Completely additive height functions
DefinitionCompletely additive functions on the positive integers#✓fully proved
A function \(H : \mathbb{N} \to \mathbb{N}\) is completely additive on the positive integers if \(H(ab) = H(a) + H(b)\) for all integers \(a, b \ge 1\).
public def IsCompletelyAdditiveNat (H : ℕ → ℕ) : Prop :=
∀ a b : ℕ, 0 < a → 0 < b → H (a * b) = H a + H b
Sources
- Hartosh Singh Bal, Completely Additive Height Functions: Profile Laws, Matula Bounds, and Inverse Growth, INTEGERS 26 (2026), Article A116, Section 2 (definition of a height function).
Used by 8
TheoremA completely additive function vanishes at one#✓fully proved
If \(H : \mathbb{N} \to \mathbb{N}\) is completely additive on the positive integers, that is, \(H(ab) = H(a) + H(b)\) for all \(a, b \ge 1\), then \(H(1) = 0\).
public theorem IsCompletelyAdditiveNat.map_one {H : ℕ → ℕ} (hH : IsCompletelyAdditiveNat H) :
H 1 = 0
public def IsCompletelyAdditiveNat (H : ℕ → ℕ) : Prop :=
∀ a b : ℕ, 0 < a → 0 < b → H (a * b) = H a + H b
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:42
Sources
- Hartosh Singh Bal, Completely Additive Height Functions: Profile Laws, Matula Bounds, and Inverse Growth, INTEGERS 26 (2026), Article A116, Section 2 (definition of a height function).
Uses 1
StructureHeight functions#✓fully proved
A function \(H : \mathbb{N} \to \mathbb{N}\) is a height function if it is completely additive on the positive integers, that is \(H(ab) = H(a) + H(b)\) for all integers \(a, b \ge 1\); if \(H(p) \ge 1\) for every prime \(p\); and if for every \(n \ge 1\) only finitely many primes \(p\) have \(H(p) = n\).
Complete additivity forces \(H(1) = 0\).
public structure IsHeightFunction (H : ℕ → ℕ) : Prop where
additive : IsCompletelyAdditiveNat H
prime_pos : ∀ p : ℕ, p.Prime → 1 ≤ H p
finite_primes : ∀ n : ℕ, 1 ≤ n → {p : ℕ | p.Prime ∧ H p = n}.Finite
public def IsCompletelyAdditiveNat (H : ℕ → ℕ) : Prop :=
∀ a b : ℕ, 0 < a → 0 < b → H (a * b) = H a + H b
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:42
Sources
- Hartosh Singh Bal, Completely Additive Height Functions: Profile Laws, Matula Bounds, and Inverse Growth, INTEGERS 26 (2026), Article A116, Section 2 (definition of a height function).
Uses 1 · Used by 7
DefinitionThe \(i\)-th prime#✓fully proved
For a natural number \(i \ge 1\), \(p_i\) denotes the \(i\)-th prime in increasing order, so that \(p_1 = 2\). For \(i = 0\) the definition also gives \(p_0 = 2\).
Sources
- Hartosh Singh Bal, Completely Additive Height Functions: Profile Laws, Matula Bounds, and Inverse Growth, INTEGERS 26 (2026), Article A116, Section 2 (notation
pᵢ).
Used by 10
TheoremThe first three primes#✓fully proved
With primes indexed from one, the first three primes are \(p_1 = 2\), \(p_2 = 3\) and \(p_3 = 5\).
public theorem nthPrimeOne_one_two_three :
nthPrimeOne 1 = 2 ∧ nthPrimeOne 2 = 3 ∧ nthPrimeOne 3 = 5
public noncomputable def nthPrimeOne (i : ℕ) : ℕ :=
Nat.nth Nat.Prime (i - 1)
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:71
Sources
- Hartosh Singh Bal, Completely Additive Height Functions: Profile Laws, Matula Bounds, and Inverse Growth, INTEGERS 26 (2026), Article A116, Section 2 (notation
pᵢ).
Uses 1
DefinitionCompletely additive function with prescribed prime values#✓fully proved
Given a function \(v : \mathbb{N} \to \mathbb{N}\), this is the function \(\mathbb{N} \to \mathbb{N}\) that sends each \(m \ge 1\) with prime factorization \(m = \prod_p p^{a_p}\) to \(\sum_p a_p\, v(p)\), and sends \(0\) to \(0\). Only the values of \(v\) at primes matter.
public noncomputable def additiveOfPrimeValues (v : ℕ → ℕ) (m : ℕ) : ℕ :=
m.factorization.sum fun p a => a * v p
Sources
- Hartosh Singh Bal, Completely Additive Height Functions: Profile Laws, Matula Bounds, and Inverse Growth, INTEGERS 26 (2026), Article A116, Proposition 3 (extension by complete additivity).
Used by 6
TheoremThe completely additive extension vanishes at one#✓fully proved
For every function \(v : \mathbb{N} \to \mathbb{N}\), the completely additive function with prime values \(v\) (which sends \(m = \prod_p p^{a_p}\) to \(\sum_p a_p\, v(p)\)) takes the value \(0\) at \(1\).
public theorem additiveOfPrimeValues_one (v : ℕ → ℕ) : additiveOfPrimeValues v 1 = 0
public noncomputable def additiveOfPrimeValues (v : ℕ → ℕ) (m : ℕ) : ℕ :=
m.factorization.sum fun p a => a * v p
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:88
Sources
- Hartosh Singh Bal, Completely Additive Height Functions: Profile Laws, Matula Bounds, and Inverse Growth, INTEGERS 26 (2026), Article A116, Proposition 3 (extension by complete additivity).
Uses 1
TheoremThe completely additive extension at a prime#✓fully proved
For every function \(v : \mathbb{N} \to \mathbb{N}\) and every prime \(p\), the completely additive function with prime values \(v\) (which sends \(m = \prod_q q^{a_q}\) to \(\sum_q a_q\, v(q)\)) takes the value \(v(p)\) at \(p\).
public theorem additiveOfPrimeValues_prime (v : ℕ → ℕ) {p : ℕ} (hp : p.Prime) :
additiveOfPrimeValues v p = v p
public noncomputable def additiveOfPrimeValues (v : ℕ → ℕ) (m : ℕ) : ℕ :=
m.factorization.sum fun p a => a * v p
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:88
Sources
- Hartosh Singh Bal, Completely Additive Height Functions: Profile Laws, Matula Bounds, and Inverse Growth, INTEGERS 26 (2026), Article A116, Proposition 3 (extension by complete additivity).
