralBinom_oddMultipleHarmonic_div_succ_sq_tsum (k : ℕ) :
((∑' n : ℕ, ((2 * n).choose n : ℝ) * oddMultipleHarmonic k n /
(4 ^ n * ((n : ℝ) + 1) ^ 2) : ℝ) : ℂ) =
2 ^ (k + 2) * (k + 1 - Real.log 2) -
∑ j ∈ Finset.Icc 1 k, (2 ^ (k + 2) - 2 ^ j) * riemannZeta ((k + 2 - j : ℕ) : ℂ)
noncomputable def oddMultipleHarmonic (k n : ℕ) : ℝ :=
∑ s ∈ (Finset.Icc 1 n).powersetCard k, ∏ i ∈ s, ((i : ℝ) - 1 / 2)⁻¹
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:98
Sources
- Paul Thomas Young, Parametric Euler Sums of Generalized Hyperharmonic Numbers, INTEGERS 26 (2026), Article A81, Corollary 3.
Uses 1
TheoremOdd harmonic central binomial sums with weight \(\binom{n}{m}\) at integer shifts#●ready to prove
Let \(k, m \ge 0\) and \(p > m\) be integers, let \(a\) be an integer with \(a > -k\), and let \(O_n(k)\) denote the depth-\(k\) odd harmonic number. Then the series of real numbers
converges, and its sum lies in the algebra \(\mathbb{Q}[\log 2, \pi^2, \zeta(3), \zeta(5), \dots]\) viewed inside \(\mathbb{C}\). The terms with \(n < k\) are \(0\), since \(O_n(k) = 0\) for \(n < k\), so this is also the sum over \(n \ge k\).
public theorem_wanted choose_centralBinom_oddMultipleHarmonic_sum_mem_eulerSumPeriodAlgebra
(k p m : ℕ) (hmp : m < p) (a : ℤ) (ha : -(k : ℤ) < a) :
Summable (fun n : ℕ => (n.choose m : ℝ) * ((2 * n).choose n : ℝ) *
oddMultipleHarmonic k n / (4 ^ n * ((n : ℝ) + a) ^ p)) ∧
((∑' n : ℕ, (n.choose m : ℝ) * ((2 * n).choose n : ℝ) * oddMultipleHarmonic k n /
(4 ^ n * ((n : ℝ) + a) ^ p) : ℝ) : ℂ) ∈ eulerSumPeriodAlgebra
noncomputable def oddMultipleHarmonic (k n : ℕ) : ℝ :=
∑ s ∈ (Finset.Icc 1 n).powersetCard k, ∏ i ∈ s, ((i : ℝ) - 1 / 2)⁻¹
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:98
noncomputable def eulerSumPeriodAlgebra : Subalgebra ℚ ℂ :=
Algebra.adjoin ℚ
({(Real.log 2 : ℂ), (π : ℂ) ^ 2} ∪ Set.range fun j : ℕ => riemannZeta (2 * j + 3))
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:129
Sources
- Paul Thomas Young, Parametric Euler Sums of Generalized Hyperharmonic Numbers, INTEGERS 26 (2026), Article A81, Corollary 4, first part.
Uses 2
TheoremCentral binomial odd harmonic sums over \(\binom{n+m}{m}\) lie in the period algebra#●ready to prove
Let \(O_n(k)\) denote the depth-\(k\) odd harmonic number and let \(A = \mathbb{Q}[\log 2, \pi^2, \zeta(3), \zeta(5), \ldots] \subseteq \mathbb{C}\) be the Euler sum period algebra. For integers \(k \ge 0\), \(p \ge 1\), \(m \ge 1\) and \(a > -k\), the series
converges and its sum lies in \(A\).
The terms with \(n < k\) are \(0\) because \(O_n(k) = 0\) there; this includes the term with \(n + a = 0\) when \(a \le 0\).
public theorem_wanted centralBinom_oddMultipleHarmonic_div_choose_sum_mem_eulerSumPeriodAlgebra
(k p m : ℕ) (hp : 0 < p) (hm : 0 < m) (a : ℤ) (ha : -(k : ℤ) < a) :
Summable (fun n : ℕ => ((2 * n).choose n : ℝ) * oddMultipleHarmonic k n /
(4 ^ n * ((n + m).choose m : ℝ) * ((n : ℝ) + a) ^ p)) ∧
((∑' n : ℕ, ((2 * n).choose n : ℝ) * oddMultipleHarmonic k n /
(4 ^ n * ((n + m).choose m : ℝ) * ((n : ℝ) + a) ^ p) : ℝ) : ℂ) ∈
eulerSumPeriodAlgebra
noncomputable def oddMultipleHarmonic (k n : ℕ) : ℝ :=
∑ s ∈ (Finset.Icc 1 n).powersetCard k, ∏ i ∈ s, ((i : ℝ) - 1 / 2)⁻¹
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:98
noncomputable def eulerSumPeriodAlgebra : Subalgebra ℚ ℂ :=
Algebra.adjoin ℚ
({(Real.log 2 : ℂ), (π : ℂ) ^ 2} ∪ Set.range fun j : ℕ => riemannZeta (2 * j + 3))
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:129
Sources
- Paul Thomas Young, Parametric Euler Sums of Generalized Hyperharmonic Numbers, INTEGERS 26 (2026), Article A81, Corollary 4, second part.