Uses 1
TheoremThe extension from prime values is completely additive#✓fully proved
For every function \(v : \mathbb{N} \to \mathbb{N}\), let \(H_v\) be the function with \(H_v(m) = \sum_p a_p v(p)\) for \(m = \prod_p p^{a_p}\). Then \(H_v\) is completely additive on the positive integers: \(H_v(ab) = H_v(a) + H_v(b)\) for all integers \(a, b \ge 1\).
public theorem isCompletelyAdditiveNat_additiveOfPrimeValues (v : ℕ → ℕ) :
IsCompletelyAdditiveNat (additiveOfPrimeValues v)
public def IsCompletelyAdditiveNat (H : ℕ → ℕ) : Prop :=
∀ a b : ℕ, 0 < a → 0 < b → H (a * b) = H a + H b
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:42
public noncomputable def additiveOfPrimeValues (v : ℕ → ℕ) (m : ℕ) : ℕ :=
m.factorization.sum fun p a => a * v p
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:88
Sources
- Hartosh Singh Bal, Completely Additive Height Functions: Profile Laws, Matula Bounds, and Inverse Growth, INTEGERS 26 (2026), Article A116, Proposition 3 (extension by complete additivity).
TheoremLevel construction for a height function#●ready to prove
Let \(H\) be a height function and let \(n \ge 1\), and for \(k \ge 0\) write \(\mathcal{H}_k = \{m \ge 1 : H(m) = k\}\). Then \(\mathcal{H}_n\) is the union of two sets: the set of products \(p m'\), where \(p\) is a prime with \(H(p) = \ell\) for some integer \(\ell\) with \(1 \le \ell \le \lfloor n/2 \rfloor\) and \(m' \in \mathcal{H}_{n - \ell}\); and the set of primes \(p\) with \(H(p) = n\).
public theorem_wanted heightLevel_eq_products_union_primes {H : ℕ → ℕ} (hH : IsHeightFunction H)
{n : ℕ} (hn : 1 ≤ n) :
{m : ℕ | 0 < m ∧ H m = n} =
{m : ℕ | ∃ ℓ, 1 ≤ ℓ ∧ ℓ ≤ n / 2 ∧ ∃ p m', p.Prime ∧ H p = ℓ ∧ 0 < m' ∧ H m' = n - ℓ ∧
m = p * m'} ∪ {p : ℕ | p.Prime ∧ H p = n}
public structure IsHeightFunction (H : ℕ → ℕ) : Prop where
additive : IsCompletelyAdditiveNat H
prime_pos : ∀ p : ℕ, p.Prime → 1 ≤ H p
finite_primes : ∀ n : ℕ, 1 ≤ n → {p : ℕ | p.Prime ∧ H p = n}.Finite
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:62
Sources
- Hartosh Singh Bal, Completely Additive Height Functions: Profile Laws, Matula Bounds, and Inverse Growth, INTEGERS 26 (2026), Article A116, Theorem 1.
Uses 1
TheoremHeight levels are counted by coloured partitions#●ready to prove
Let \(H\) be a height function and, for \(\ell \ge 1\), let \(\pi_\ell\) be the number of primes \(p\) with \(H(p) = \ell\). For every \(n \ge 0\), the number of positive integers \(m\) with \(H(m) = n\) equals the number of families \((\alpha_{\ell,j})\) of nonnegative integers indexed by \(1 \le \ell \le n\) and \(0 \le j < \pi_\ell\) with \(\sum_{\ell,j} \ell\, \alpha_{\ell,j} = n\), that is, the number of partitions of \(n\) in which a part of size \(\ell\) comes in \(\pi_\ell\) colours.
This is the level-by-level form of the statement that \(H\) induces a height-preserving bijection between the positive integers and such coloured partitions.
public theorem_wanted heightLevel_card_eq_multipartitions {H : ℕ → ℕ} (hH : IsHeightFunction H)
(n : ℕ) :
Nat.card {m : ℕ // 0 < m ∧ H m = n} =
Nat.card {α : (Σ ℓ : Fin n, Fin {p : ℕ | p.Prime ∧ H p = ℓ.1 + 1}.ncard) → ℕ //
∑ x, (x.1.1 + 1) * α x = n}
public structure IsHeightFunction (H : ℕ → ℕ) : Prop where
additive : IsCompletelyAdditiveNat H
prime_pos : ∀ p : ℕ, p.Prime → 1 ≤ H p
finite_primes : ∀ n : ℕ, 1 ≤ n → {p : ℕ | p.Prime ∧ H p = n}.Finite
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:62
Sources
- Hartosh Singh Bal, Completely Additive Height Functions: Profile Laws, Matula Bounds, and Inverse Growth, INTEGERS 26 (2026), Article A116, Proposition 1.
Uses 1
TheoremRealizing a prescribed profile#●ready to prove
Let \((\pi_k)_{k \ge 1}\) be nonnegative integers with \(\pi_k \ne 0\) for infinitely many \(k\) (equivalently, \(\sum_k \pi_k = \infty\)). Then there is a height function \(H\) such that for every \(k \ge 1\) exactly \(\pi_k\) primes \(p\) satisfy \(H(p) = k\).
public theorem_wanted exists_heightFunction_with_profile (π : ℕ → ℕ)
(hπ : {k : ℕ | 1 ≤ k ∧ π k ≠ 0}.Infinite) :
∃ H : ℕ → ℕ, IsHeightFunction H ∧ ∀ k, 1 ≤ k → {p : ℕ | p.Prime ∧ H p = k}.ncard = π k
public structure IsHeightFunction (H : ℕ → ℕ) : Prop where
additive : IsCompletelyAdditiveNat H
prime_pos : ∀ p : ℕ, p.Prime → 1 ≤ H p
finite_primes : ∀ n : ℕ, 1 ≤ n → {p : ℕ | p.Prime ∧ H p = n}.Finite
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:62
Sources
- Hartosh Singh Bal, Completely Additive Height Functions: Profile Laws, Matula Bounds, and Inverse Growth, INTEGERS 26 (2026), Article A116, Proposition 2.
Uses 1
TheoremHeight functions with prescribed values at the primes#●ready to prove
Let \(f : \mathbb{N} \to \mathbb{N}\) satisfy \(f(i) \ge 1\) for every \(i \ge 1\), and suppose that for every \(t \ge 1\) only finitely many \(i \ge 1\) have \(f(i) = t\). Let \(H\) be the completely additive function on the positive integers with \(H(p) = f(i)\) when \(p\) is the \(i\)-th prime (so \(H(2) = f(1)\)). Then \(H\) is a height function, and for every \(n \ge 1\) the number of primes \(p\) with \(H(p) = n\) equals the number of integers \(i \ge 1\) with \(f(i) = n\).
public theorem_wanted isHeightFunction_of_finite_fibers (f : ℕ → ℕ)
(hpos : ∀ i, 1 ≤ i → 1 ≤ f i) (hfin : ∀ t, 1 ≤ t → {i : ℕ | 1 ≤ i ∧ f i = t}.Finite) :
let H := additiveOfPrimeValues fun p => f (Nat.count Nat.Prime p + 1)
IsHeightFunction H ∧
∀ n, 1 ≤ n → {p : ℕ | p.Prime ∧ H p = n}.ncard = {i : ℕ | 1 ≤ i ∧ f i = n}.ncard
public noncomputable def additiveOfPrimeValues (v : ℕ → ℕ) (m : ℕ) : ℕ :=
m.factorization.sum fun p a => a * v p
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:88
public structure IsHeightFunction (H : ℕ → ℕ) : Prop where
additive : IsCompletelyAdditiveNat H
prime_pos : ∀ p : ℕ, p.Prime → 1 ≤ H p
finite_primes : ∀ n : ℕ, 1 ≤ n → {p : ℕ | p.Prime ∧ H p = n}.Finite
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:62
Sources
- Hartosh Singh Bal, Completely Additive Height Functions: Profile Laws, Matula Bounds, and Inverse Growth, INTEGERS 26 (2026), Article A116, Proposition 3.
Uses 2
TheoremThe largest integer at each height for the prime-index height#●ready to prove
Let \(H\) be the completely additive function on the positive integers with \(H(p) = i\) when \(p\) is the \(i\)-th prime, so that \(H(2) = 1\), \(H(3) = 2\) and \(H(5) = 3\). For every natural number \(n\), \(2^n\) is the largest integer \(m \ge 1\) with \(H(m) = n\).
public theorem_wanted isGreatest_primeIndexHeight_level (n : ℕ) :
IsGreatest
{m : ℕ | 0 < m ∧ additiveOfPrimeValues (fun p => Nat.count Nat.Prime p + 1) m = n} (2 ^ n)
public noncomputable def additiveOfPrimeValues (v : ℕ → ℕ) (m : ℕ) : ℕ :=
m.factorization.sum fun p a => a * v p
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:88
Sources
- Hartosh Singh Bal, Completely Additive Height Functions: Profile Laws, Matula Bounds, and Inverse Growth, INTEGERS 26 (2026), Article A116, Proposition 4.
Uses 1
DefinitionGutman–Ivić maximal Matula numbers#✓fully proved
Write \(p_i\) for the \(i\)-th prime, indexed from one (\(p_1 = 2\)). For a natural number \(n\), \(P(n)\) is the result of applying \(i \mapsto p_i\) to \(8\) exactly \(n - 3\) times, so \(P(4) = p_8\) and \(P(n) = p_{P(n-1)}\) for \(n \ge 5\). For \(n \le 3\) there are no applications and \(P(n) = 8\).
public noncomputable def matulaMax (n : ℕ) : ℕ :=
nthPrimeOne^[n - 3] 8
public noncomputable def nthPrimeOne (i : ℕ) : ℕ :=
Nat.nth Nat.Prime (i - 1)
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:71
Sources
- Hartosh Singh Bal, Completely Additive Height Functions: Profile Laws, Matula Bounds, and Inverse Growth, INTEGERS 26 (2026), Article A116, Section 3.2 (definition of
P(n)).
Uses 1 · Used by 3
TheoremThe first maximal Matula number#✓fully proved
The Gutman–Ivić maximal Matula number \(P(4)\) equals \(p_8\), the eighth prime.
public theorem matulaMax_four : matulaMax 4 = nthPrimeOne 8
public noncomputable def matulaMax (n : ℕ) : ℕ :=
nthPrimeOne^[n - 3] 8
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:185
public noncomputable def nthPrimeOne (i : ℕ) : ℕ :=
Nat.nth Nat.Prime (i - 1)
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:71
Sources
- Hartosh Singh Bal, Completely Additive Height Functions: Profile Laws, Matula Bounds, and Inverse Growth, INTEGERS 26 (2026), Article A116, Section 3.2 (definition of
P(n)).
Uses 2
TheoremRecursion for the maximal Matula numbers#✓fully proved
Write \(p_i\) for the \(i\)-th prime, indexed from one (\(p_1 = 2\)). For every \(n \ge 4\) the Gutman–Ivić maximal Matula numbers satisfy \(P(n + 1) = p_{P(n)}\).
public theorem matulaMax_succ {n : ℕ} (hn : 4 ≤ n) :
matulaMax (n + 1) = nthPrimeOne (matulaMax n)
public noncomputable def matulaMax (n : ℕ) : ℕ :=
nthPrimeOne^[n - 3] 8
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:185
public noncomputable def nthPrimeOne (i : ℕ) : ℕ :=
Nat.nth Nat.Prime (i - 1)
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:71
Sources
- Hartosh Singh Bal, Completely Additive Height Functions: Profile Laws, Matula Bounds, and Inverse Growth, INTEGERS 26 (2026), Article A116, Section 3.2 (definition of
P(n)).
Uses 2
TheoremLargest number at a given Matula height#●ready to prove
Write \(p_i\) for the \(i\)-th prime, indexed from one (\(p_1 = 2\)), and let \(H\) be a function from the natural numbers to themselves that is completely additive on the positive integers and satisfies \(H(p_i) = H(i) + 1\) for every \(i \ge 1\) (Matula's height function). Then for every \(n \ge 4\) the largest positive integer \(m\) with \(H(m) = n\) is the Gutman–Ivić maximal Matula number \(P(n)\).
public theorem_wanted matula_isGreatest_level {H : ℕ → ℕ} (hH : IsCompletelyAdditiveNat H)
(hM : ∀ i, 1 ≤ i → H (nthPrimeOne i) = H i + 1) {n : ℕ} (hn : 4 ≤ n) :
IsGreatest {m : ℕ | 0 < m ∧ H m = n} (matulaMax n)
public def IsCompletelyAdditiveNat (H : ℕ → ℕ) : Prop :=
∀ a b : ℕ, 0 < a → 0 < b → H (a * b) = H a + H b
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:42
public noncomputable def nthPrimeOne (i : ℕ) : ℕ :=
Nat.nth Nat.Prime (i - 1)
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:71
public noncomputable def matulaMax (n : ℕ) : ℕ :=
nthPrimeOne^[n - 3] 8
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:185
Sources
- Hartosh Singh Bal, Completely Additive Height Functions: Profile Laws, Matula Bounds, and Inverse Growth, INTEGERS 26 (2026), Article A116, Proposition 5.
Uses 3
TheoremSmallest number at a given Matula height#●ready to prove
Write \(p_i\) for the \(i\)-th prime, indexed from one (\(p_1 = 2\)), and let \(H\) be a function from the natural numbers to themselves that is completely additive on the positive integers and satisfies \(H(p_i) = H(i) + 1\) for every \(i \ge 1\) (Matula's height function). Then for every \(n > 1\) the smallest positive integer \(m\) with \(H(m) = n\) is \(3 \cdot 5^{(n-2)/3}\) if \(n \equiv 2 \pmod 3\), \(5^{n/3}\) if \(n \equiv 0 \pmod 3\), and \(9 \cdot 5^{(n-4)/3}\) if \(n \equiv 1 \pmod 3\).
public theorem_wanted matula_isLeast_level {H : ℕ → ℕ} (hH : IsCompletelyAdditiveNat H)
(hM : ∀ i, 1 ≤ i → H (nthPrimeOne i) = H i + 1) {n : ℕ} (hn : 1 < n) :
IsLeast {m : ℕ | 0 < m ∧ H m = n}
(if n % 3 = 2 then 3 * 5 ^ ((n - 2) / 3)
else if n % 3 = 0 then 5 ^ (n / 3) else 9 * 5 ^ ((n - 4) / 3))
public def IsCompletelyAdditiveNat (H : ℕ → ℕ) : Prop :=
∀ a b : ℕ, 0 < a → 0 < b → H (a * b) = H a + H b
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:42
public noncomputable def nthPrimeOne (i : ℕ) : ℕ :=
Nat.nth Nat.Prime (i - 1)
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:71
Sources
- Hartosh Singh Bal, Completely Additive Height Functions: Profile Laws, Matula Bounds, and Inverse Growth, INTEGERS 26 (2026), Article A116, Proposition 6.