Uses 2
TheoremCatalan–odd harmonic sum with denominator \(4^n n\)#●ready to prove
Let \(k \ge 1\) be an integer, let \(C_n\) denote the \(n\)-th Catalan number and \(O_n(k)\) the depth-\(k\) odd harmonic number. Then
Since \(O_n(k) = 0\) for \(n < k\), the sum can equally be taken over \(n \ge k\).
public theorem_wanted catalan_oddMultipleHarmonic_div_tsum (k : ℕ) (hk : 0 < k) :
((∑' n : ℕ, (catalan n : ℝ) * oddMultipleHarmonic k n / (4 ^ n * n) : ℝ) : ℂ) =
(2 ^ (k + 1) - 1) * riemannZeta (k + 1) - 2 ^ (k + 1)
noncomputable def oddMultipleHarmonic (k n : ℕ) : ℝ :=
∑ s ∈ (Finset.Icc 1 n).powersetCard k, ∏ i ∈ s, ((i : ℝ) - 1 / 2)⁻¹
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:98
Sources
- the denominator \(4^n n\) is read from the record's two-row layout (\(C_n O_n(k)\) over \(4^n n\)).
- Paul Thomas Young, Parametric Euler Sums of Generalized Hyperharmonic Numbers, INTEGERS 26 (2026), Article A81, Corollary 5.
Uses 1
TheoremQuarter-shift hyperharmonic Euler sums#●ready to prove
Let \(k \ge 0\) and \(0 \le r < p\) be integers, and let \(a\) be a complex number with \(a \in \tfrac14 + \mathbb{Z}\) or \(a \in \tfrac34 + \mathbb{Z}\). Write \(H_n^{[r]}(k)\) for the generalized hyperharmonic numbers. Then the generalized hyperharmonic Euler sum \(S(k, r; p; a) = \sum_{n \ge k} H_n^{[r]}(k) / (n + a)^p\) converges, and its value lies in the \(\mathbb{Q}\)-algebra \(\mathbb{Q}[\log 2, \pi, \zeta(3), \zeta(5), \dots, \beta(2), \beta(4), \dots]\) of complex numbers, where \(\beta(s) = \sum_{n \ge 0} (-1)^n / (2n+1)^s\) is the Dirichlet beta function.
The source prints \(0 \le r \le p\); at \(r = p \ge 1\) the series diverges (for \(k = 0\) and \(r = p = 1\) it is \(\sum_{n \ge 0} 1 / (n + a)\)), so \(r < p\) is used.
public theorem_wanted hyperharmonicEulerSum_quarter_mem_eulerSumBetaPeriodAlgebra
(k p r : ℕ) (hrp : r < p) (a : ℂ) (ha : ∃ c : ℤ, a = c + 1 / 4 ∨ a = c + 3 / 4) :
Summable (fun n : ℕ =>
generalizedHyperharmonic (n + k) k r / (((n + k : ℕ) : ℂ) + a) ^ (p : ℤ)) ∧
hyperharmonicEulerSum k r p a ∈ eulerSumBetaPeriodAlgebra
noncomputable def generalizedHyperharmonic (n k : ℕ) (r : ℂ) : ℂ :=
if k ≤ n then
(∑ s ∈ (Finset.range n).powersetCard (n - k), ∏ j ∈ s, (r + j)) / n.factorial
else 0
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:34
noncomputable def hyperharmonicEulerSum (k : ℕ) (r : ℂ) (p : ℤ) (a : ℂ) : ℂ :=
∑' n : ℕ, generalizedHyperharmonic (n + k) k r / (((n + k : ℕ) : ℂ) + a) ^ p
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:74
noncomputable def eulerSumBetaPeriodAlgebra : Subalgebra ℚ ℂ :=
Algebra.adjoin ℚ
({(Real.log 2 : ℂ), (π : ℂ)} ∪ Set.range (fun j : ℕ => riemannZeta (2 * j + 3)) ∪
Set.range fun j : ℕ =>
((∑' n : ℕ, (-1 : ℝ) ^ n / (2 * n + 1 : ℝ) ^ (2 * j + 2) : ℝ) : ℂ))
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:148
Sources
- Paul Thomas Young, Parametric Euler Sums of Generalized Hyperharmonic Numbers, INTEGERS 26 (2026), Article A81, Corollary 6.