Uses 2
TheoremRosser–Schoenfeld bounds for the \(k\)-th prime#●ready to prove
Write \(p_k\) for the \(k\)-th prime, indexed from one (\(p_1 = 2\)). There is a natural number \(k_0\) such that every \(k \ge k_0\) satisfies \(k (\log k + \log \log k - 1) \le p_k \le k (\log k + \log \log k)\).
public theorem_wanted nthPrimeOne_rosser_schoenfeld :
∃ k₀ : ℕ, ∀ k ≥ k₀,
(k : ℝ) * (Real.log k + Real.log (Real.log k) - 1) ≤ nthPrimeOne k ∧
(nthPrimeOne k : ℝ) ≤ k * (Real.log k + Real.log (Real.log k))
public noncomputable def nthPrimeOne (i : ℕ) : ℕ :=
Nat.nth Nat.Prime (i - 1)
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:71
Sources
- Hartosh Singh Bal, Completely Additive Height Functions: Profile Laws, Matula Bounds, and Inverse Growth, INTEGERS 26 (2026), Article A116, Lemma 1, quoting Rosser and Schoenfeld.
Uses 1
TheoremThe \(k\)-th prime exceeds \(2k\) for \(k \ge 6\)#●ready to prove
For every natural number \(k \ge 6\), the \(k\)-th prime \(p_k\), indexed from one so that \(p_1 = 2\), satisfies \(p_k > 2k\).
public theorem_wanted two_mul_lt_nthPrimeOne {k : ℕ} (hk : 6 ≤ k) : 2 * k < nthPrimeOne k
public noncomputable def nthPrimeOne (i : ℕ) : ℕ :=
Nat.nth Nat.Prime (i - 1)
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:71
Sources
- Hartosh Singh Bal, Completely Additive Height Functions: Profile Laws, Matula Bounds, and Inverse Growth, INTEGERS 26 (2026), Article A116, Lemma 2.
Uses 1
TheoremPrime-power decomposition of summatory heights#●ready to prove
Let \(H : \mathbb{N} \to \mathbb{N}\) be completely additive on the positive integers, and let \(x \ge 2\) be a natural number. Then
where \(p\) runs over the primes \(p \le x\).
The paper states this for real \(x \ge 2\); both sides depend on \(x\) only through \(\lfloor x \rfloor\).
public theorem_wanted sum_height_eq_sum_prime_powers {H : ℕ → ℕ} (hH : IsCompletelyAdditiveNat H)
{x : ℕ} (hx : 2 ≤ x) :
∑ m ∈ Finset.Icc 1 x, H m =
∑ p ∈ (Finset.Iic x).filter Nat.Prime,
∑ α ∈ (Finset.Icc 1 x).filter (fun α => p ^ α ≤ x), H p * (x / p ^ α)
public def IsCompletelyAdditiveNat (H : ℕ → ℕ) : Prop :=
∀ a b : ℕ, 0 < a → 0 < b → H (a * b) = H a + H b
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:42
Sources
- Hartosh Singh Bal, Completely Additive Height Functions: Profile Laws, Matula Bounds, and Inverse Growth, INTEGERS 26 (2026), Article A116, Proposition 7, Equation (6).
Uses 1
TheoremPrime reduction for summatory heights#●ready to prove
Let \(H : \mathbb{N} \to \mathbb{N}\) be completely additive on the positive integers, and for real \(x\) let \(F(x) = \sum_{1 \le m \le x} H(m)\). Then there is a real constant \(C\) such that for every real \(x \ge 2\),
where both sums run over the primes \(p \le x\).
public theorem_wanted sum_height_asymptotic {H : ℕ → ℕ} (hH : IsCompletelyAdditiveNat H) :
∃ C : ℝ, ∀ x : ℝ, 2 ≤ x →
|(∑ m ∈ Finset.Icc 1 ⌊x⌋₊, (H m : ℝ)) -
x * ∑ p ∈ (Finset.Iic ⌊x⌋₊).filter Nat.Prime, (H p : ℝ) / ((p : ℝ) - 1)| ≤
C * Real.log x * ∑ p ∈ (Finset.Iic ⌊x⌋₊).filter Nat.Prime, (H p : ℝ)
public def IsCompletelyAdditiveNat (H : ℕ → ℕ) : Prop :=
∀ a b : ℕ, 0 < a → 0 < b → H (a * b) = H a + H b
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:42
Sources
- Hartosh Singh Bal, Completely Additive Height Functions: Profile Laws, Matula Bounds, and Inverse Growth, INTEGERS 26 (2026), Article A116, Proposition 7, "In particular" display.
Uses 1
TheoremIteratively defined height functions#●ready to prove
Write \(p_i\) for the \(i\)-th prime, so \(p_1 = 2\). Let \(f : \mathbb{N} \to \mathbb{N}\) satisfy \(f(1) = 1\); suppose that for all integers \(i, r \ge 1\), if \(p_r\) divides \(f(i)\) then \(r < i\) (this forces \(f(i) \ge 1\) for \(i \ge 1\)); and suppose that for every \(t \ge 1\) only finitely many \(i \ge 1\) have \(f(i) = t\). Let \(j \ge 1\). Then there is a function \(H : \mathbb{N} \to \mathbb{N}\) that is completely additive on the positive integers and satisfies \(H(p_i) = H(f(i)) + j\) for every \(i \ge 1\), and every such \(H\) is a height function.
The case \(i = 1\) gives \(H(2) = j\).
public theorem_wanted isHeightFunction_of_iterative (f : ℕ → ℕ) (hf1 : f 1 = 1)
(hdiv : ∀ i r, 1 ≤ i → 1 ≤ r → nthPrimeOne r ∣ f i → r < i)
(hfin : ∀ t, 1 ≤ t → {i : ℕ | 1 ≤ i ∧ f i = t}.Finite) {j : ℕ} (hj : 1 ≤ j) :
(∃ H : ℕ → ℕ, IsCompletelyAdditiveNat H ∧ ∀ i, 1 ≤ i → H (nthPrimeOne i) = H (f i) + j) ∧
∀ H : ℕ → ℕ, IsCompletelyAdditiveNat H → (∀ i, 1 ≤ i → H (nthPrimeOne i) = H (f i) + j) →
IsHeightFunction H
public noncomputable def nthPrimeOne (i : ℕ) : ℕ :=
Nat.nth Nat.Prime (i - 1)
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:71
public def IsCompletelyAdditiveNat (H : ℕ → ℕ) : Prop :=
∀ a b : ℕ, 0 < a → 0 < b → H (a * b) = H a + H b
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:42
public structure IsHeightFunction (H : ℕ → ℕ) : Prop where
additive : IsCompletelyAdditiveNat H
prime_pos : ∀ p : ℕ, p.Prime → 1 ≤ H p
finite_primes : ∀ n : ℕ, 1 ≤ n → {p : ℕ | p.Prime ∧ H p = n}.Finite
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:62
Sources
- Hartosh Singh Bal, Completely Additive Height Functions: Profile Laws, Matula Bounds, and Inverse Growth, INTEGERS 26 (2026), Article A116, Theorem 2.
Uses 3
DefinitionWeighted partition coefficients#✓fully proved
Given real weights \((b_k)_{k \ge 0}\), the coefficients \(a_n\) of the weighted partition product \(\prod_{k \ge 1} (1 - q^k)^{-b_k}\) are defined by \(a_0 = 1\) and, for \(n \ge 1\), the logarithmic-derivative recurrence \(n a_n = \sum_{m=1}^{n} \big(\sum_{k \mid m} k b_k\big) a_{n-m}\) (valid for real exponents \(b_k\)). The value \(b_0\) is unused.
public noncomputable def weightedPartitionCoeff (b : ℕ → ℝ) : ℕ → ℝ
| 0 => 1
| n + 1 => (∑ m : Fin (n + 1), (∑ k ∈ (m.1 + 1).divisors, (k : ℝ) * b k) *
weightedPartitionCoeff b (n - m.1)) / (n + 1)
decreasing_by omega
Sources
- Hartosh Singh Bal, Completely Additive Height Functions: Profile Laws, Matula Bounds, and Inverse Growth, INTEGERS 26 (2026), Article A116, Section 6.1 (weighted partitions
F(q) = ∑ aₙ qⁿ).
Used by 3
TheoremZeroth weighted partition coefficient#✓fully proved
For every sequence of real weights \((b_k)\), the weighted partition coefficient \(a_0\) attached to \((b_k)\) equals \(1\).
public theorem weightedPartitionCoeff_zero (b : ℕ → ℝ) : weightedPartitionCoeff b 0 = 1
public noncomputable def weightedPartitionCoeff (b : ℕ → ℝ) : ℕ → ℝ
| 0 => 1
| n + 1 => (∑ m : Fin (n + 1), (∑ k ∈ (m.1 + 1).divisors, (k : ℝ) * b k) *
weightedPartitionCoeff b (n - m.1)) / (n + 1)
decreasing_by omega
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:291
Sources
- Hartosh Singh Bal, Completely Additive Height Functions: Profile Laws, Matula Bounds, and Inverse Growth, INTEGERS 26 (2026), Article A116, Section 6.1 (weighted partitions
F(q) = ∑ aₙ qⁿ).
Uses 1
TheoremFirst weighted partition coefficient#✓fully proved
For every sequence of real weights \((b_k)\), the weighted partition coefficient \(a_1\) attached to \((b_k)\) equals \(b_1\).
public theorem weightedPartitionCoeff_one (b : ℕ → ℝ) : weightedPartitionCoeff b 1 = b 1
public noncomputable def weightedPartitionCoeff (b : ℕ → ℝ) : ℕ → ℝ
| 0 => 1
| n + 1 => (∑ m : Fin (n + 1), (∑ k ∈ (m.1 + 1).divisors, (k : ℝ) * b k) *
weightedPartitionCoeff b (n - m.1)) / (n + 1)
decreasing_by omega
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:291
Sources
- Hartosh Singh Bal, Completely Additive Height Functions: Profile Laws, Matula Bounds, and Inverse Growth, INTEGERS 26 (2026), Article A116, Section 6.1 (weighted partitions
F(q) = ∑ aₙ qⁿ).
Uses 1
DefinitionMeinardus' hypotheses#✓fully proved
Let \(b_1, b_2, \dots\) be real weights (the value \(b_0\) is unused), let \(\alpha\) and \(A\) be reals and let \(G : \mathbb{C} \to \mathbb{C}\). The weights satisfy Meinardus' hypotheses (M1)–(M3) with exponent \(\alpha\), residue \(A\) and continuation \(G\) if there is \(C_0\) with \(0 < C_0 < 1\) such that: (M1) for \(\operatorname{Re} s > \alpha\) the Dirichlet series \(B(s) = \sum_{k \ge 1} b_k k^{-s}\) converges absolutely and equals \(G(s)\), and there is a function \(R : \mathbb{C} \to \mathbb{C}\) that is complex differentiable at every point of the closed half-plane \(\{s : \operatorname{Re} s \ge -C_0\}\) relative to that half-plane, with \(G(s) = A/(s - \alpha) + R(s)\) for all \(s \ne \alpha\) with \(\operatorname{Re} s \ge -C_0\); (M2) there are \(c > 0\) and reals \(K\), \(T\) with \(|G(s)| \le K (1 + |\operatorname{Im} s|)^c\) whenever \(\operatorname{Re} s \ge -C_0\) and \(|\operatorname{Im} s| \ge T\); (M3) there are \(\varepsilon > 0\), \(C_3 > 0\) and \(y_0 > 0\) such that, with \(g(\tau) = \sum_{k \ge 1} b_k e^{-k\tau}\), the inequality \(\operatorname{Re} g(y + 2\pi i x) - g(y) \le -C_3 y^{-\varepsilon}\) holds for all reals \(x\), \(y\) with \(0 < y < y_0\) and \(y^{1 + \varepsilon/2} \le |x| \le 1/2\).
public def MeinardusHypotheses (b : ℕ → ℝ) (α A : ℝ) (G : ℂ → ℂ) : Prop :=
∃ C₀ : ℝ, 0 < C₀ ∧ C₀ < 1 ∧
(∀ s : ℂ, α < s.re →
LSeriesSummable (fun k => (b k : ℂ)) s ∧ LSeries (fun k => (b k : ℂ)) s = G s) ∧
(∃ R : ℂ → ℂ, DifferentiableOn ℂ R {s | -C₀ ≤ s.re} ∧
∀ s : ℂ, -C₀ ≤ s.re → s ≠ α → G s = A / (s - α) + R s) ∧
(∃ c : ℝ, 0 < c ∧ ∃ K T : ℝ, ∀ s : ℂ, -C₀ ≤ s.re → T ≤ |s.im| →
‖G s‖ ≤ K * (1 + |s.im|) ^ c) ∧
(∃ ε C₃ y₀ : ℝ, 0 < ε ∧ 0 < C₃ ∧ 0 < y₀ ∧ ∀ y x : ℝ, 0 < y → y < y₀ →
y ^ (1 + ε / 2) ≤ |x| → |x| ≤ 1 / 2 →
(∑' k : ℕ, (b (k + 1) : ℂ) *
Complex.exp (-((k : ℂ) + 1) * ((y : ℂ) + 2 * Real.pi * Complex.I * x))).re -
∑' k : ℕ, b (k + 1) * Real.exp (-((k : ℝ) + 1) * y) ≤ -C₃ * y ^ (-ε))
Sources
- Hartosh Singh Bal, Completely Additive Height Functions: Profile Laws, Matula Bounds, and Inverse Growth, INTEGERS 26 (2026), Article A116, Section 6.1 (hypotheses of Theorem 3).