TheoremHyperharmonic Euler sums with half-integer parameter and shift#●ready to prove
Let \(k \ge 0\) be an integer, let \(r, a \in \frac{1}{2} + \mathbb{Z}\), and let \(p\) be an integer with \(p > r\). Write \(H_n^{[r]}(k)\) for the generalized hyperharmonic number. Then the series \(S(k, r; p; a) = \sum_{n \ge k} H_n^{[r]}(k) / (n + a)^p\) converges absolutely, and its sum equals \(\pi x\) for some \(x\) in the \(\mathbb{Q}\)-subalgebra \(\mathbb{Q}[\log 2, \pi^2, \zeta(2j+1)_{j \ge 1}]\) of \(\mathbb{C}\) generated by \(\log 2\), \(\pi^2\) and the odd zeta values.
public theorem_wanted hyperharmonicEulerSum_half_half_mem_pi_mul (k : ℕ) (c d p : ℤ)
(hp : c < p) :
Summable (fun n : ℕ => generalizedHyperharmonic (n + k) k (c + 1 / 2) /
(((n + k : ℕ) : ℂ) + (d + 1 / 2)) ^ p) ∧
∃ x ∈ eulerSumPeriodAlgebra,
hyperharmonicEulerSum k (c + 1 / 2) p (d + 1 / 2) = π * x
noncomputable def generalizedHyperharmonic (n k : ℕ) (r : ℂ) : ℂ :=
if k ≤ n then
(∑ s ∈ (Finset.range n).powersetCard (n - k), ∏ j ∈ s, (r + j)) / n.factorial
else 0
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:34
noncomputable def eulerSumPeriodAlgebra : Subalgebra ℚ ℂ :=
Algebra.adjoin ℚ
({(Real.log 2 : ℂ), (π : ℂ) ^ 2} ∪ Set.range fun j : ℕ => riemannZeta (2 * j + 3))
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:129
noncomputable def hyperharmonicEulerSum (k : ℕ) (r : ℂ) (p : ℤ) (a : ℂ) : ℂ :=
∑' n : ℕ, generalizedHyperharmonic (n + k) k r / (((n + k : ℕ) : ℂ) + a) ^ p
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:74
Sources
- Paul Thomas Young, Parametric Euler Sums of Generalized Hyperharmonic Numbers, INTEGERS 26 (2026), Article A81, Corollary 7, first part.
TheoremOdd harmonic central binomial sums with weight \(\binom{n}{m}\) at half-integer shifts#●ready to prove
Let \(k, m \ge 0\) and \(p > m\) be integers, let \(a = d + 1/2\) with \(d\) an integer, and let \(O_n(k)\) denote the depth-\(k\) odd harmonic number. Then the series of real numbers
converges, and its sum equals \(\pi x\) for some \(x\) in the algebra \(\mathbb{Q}[\log 2, \pi^2, \zeta(3), \zeta(5), \dots] \subseteq \mathbb{C}\). The terms with \(n < k\) are \(0\), since \(O_n(k) = 0\) for \(n < k\), so this is also the sum over \(n \ge k\).
public theorem_wanted choose_centralBinom_oddMultipleHarmonic_half_sum_mem_pi_mul
(k p m : ℕ) (hmp : m < p) (d : ℤ) :
Summable (fun n : ℕ => (n.choose m : ℝ) * ((2 * n).choose n : ℝ) *
oddMultipleHarmonic k n / (4 ^ n * ((n : ℝ) + (d + 1 / 2)) ^ p)) ∧
∃ x ∈ eulerSumPeriodAlgebra,
((∑' n : ℕ, (n.choose m : ℝ) * ((2 * n).choose n : ℝ) * oddMultipleHarmonic k n /
(4 ^ n * ((n : ℝ) + (d + 1 / 2)) ^ p) : ℝ) : ℂ) = π * x
noncomputable def oddMultipleHarmonic (k n : ℕ) : ℝ :=
∑ s ∈ (Finset.Icc 1 n).powersetCard k, ∏ i ∈ s, ((i : ℝ) - 1 / 2)⁻¹
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:98
noncomputable def eulerSumPeriodAlgebra : Subalgebra ℚ ℂ :=
Algebra.adjoin ℚ
({(Real.log 2 : ℂ), (π : ℂ) ^ 2} ∪ Set.range fun j : ℕ => riemannZeta (2 * j + 3))
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:129
Sources
- Paul Thomas Young, Parametric Euler Sums of Generalized Hyperharmonic Numbers, INTEGERS 26 (2026), Article A81, Corollary 7, second part.
Uses 2
TheoremCentral binomial odd harmonic sums over \(\binom{n+m}{m}\) at half-integer shifts#●ready to prove
Let \(O_n(k)\) denote the depth-\(k\) odd harmonic number and let \(A = \mathbb{Q}[\log 2, \pi^2, \zeta(3), \zeta(5), \ldots] \subseteq \mathbb{C}\) be the Euler sum period algebra. For integers \(k \ge 0\), \(p \ge 1\), \(m \ge 1\) and \(d\), put \(a = d + 1/2\). Then the series
converges, and its sum equals \(q + \pi x\) for some rational number \(q\) and some \(x \in A\).