TheoremMeinardus' theorem#●ready to prove
Let \(b_0, b_1, b_2, \dots\) be nonnegative reals with \(b_k \ne 0\) for some \(k \ge 1\), and let \(a_n\) be the coefficients of the weighted partition product \(\prod_{k \ge 1} (1 - q^k)^{-b_k}\). Suppose that the weights satisfy Meinardus' hypotheses (M1)–(M3) with exponent \(\alpha > 0\), residue \(A > 0\) and continuation \(G\), and put \(\kappa = (\operatorname{Re} G(0) - 1 - \alpha/2)/(\alpha + 1)\) and \(C_2 = (1 + 1/\alpha)\,(A\,\Gamma(\alpha + 1)\,\zeta(\alpha + 1))^{1/(\alpha + 1)}\). Then there is \(C_1 > 0\) with \(a_n \sim C_1 n^{\kappa} \exp(C_2 n^{\alpha/(\alpha + 1)})\) as \(n \to \infty\), and \(\log a_n \sim C_2 n^{\alpha/(\alpha + 1)}\) as \(n \to \infty\).
public theorem_wanted meinardus_asymptotic {b : ℕ → ℝ} (hb : ∀ k, 0 ≤ b k)
(hb0 : ∃ k, 1 ≤ k ∧ b k ≠ 0) {α A : ℝ} (hα : 0 < α) (hA : 0 < A) {G : ℂ → ℂ}
(hM : MeinardusHypotheses b α A G) :
let κ := ((G 0).re - 1 - α / 2) / (α + 1)
let C₂ := (1 + 1 / α) *
(A * Real.Gamma (α + 1) * (riemannZeta (α + 1)).re) ^ (1 / (α + 1))
(∃ C₁ : ℝ, 0 < C₁ ∧ Asymptotics.IsEquivalent atTop (fun n : ℕ => weightedPartitionCoeff b n)
(fun n : ℕ => C₁ * (n : ℝ) ^ κ * Real.exp (C₂ * (n : ℝ) ^ (α / (α + 1))))) ∧
Asymptotics.IsEquivalent atTop (fun n : ℕ => Real.log (weightedPartitionCoeff b n))
(fun n : ℕ => C₂ * (n : ℝ) ^ (α / (α + 1)))
public def MeinardusHypotheses (b : ℕ → ℝ) (α A : ℝ) (G : ℂ → ℂ) : Prop :=
∃ C₀ : ℝ, 0 < C₀ ∧ C₀ < 1 ∧
(∀ s : ℂ, α < s.re →
LSeriesSummable (fun k => (b k : ℂ)) s ∧ LSeries (fun k => (b k : ℂ)) s = G s) ∧
(∃ R : ℂ → ℂ, DifferentiableOn ℂ R {s | -C₀ ≤ s.re} ∧
∀ s : ℂ, -C₀ ≤ s.re → s ≠ α → G s = A / (s - α) + R s) ∧
(∃ c : ℝ, 0 < c ∧ ∃ K T : ℝ, ∀ s : ℂ, -C₀ ≤ s.re → T ≤ |s.im| →
‖G s‖ ≤ K * (1 + |s.im|) ^ c) ∧
(∃ ε C₃ y₀ : ℝ, 0 < ε ∧ 0 < C₃ ∧ 0 < y₀ ∧ ∀ y x : ℝ, 0 < y → y < y₀ →
y ^ (1 + ε / 2) ≤ |x| → |x| ≤ 1 / 2 →
(∑' k : ℕ, (b (k + 1) : ℂ) *
Complex.exp (-((k : ℂ) + 1) * ((y : ℂ) + 2 * Real.pi * Complex.I * x))).re -
⋯
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:322
public noncomputable def weightedPartitionCoeff (b : ℕ → ℝ) : ℕ → ℝ
| 0 => 1
| n + 1 => (∑ m : Fin (n + 1), (∑ k ∈ (m.1 + 1).divisors, (k : ℝ) * b k) *
weightedPartitionCoeff b (n - m.1)) / (n + 1)
decreasing_by omega
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:291
Sources
- Hartosh Singh Bal, Completely Additive Height Functions: Profile Laws, Matula Bounds, and Inverse Growth, INTEGERS 26 (2026), Article A116, Theorem 3, quoting Meinardus.
Uses 2
TheoremA simple pole of a Dirichlet series from the growth of its partial sums#●ready to prove
Let \((\pi_k)_{k \ge 0}\) be nonnegative real numbers and let \(C\), \(\alpha\), \(\eta\) be positive reals. Suppose there is a constant \(K\) with \(\bigl|\sum_{k=1}^{\lfloor x \rfloor} \pi_k - (C/\alpha) x^{\alpha}\bigr| \le K x^{\alpha - \eta}\) for every real \(x \ge 1\). Then the Dirichlet series \(B(s) = \sum_{k \ge 1} \pi_k k^{-s}\) converges absolutely for \(\operatorname{Re} s > \alpha\), and there is a function \(R\), holomorphic on the half-plane \(\operatorname{Re} s > \alpha - \eta\), with \(B(s) = C/(s - \alpha) + R(s)\) for \(\operatorname{Re} s > \alpha\). Moreover, for every \(\delta\) with \(0 < \delta < \eta\) there is a constant \(K'\) with \(|C/(s - \alpha) + R(s)| \le K'(1 + |\operatorname{Im} s|)\) for all complex \(s\) with \(\operatorname{Re} s \ge \alpha - \eta + \delta\) and \(|\operatorname{Im} s| \ge 1\).
public theorem_wanted lseries_profile_simple_pole {π : ℕ → ℝ} (hπ : ∀ k, 0 ≤ π k) {C α η : ℝ}
(hC : 0 < C) (hα : 0 < α) (hη : 0 < η)
(hcum : ∃ K : ℝ, ∀ x : ℝ, 1 ≤ x →
|∑ k ∈ Finset.Icc 1 ⌊x⌋₊, π k - C / α * x ^ α| ≤ K * x ^ (α - η)) :
(∀ s : ℂ, α < s.re → LSeriesSummable (fun k => (π k : ℂ)) s) ∧
∃ R : ℂ → ℂ, DifferentiableOn ℂ R {s | α - η < s.re} ∧
(∀ s : ℂ, α < s.re → LSeries (fun k => (π k : ℂ)) s = C / (s - α) + R s) ∧
∀ δ : ℝ, 0 < δ → δ < η → ∃ K : ℝ, ∀ s : ℂ, α - η + δ ≤ s.re → 1 ≤ |s.im| →
‖C / (s - α) + R s‖ ≤ K * (1 + |s.im|)
Sources
- Hartosh Singh Bal, Completely Additive Height Functions: Profile Laws, Matula Bounds, and Inverse Growth, INTEGERS 26 (2026), Article A116, Lemma 3.