The terms with \(n < k\) vanish, since \(O_n(k) = 0\) for \(n < k\).
public theorem_wanted centralBinom_oddMultipleHarmonic_div_choose_half_sum_mem
(k p m : ℕ) (hp : 0 < p) (hm : 0 < m) (d : ℤ) :
Summable (fun n : ℕ => ((2 * n).choose n : ℝ) * oddMultipleHarmonic k n /
(4 ^ n * ((n + m).choose m : ℝ) * ((n : ℝ) + (d + 1 / 2)) ^ p)) ∧
∃ q : ℚ, ∃ x ∈ eulerSumPeriodAlgebra,
((∑' n : ℕ, ((2 * n).choose n : ℝ) * oddMultipleHarmonic k n /
(4 ^ n * ((n + m).choose m : ℝ) * ((n : ℝ) + (d + 1 / 2)) ^ p) : ℝ) : ℂ) =
q + π * x
noncomputable def oddMultipleHarmonic (k n : ℕ) : ℝ :=
∑ s ∈ (Finset.Icc 1 n).powersetCard k, ∏ i ∈ s, ((i : ℝ) - 1 / 2)⁻¹
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:98
noncomputable def eulerSumPeriodAlgebra : Subalgebra ℚ ℂ :=
Algebra.adjoin ℚ
({(Real.log 2 : ℂ), (π : ℂ) ^ 2} ∪ Set.range fun j : ℕ => riemannZeta (2 * j + 3))
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:129
Sources
- Paul Thomas Young, Parametric Euler Sums of Generalized Hyperharmonic Numbers, INTEGERS 26 (2026), Article A81, Corollary 7, third part.
Uses 2
TheoremCatalan–odd harmonic sums at the shift \(1/2\)#●ready to prove
Let \(k \ge 0\) and \(p \ge 1\) be integers, let \(C_n\) denote the \(n\)-th Catalan number and \(O_n(k)\) the depth-\(k\) odd harmonic number, and let \(A = \mathbb{Q}[\log 2, \pi^2, \zeta(3), \zeta(5), \ldots] \subseteq \mathbb{C}\) be the Euler sum period algebra. Then the series
converges, and \(S - (-2)^p 2^{k+1} = \pi x\) for some \(x \in A\).
The terms with \(n < k\) vanish, since \(O_n(k) = 0\) for \(n < k\).
public theorem_wanted catalan_oddMultipleHarmonic_half_sum_sub_mem_pi_mul (k p : ℕ)
(hp : 0 < p) :
Summable (fun n : ℕ => (catalan n : ℝ) * oddMultipleHarmonic k n /
(4 ^ n * ((n : ℝ) + 1 / 2) ^ p)) ∧
∃ x ∈ eulerSumPeriodAlgebra,
((∑' n : ℕ, (catalan n : ℝ) * oddMultipleHarmonic k n /
(4 ^ n * ((n : ℝ) + 1 / 2) ^ p) : ℝ) : ℂ) - (-2) ^ p * 2 ^ (k + 1) = π * x
noncomputable def oddMultipleHarmonic (k n : ℕ) : ℝ :=
∑ s ∈ (Finset.Icc 1 n).powersetCard k, ∏ i ∈ s, ((i : ℝ) - 1 / 2)⁻¹
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:98
noncomputable def eulerSumPeriodAlgebra : Subalgebra ℚ ℂ :=
Algebra.adjoin ℚ
({(Real.log 2 : ℂ), (π : ℂ) ^ 2} ∪ Set.range fun j : ℕ => riemannZeta (2 * j + 3))
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:129
Sources
- Paul Thomas Young, Parametric Euler Sums of Generalized Hyperharmonic Numbers, INTEGERS 26 (2026), Article A81, Corollary 8.
Uses 2
TheoremCentral binomial sum over \(\binom{n+m}{m} (n + 1/2)^2\)#●ready to prove
Let \(m \ge 1\) be an integer and write \(O_m = O_m(1) = \sum_{i=1}^{m} 1/(i - 1/2)\) for the depth-one odd harmonic number. Then
public theorem_wanted centralBinom_div_choose_half_sq_tsum (m : ℕ) (hm : 0 < m) :
∑' n : ℕ, ((2 * n).choose n : ℝ) / (4 ^ n * ((n + m).choose m : ℝ) * ((n : ℝ) + 1 / 2) ^ 2) =
4 ^ m * π / ((2 * m).choose m : ℝ) * (2 * Real.log 2 - oddMultipleHarmonic 1 m) -
∑ q ∈ Finset.Icc 1 m,
(-4 : ℝ) ^ q * (m.choose q : ℝ) / (((2 * q).choose q : ℝ) * ((q : ℝ) - 1 / 2) ^ 2)
noncomputable def oddMultipleHarmonic (k n : ℕ) : ℝ :=
∑ s ∈ (Finset.Icc 1 n).powersetCard k, ∏ i ∈ s, ((i : ℝ) - 1 / 2)⁻¹
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:98
Sources
- Paul Thomas Young, Parametric Euler Sums of Generalized Hyperharmonic Numbers, INTEGERS 26 (2026), Article A81, Corollary 9, first identity.
Uses 1
TheoremOdd harmonic central binomial sum over \(\binom{n+m}{m} (n + 1/2)^2\)#●ready to prove
Let \(m \ge 1\) be an integer and write \(O_j = O_j(1) = \sum_{i=1}^{j} 1/(i - 1/2)\) for the depth-one odd harmonic number. Then
Here \(\pi^2 / 6 = \zeta(2)\).
public theorem_wanted centralBinom_oddHarmonic_sub_div_choose_half_sq_tsum (m : ℕ)
(hm : 0 < m) :
∑' n : ℕ, ((2 * n).choose n : ℝ) * (oddMultipleHarmonic 1 n - oddMultipleHarmonic 1 m) /
(4 ^ n * ((n + m).choose m : ℝ) * ((n : ℝ) + 1 / 2) ^ 2) =
4 ^ m * π / ((2 * m).choose m : ℝ) *
((oddMultipleHarmonic 1 m - 2 * Real.log 2) ^ 2 - π ^ 2 / 6) +
∑ q ∈ Finset.Icc 1 m,
(-4 : ℝ) ^ q * (m.choose q : ℝ) * (oddMultipleHarmonic 1 m - oddMultipleHarmonic 1 q) /
(((2 * q).choose q : ℝ) * ((q : ℝ) - 1 / 2) ^ 2)
noncomputable def oddMultipleHarmonic (k n : ℕ) : ℝ :=
∑ s ∈ (Finset.Icc 1 n).powersetCard k, ∏ i ∈ s, ((i : ℝ) - 1 / 2)⁻¹
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:98
Sources
- Paul Thomas Young, Parametric Euler Sums of Generalized Hyperharmonic Numbers, INTEGERS 26 (2026), Article A81, Corollary 9, second identity.
Uses 1
TheoremProducts of quarter-shift hyperharmonic Euler sums#●ready to prove
Let \(k_1, k_2 \ge 0\) be integers, let \(r_1, r_2 \in \tfrac12 + \mathbb{Z}\), \(a \in \tfrac14 + \mathbb{Z}\) and \(b \in \tfrac34 + \mathbb{Z}\), and let \(p_1, p_2\) be integers with \(p_1 > r_1\) and \(p_2 > r_2\). Write \(H_n^{[r]}(k)\) for the generalized hyperharmonic numbers and \(S(k, r; p; a) = \sum_{n \ge k} H_n^{[r]}(k) / (n + a)^p\) for the generalized hyperharmonic Euler sum. Then the series \(S(k_1, r_1; p_1; a)\) and \(S(k_2, r_2; p_2; b)\) converge, and \(S(k_1, r_1; p_1; a) \, S(k_2, r_2; p_2; b) = \pi x\) for some \(x\) in the \(\mathbb{Q}\)-algebra \(\mathbb{Q}[\log 2, \pi, \zeta(3), \zeta(5), \dots, \beta(2), \beta(4), \dots]\) of complex numbers, where \(\beta(s) = \sum_{n \ge 0} (-1)^n / (2n+1)^s\) is the Dirichlet beta function.
public theorem_wanted hyperharmonicEulerSum_quarter_mul_mem_pi_mul (k₁ k₂ : ℕ)
(c₁ c₂ d e p₁ p₂ : ℤ) (h₁ : c₁ < p₁) (h₂ : c₂ < p₂) :
Summable (fun n : ℕ => generalizedHyperharmonic (n + k₁) k₁ (c₁ + 1 / 2) /
(((n + k₁ : ℕ) : ℂ) + (d + 1 / 4)) ^ p₁) ∧
Summable (fun n : ℕ => generalizedHyperharmonic (n + k₂) k₂ (c₂ + 1 / 2) /
(((n + k₂ : ℕ) : ℂ) + (e + 3 / 4)) ^ p₂) ∧
∃ x ∈ eulerSumBetaPeriodAlgebra,
hyperharmonicEulerSum k₁ (c₁ + 1 / 2) p₁ (d + 1 / 4) *
hyperharmonicEulerSum k₂ (c₂ + 1 / 2) p₂ (e + 3 / 4) = π * x
noncomputable def generalizedHyperharmonic (n k : ℕ) (r : ℂ) : ℂ :=
if k ≤ n then
(∑ s ∈ (Finset.range n).powersetCard (n - k), ∏ j ∈ s, (r + j)) / n.factorial
else 0
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:34
noncomputable def eulerSumBetaPeriodAlgebra : Subalgebra ℚ ℂ :=
Algebra.adjoin ℚ
({(Real.log 2 : ℂ), (π : ℂ)} ∪ Set.range (fun j : ℕ => riemannZeta (2 * j + 3)) ∪
Set.range fun j : ℕ =>
((∑' n : ℕ, (-1 : ℝ) ^ n / (2 * n + 1 : ℝ) ^ (2 * j + 2) : ℝ) : ℂ))
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:148
noncomputable def hyperharmonicEulerSum (k : ℕ) (r : ℂ) (p : ℤ) (a : ℂ) : ℂ :=
∑' n : ℕ, generalizedHyperharmonic (n + k) k r / (((n + k : ℕ) : ℂ) + a) ^ p
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:74
Sources
- Paul Thomas Young, Parametric Euler Sums of Generalized Hyperharmonic Numbers, INTEGERS 26 (2026), Article A81, Corollary 11, first part.