TheoremMeinardus asymptotics for the levels of a height function#●ready to prove
Let \(H\) be a height function, for \(k \ge 0\) let \(\pi_k\) be the number of primes \(p\) with \(H(p) = k\), and for \(n \ge 0\) let \(N_n\) be the number of integers \(m \ge 1\) with \(H(m) = n\). Let \(\alpha > 0\) and \(C > 0\) be real numbers and suppose that the weights \((\pi_k)\) satisfy Meinardus' hypotheses with exponent \(\alpha\), residue \(C\) and some continuation \(G\). Put \(C_2 = (1 + 1/\alpha)\bigl(C\,\Gamma(\alpha + 1)\,\zeta(\alpha + 1)\bigr)^{1/(\alpha+1)}\). Then there are real numbers \(C_1 > 0\) and \(\kappa\) with \(N_n \sim C_1 n^{\kappa} \exp\bigl(C_2 n^{\alpha/(\alpha+1)}\bigr)\) as \(n \to \infty\), and \(\log N_n \sim C_2 n^{\alpha/(\alpha+1)}\) as \(n \to \infty\).
public theorem_wanted heightLevel_card_meinardus {H : ℕ → ℕ} (hH : IsHeightFunction H)
{α C : ℝ} (hα : 0 < α) (hC : 0 < C) {G : ℂ → ℂ}
(hM : MeinardusHypotheses (fun k => ({p : ℕ | p.Prime ∧ H p = k}.ncard : ℝ)) α C G) :
let C₂ := (1 + 1 / α) *
(C * Real.Gamma (α + 1) * (riemannZeta (α + 1)).re) ^ (1 / (α + 1))
(∃ C₁ κ : ℝ, 0 < C₁ ∧ Asymptotics.IsEquivalent atTop
(fun n : ℕ => (Nat.card {m : ℕ // 0 < m ∧ H m = n} : ℝ))
(fun n : ℕ => C₁ * (n : ℝ) ^ κ * Real.exp (C₂ * (n : ℝ) ^ (α / (α + 1))))) ∧
Asymptotics.IsEquivalent atTop
(fun n : ℕ => Real.log (Nat.card {m : ℕ // 0 < m ∧ H m = n} : ℝ))
(fun n : ℕ => C₂ * (n : ℝ) ^ (α / (α + 1)))
public structure IsHeightFunction (H : ℕ → ℕ) : Prop where
additive : IsCompletelyAdditiveNat H
prime_pos : ∀ p : ℕ, p.Prime → 1 ≤ H p
finite_primes : ∀ n : ℕ, 1 ≤ n → {p : ℕ | p.Prime ∧ H p = n}.Finite
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:62
public def MeinardusHypotheses (b : ℕ → ℝ) (α A : ℝ) (G : ℂ → ℂ) : Prop :=
∃ C₀ : ℝ, 0 < C₀ ∧ C₀ < 1 ∧
(∀ s : ℂ, α < s.re →
LSeriesSummable (fun k => (b k : ℂ)) s ∧ LSeries (fun k => (b k : ℂ)) s = G s) ∧
(∃ R : ℂ → ℂ, DifferentiableOn ℂ R {s | -C₀ ≤ s.re} ∧
∀ s : ℂ, -C₀ ≤ s.re → s ≠ α → G s = A / (s - α) + R s) ∧
(∃ c : ℝ, 0 < c ∧ ∃ K T : ℝ, ∀ s : ℂ, -C₀ ≤ s.re → T ≤ |s.im| →
‖G s‖ ≤ K * (1 + |s.im|) ^ c) ∧
(∃ ε C₃ y₀ : ℝ, 0 < ε ∧ 0 < C₃ ∧ 0 < y₀ ∧ ∀ y x : ℝ, 0 < y → y < y₀ →
y ^ (1 + ε / 2) ≤ |x| → |x| ≤ 1 / 2 →
(∑' k : ℕ, (b (k + 1) : ℂ) *
Complex.exp (-((k : ℂ) + 1) * ((y : ℂ) + 2 * Real.pi * Complex.I * x))).re -
⋯
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:322
Sources
- Hartosh Singh Bal, Completely Additive Height Functions: Profile Laws, Matula Bounds, and Inverse Growth, INTEGERS 26 (2026), Article A116, Theorem 4.
Uses 2
TheoremIdentifying the profile from the growth of the levels#●ready to prove
Let \(H\) be a height function, for \(k \ge 0\) let \(\pi_k\) be the number of primes \(p\) with \(H(p) = k\), and for \(n \ge 0\) let \(N_n\) be the number of integers \(m \ge 1\) with \(H(m) = n\). Let \(r > 0\) and \(A > 0\) be real numbers and suppose that the weights \((\pi_k)\) satisfy Meinardus' hypotheses with exponent \(r\), residue \(A\) and some continuation \(G\). Suppose further that \(N_n \sim C_1 n^{\kappa} \exp(C_2 n^{\beta})\) as \(n \to \infty\) for real numbers \(C_1 > 0\), \(C_2 > 0\), \(\kappa\) and \(\beta\) with \(0 < \beta < 1\). Then \(\beta = r/(r + 1)\), \(r = \beta/(1 - \beta)\), \(C_2 = (1 + 1/r)\bigl(A\,\Gamma(r + 1)\,\zeta(r + 1)\bigr)^{1/(r+1)}\), and \(\sum_{k=1}^{\lfloor x \rfloor} \pi_k \sim (A/r)\, x^r\) as the real number \(x\) tends to \(\infty\).