TheoremProducts of quarter-shift central binomial odd harmonic sums#●ready to prove
Let \(O_n(k)\) denote the depth-\(k\) odd harmonic number, let \(\beta(s) = \sum_{n \ge 0} (-1)^n / (2n+1)^s\) be the Dirichlet beta function, and let \(B = \mathbb{Q}[\log 2, \pi, \zeta(3), \zeta(5), \ldots, \beta(2), \beta(4), \ldots] \subseteq \mathbb{C}\). For integers \(k_1, k_2 \ge 0\) and \(p_1, p_2 \ge 1\), the series
both converge, and \(S_1 S_2 = \pi x\) for some \(x \in B\).
The terms with \(n < k_i\) vanish, since \(O_n(k_i) = 0\) for \(n < k_i\).
public theorem_wanted centralBinom_oddMultipleHarmonic_quarter_mul_mem_pi_mul
(k₁ k₂ p₁ p₂ : ℕ) (hp₁ : 0 < p₁) (hp₂ : 0 < p₂) :
Summable (fun n : ℕ => ((2 * n).choose n : ℝ) * oddMultipleHarmonic k₁ n /
(4 ^ n * ((n : ℝ) + 1 / 4) ^ p₁)) ∧
Summable (fun n : ℕ => ((2 * n).choose n : ℝ) * oddMultipleHarmonic k₂ n /
(4 ^ n * ((n : ℝ) + 3 / 4) ^ p₂)) ∧
∃ x ∈ eulerSumBetaPeriodAlgebra,
((∑' n : ℕ, ((2 * n).choose n : ℝ) * oddMultipleHarmonic k₁ n /
(4 ^ n * ((n : ℝ) + 1 / 4) ^ p₁) : ℝ) : ℂ) *
((∑' n : ℕ, ((2 * n).choose n : ℝ) * oddMultipleHarmonic k₂ n /
(4 ^ n * ((n : ℝ) + 3 / 4) ^ p₂) : ℝ) : ℂ) = π * x
noncomputable def oddMultipleHarmonic (k n : ℕ) : ℝ :=
∑ s ∈ (Finset.Icc 1 n).powersetCard k, ∏ i ∈ s, ((i : ℝ) - 1 / 2)⁻¹
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:98
noncomputable def eulerSumBetaPeriodAlgebra : Subalgebra ℚ ℂ :=
Algebra.adjoin ℚ
({(Real.log 2 : ℂ), (π : ℂ)} ∪ Set.range (fun j : ℕ => riemannZeta (2 * j + 3)) ∪
Set.range fun j : ℕ =>
((∑' n : ℕ, (-1 : ℝ) ^ n / (2 * n + 1 : ℝ) ^ (2 * j + 2) : ℝ) : ℂ))
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:148
Sources
- Paul Thomas Young, Parametric Euler Sums of Generalized Hyperharmonic Numbers, INTEGERS 26 (2026), Article A81, Corollary 11, second part.
TheoremOdd harmonic central binomial sums with weight \(\binom{n}{m}\) at shift \(1/4\)#●ready to prove
Let \(k, m \ge 0\) and \(p > m\) be integers, let \(O_n(k)\) denote the depth-\(k\) odd harmonic number, let \(\beta\) be the Dirichlet beta function, and let
be the lemniscate constant. Then the series of real numbers
converges, and its sum equals \(\varpi x\) for some \(x\) in the algebra \(\mathbb{Q}[\log 2, \pi, \zeta(2j+1)_{j \ge 1}, \beta(2j)_{j \ge 1}] \subseteq \mathbb{C}\). The terms with \(n < k\) are \(0\), since \(O_n(k) = 0\) for \(n < k\), so this is also the sum over \(n \ge k\).
public theorem_wanted choose_centralBinom_oddMultipleHarmonic_quarter_sum_mem
(k p m : ℕ) (hmp : m < p) :
Summable (fun n : ℕ => (n.choose m : ℝ) * ((2 * n).choose n : ℝ) *
oddMultipleHarmonic k n / (4 ^ n * ((n : ℝ) + 1 / 4) ^ p)) ∧
∃ x ∈ eulerSumBetaPeriodAlgebra,
((∑' n : ℕ, (n.choose m : ℝ) * ((2 * n).choose n : ℝ) * oddMultipleHarmonic k n /
(4 ^ n * ((n : ℝ) + 1 / 4) ^ p) : ℝ) : ℂ) =
(((2 * ∫ t in (0 : ℝ)..1, (Real.sqrt (1 - t ^ 4))⁻¹) : ℝ) : ℂ) * x
noncomputable def oddMultipleHarmonic (k n : ℕ) : ℝ :=
∑ s ∈ (Finset.Icc 1 n).powersetCard k, ∏ i ∈ s, ((i : ℝ) - 1 / 2)⁻¹
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:98
noncomputable def eulerSumBetaPeriodAlgebra : Subalgebra ℚ ℂ :=
Algebra.adjoin ℚ
({(Real.log 2 : ℂ), (π : ℂ)} ∪ Set.range (fun j : ℕ => riemannZeta (2 * j + 3)) ∪
Set.range fun j : ℕ =>
((∑' n : ℕ, (-1 : ℝ) ^ n / (2 * n + 1 : ℝ) ^ (2 * j + 2) : ℝ) : ℂ))
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:148
Sources
- Paul Thomas Young, Parametric Euler Sums of Generalized Hyperharmonic Numbers, INTEGERS 26 (2026), Article A81, Corollary 10, first statement.
TheoremOdd harmonic central binomial sums with weight \(\binom{n}{m}\) at shift \(3/4\)#●ready to prove
Let \(k, m \ge 0\) and \(p > m\) be integers, let \(O_n(k)\) denote the depth-\(k\) odd harmonic number, let \(\beta\) be the Dirichlet beta function, and let
be the lemniscate constant. Then the series of real numbers
converges, and its sum equals \((\pi / \varpi) x\) for some \(x\) in the algebra \(\mathbb{Q}[\log 2, \pi, \zeta(2j+1)_{j \ge 1}, \beta(2j)_{j \ge 1}] \subseteq \mathbb{C}\). The terms with \(n < k\) are \(0\), since \(O_n(k) = 0\) for \(n < k\), so this is also the sum over \(n \ge k\).
public theorem_wanted choose_centralBinom_oddMultipleHarmonic_three_quarter_sum_mem
(k p m : ℕ) (hmp : m < p) :
Summable (fun n : ℕ => (n.choose m : ℝ) * ((2 * n).choose n : ℝ) *
oddMultipleHarmonic k n / (4 ^ n * ((n : ℝ) + 3 / 4) ^ p)) ∧
∃ x ∈ eulerSumBetaPeriodAlgebra,
((∑' n : ℕ, (n.choose m : ℝ) * ((2 * n).choose n : ℝ) * oddMultipleHarmonic k n /
(4 ^ n * ((n : ℝ) + 3 / 4) ^ p) : ℝ) : ℂ) =
((π / (2 * ∫ t in (0 : ℝ)..1, (Real.sqrt (1 - t ^ 4))⁻¹) : ℝ) : ℂ) * x
noncomputable def oddMultipleHarmonic (k n : ℕ) : ℝ :=
∑ s ∈ (Finset.Icc 1 n).powersetCard k, ∏ i ∈ s, ((i : ℝ) - 1 / 2)⁻¹
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:98
noncomputable def eulerSumBetaPeriodAlgebra : Subalgebra ℚ ℂ :=
Algebra.adjoin ℚ
({(Real.log 2 : ℂ), (π : ℂ)} ∪ Set.range (fun j : ℕ => riemannZeta (2 * j + 3)) ∪
Set.range fun j : ℕ =>
((∑' n : ℕ, (-1 : ℝ) ^ n / (2 * n + 1 : ℝ) ^ (2 * j + 2) : ℝ) : ℂ))
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:148
Sources
- Paul Thomas Young, Parametric Euler Sums of Generalized Hyperharmonic Numbers, INTEGERS 26 (2026), Article A81, Corollary 10, second statement.
TheoremCentral binomial odd harmonic sums over \(\binom{n+m}{m}\) at the shift \(1/4\)#●ready to prove
Let \(O_n(k)\) denote the depth-\(k\) odd harmonic number, let \(\beta(s) = \sum_{n \ge 0} (-1)^n / (2n+1)^s\) be the Dirichlet beta function, let \(B = \mathbb{Q}[\log 2, \pi, \zeta(3), \zeta(5), \ldots, \beta(2), \beta(4), \ldots] \subseteq \mathbb{C}\), and let \(\varpi = 2 \int_0^1 dt / \sqrt{1 - t^4}\) be the lemniscate constant. For integers \(k \ge 0\), \(p \ge 1\) and \(m \ge 1\), the series
converges, and its sum equals \(q + \varpi x\) for some rational number \(q\) and some \(x \in B\).
The terms with \(n < k\) vanish, since \(O_n(k) = 0\) for \(n < k\).
public theorem_wanted centralBinom_oddMultipleHarmonic_div_choose_quarter_sum_mem
(k p m : ℕ) (hp : 0 < p) (hm : 0 < m) :
Summable (fun n : ℕ => ((2 * n).choose n : ℝ) * oddMultipleHarmonic k n /
(4 ^ n * ((n + m).choose m : ℝ) * ((n : ℝ) + 1 / 4) ^ p)) ∧
∃ q : ℚ, ∃ x ∈ eulerSumBetaPeriodAlgebra,
((∑' n : ℕ, ((2 * n).choose n : ℝ) * oddMultipleHarmonic k n /
(4 ^ n * ((n + m).choose m : ℝ) * ((n : ℝ) + 1 / 4) ^ p) : ℝ) : ℂ) =
q + (((2 * ∫ t in (0 : ℝ)..1, (Real.sqrt (1 - t ^ 4))⁻¹) : ℝ) : ℂ) * x
noncomputable def oddMultipleHarmonic (k n : ℕ) : ℝ :=
∑ s ∈ (Finset.Icc 1 n).powersetCard k, ∏ i ∈ s, ((i : ℝ) - 1 / 2)⁻¹
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:98
noncomputable def eulerSumBetaPeriodAlgebra : Subalgebra ℚ ℂ :=
Algebra.adjoin ℚ
({(Real.log 2 : ℂ), (π : ℂ)} ∪ Set.range (fun j : ℕ => riemannZeta (2 * j + 3)) ∪
Set.range fun j : ℕ =>
((∑' n : ℕ, (-1 : ℝ) ^ n / (2 * n + 1 : ℝ) ^ (2 * j + 2) : ℝ) : ℂ))
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:148
Sources
- Paul Thomas Young, Parametric Euler Sums of Generalized Hyperharmonic Numbers, INTEGERS 26 (2026), Article A81, Corollary 10, third statement.
TheoremCentral binomial odd harmonic sums over \(\binom{n+m}{m}\) at the shift \(3/4\)#●ready to prove
Let \(O_n(k)\) denote the depth-\(k\) odd harmonic number, let \(\beta(s) = \sum_{n \ge 0} (-1)^n / (2n+1)^s\) be the Dirichlet beta function, let \(B = \mathbb{Q}[\log 2, \pi, \zeta(3), \zeta(5), \ldots, \beta(2), \beta(4), \ldots] \subseteq \mathbb{C}\), and let \(\varpi = 2 \int_0^1 dt / \sqrt{1 - t^4}\) be the lemniscate constant. For integers \(k \ge 0\), \(p \ge 1\) and \(m \ge 1\), the series
converges, and its sum equals \(q + (\pi / \varpi) x\) for some rational number \(q\) and some \(x \in B\).
The terms with \(n < k\) vanish, since \(O_n(k) = 0\) for \(n < k\).
public theorem_wanted centralBinom_oddMultipleHarmonic_div_choose_three_quarter_sum_mem
(k p m : ℕ) (hp : 0 < p) (hm : 0 < m) :
Summable (fun n : ℕ => ((2 * n).choose n : ℝ) * oddMultipleHarmonic k n /
(4 ^ n * ((n + m).choose m : ℝ) * ((n : ℝ) + 3 / 4) ^ p)) ∧
∃ q : ℚ, ∃ x ∈ eulerSumBetaPeriodAlgebra,
((∑' n : ℕ, ((2 * n).choose n : ℝ) * oddMultipleHarmonic k n /
(4 ^ n * ((n + m).choose m : ℝ) * ((n : ℝ) + 3 / 4) ^ p) : ℝ) : ℂ) =
q + ((π / (2 * ∫ t in (0 : ℝ)..1, (Real.sqrt (1 - t ^ 4))⁻¹) : ℝ) : ℂ) * x
noncomputable def oddMultipleHarmonic (k n : ℕ) : ℝ :=
∑ s ∈ (Finset.Icc 1 n).powersetCard k, ∏ i ∈ s, ((i : ℝ) - 1 / 2)⁻¹
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:98
noncomputable def eulerSumBetaPeriodAlgebra : Subalgebra ℚ ℂ :=
Algebra.adjoin ℚ
({(Real.log 2 : ℂ), (π : ℂ)} ∪ Set.range (fun j : ℕ => riemannZeta (2 * j + 3)) ∪
Set.range fun j : ℕ =>
((∑' n : ℕ, (-1 : ℝ) ^ n / (2 * n + 1 : ℝ) ^ (2 * j + 2) : ℝ) : ℂ))
WantedExt/NumberTheory/HyperharmonicEulerSumsWanted.lean:148
Sources
- Paul Thomas Young, Parametric Euler Sums of Generalized Hyperharmonic Numbers, INTEGERS 26 (2026), Article A81, Corollary 10, fourth statement.
Loading dependency explorer…
The dependency explorer needs JavaScript and its graph data.