public theorem_wanted heightProfile_identification {H : ℕ → ℕ} (hH : IsHeightFunction H)
{r A : ℝ} (hr : 0 < r) (hA : 0 < A) {G : ℂ → ℂ}
(hM : MeinardusHypotheses (fun k => ({p : ℕ | p.Prime ∧ H p = k}.ncard : ℝ)) r A G)
{C₁ C₂ κ β : ℝ} (hC₁ : 0 < C₁) (hC₂ : 0 < C₂) (hβ0 : 0 < β) (hβ1 : β < 1)
(hN : Asymptotics.IsEquivalent atTop
(fun n : ℕ => (Nat.card {m : ℕ // 0 < m ∧ H m = n} : ℝ))
(fun n : ℕ => C₁ * (n : ℝ) ^ κ * Real.exp (C₂ * (n : ℝ) ^ β))) :
β = r / (r + 1) ∧ r = β / (1 - β) ∧
C₂ = (1 + 1 / r) * (A * Real.Gamma (r + 1) * (riemannZeta (r + 1)).re) ^ (1 / (r + 1)) ∧
Asymptotics.IsEquivalent atTop
(fun x : ℝ => (∑ k ∈ Finset.Icc 1 ⌊x⌋₊, {p : ℕ | p.Prime ∧ H p = k}.ncard : ℝ))
(fun x : ℝ => A / r * x ^ r)
public structure IsHeightFunction (H : ℕ → ℕ) : Prop where
additive : IsCompletelyAdditiveNat H
prime_pos : ∀ p : ℕ, p.Prime → 1 ≤ H p
finite_primes : ∀ n : ℕ, 1 ≤ n → {p : ℕ | p.Prime ∧ H p = n}.Finite
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:62
public def MeinardusHypotheses (b : ℕ → ℝ) (α A : ℝ) (G : ℂ → ℂ) : Prop :=
∃ C₀ : ℝ, 0 < C₀ ∧ C₀ < 1 ∧
(∀ s : ℂ, α < s.re →
LSeriesSummable (fun k => (b k : ℂ)) s ∧ LSeries (fun k => (b k : ℂ)) s = G s) ∧
(∃ R : ℂ → ℂ, DifferentiableOn ℂ R {s | -C₀ ≤ s.re} ∧
∀ s : ℂ, -C₀ ≤ s.re → s ≠ α → G s = A / (s - α) + R s) ∧
(∃ c : ℝ, 0 < c ∧ ∃ K T : ℝ, ∀ s : ℂ, -C₀ ≤ s.re → T ≤ |s.im| →
‖G s‖ ≤ K * (1 + |s.im|) ^ c) ∧
(∃ ε C₃ y₀ : ℝ, 0 < ε ∧ 0 < C₃ ∧ 0 < y₀ ∧ ∀ y x : ℝ, 0 < y → y < y₀ →
y ^ (1 + ε / 2) ≤ |x| → |x| ≤ 1 / 2 →
(∑' k : ℕ, (b (k + 1) : ℂ) *
Complex.exp (-((k : ℂ) + 1) * ((y : ℂ) + 2 * Real.pi * Complex.I * x))).re -
⋯
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:322
Sources
- Hartosh Singh Bal, Completely Additive Height Functions: Profile Laws, Matula Bounds, and Inverse Growth, INTEGERS 26 (2026), Article A116, Corollary 1.
Uses 2
DefinitionCumulative profile#✓fully proved
For a sequence \(\pi : \mathbb{N} \to \mathbb{N}\) and a natural number \(k\), the cumulative profile is \(\Pi_k = \sum_{j=1}^{k} \pi_j\). The value \(\pi_0\) is not used, and \(\Pi_0 = 0\).
public def profilePartialSum (π : ℕ → ℕ) (k : ℕ) : ℕ :=
∑ j ∈ Finset.Icc 1 k, π j
Sources
- Hartosh Singh Bal, Completely Additive Height Functions: Profile Laws, Matula Bounds, and Inverse Growth, INTEGERS 26 (2026), Article A116, Section 7.2 (canonical sequential realization).
TheoremRecursion for the cumulative profile#✓fully proved
For every sequence \(\pi : \mathbb{N} \to \mathbb{N}\) with cumulative profile \(\Pi_k = \sum_{j=1}^{k} \pi_j\) and every natural number \(k\), \(\Pi_{k+1} = \Pi_k + \pi_{k+1}\).
public theorem profilePartialSum_succ (π : ℕ → ℕ) (k : ℕ) :
profilePartialSum π (k + 1) = profilePartialSum π k + π (k + 1)
public def profilePartialSum (π : ℕ → ℕ) (k : ℕ) : ℕ :=
∑ j ∈ Finset.Icc 1 k, π j
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:420
Sources
- Hartosh Singh Bal, Completely Additive Height Functions: Profile Laws, Matula Bounds, and Inverse Growth, INTEGERS 26 (2026), Article A116, Section 7.2 (canonical sequential realization).
Uses 1
DefinitionBlock index of a profile#✓fully proved
Let \(\pi : \mathbb{N} \to \mathbb{N}\) be a sequence with cumulative profile \(\Pi_k = \sum_{j=1}^{k} \pi_j\). For a natural number \(i\), the block index \(k(i)\) is the least natural number \(k\) with \(i \le \Pi_k\), and \(k(i) = 0\) if there is no such \(k\). When \(i \ge 1\) and \(\Pi\) is unbounded, \(k(i)\) is the unique \(k\) with \(\Pi_{k-1} < i \le \Pi_k\).
public noncomputable def seqHeightIndex (π : ℕ → ℕ) (i : ℕ) : ℕ :=
sInf {k | i ≤ profilePartialSum π k}
public def profilePartialSum (π : ℕ → ℕ) (k : ℕ) : ℕ :=
∑ j ∈ Finset.Icc 1 k, π j
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:420
Sources
- Hartosh Singh Bal, Completely Additive Height Functions: Profile Laws, Matula Bounds, and Inverse Growth, INTEGERS 26 (2026), Article A116, Section 7.2 (canonical sequential realization).
Uses 1 · Used by 3
TheoremBlock index of the constant profile#✓fully proved
For the constant profile \(\pi_k = 1\), whose cumulative profile is \(\Pi_k = \sum_{j=1}^{k} \pi_j = k\), the block index \(k(i)\), the least \(k\) with \(i \le \Pi_k\), equals \(i\) for every natural number \(i\).
public theorem seqHeightIndex_const_one (i : ℕ) : seqHeightIndex (fun _ => 1) i = i
public noncomputable def seqHeightIndex (π : ℕ → ℕ) (i : ℕ) : ℕ :=
sInf {k | i ≤ profilePartialSum π k}
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:438
Sources
- Hartosh Singh Bal, Completely Additive Height Functions: Profile Laws, Matula Bounds, and Inverse Growth, INTEGERS 26 (2026), Article A116, Section 7.2 (canonical sequential realization).
Uses 2
DefinitionCanonical sequential realization of a profile#✓fully proved
Let \((\pi_k)_{k \ge 1}\) be a sequence of nonnegative integers with partial sums \(\Pi_k = \sum_{1 \le j \le k} \pi_j\). For \(i \ge 1\) let \(k(i)\) be the least \(k \ge 0\) with \(i \le \Pi_k\), or \(0\) if there is none, and let \(p_1 = 2, p_2 = 3, \dots\) be the primes in increasing order. The canonical sequential realization \(H_{\mathrm{seq}}\) of \(\pi\) is the function \(\mathbb{N} \to \mathbb{N}\) that sends \(m \ge 1\) to \(\sum_{i \ge 1} a_i\, k(i)\), where \(a_i\) is the exponent of \(p_i\) in \(m\), and sends \(0\) to \(0\); thus \(H_{\mathrm{seq}}\) is completely additive with \(H_{\mathrm{seq}}(p_i) = k(i)\).
public noncomputable def canonicalSeqHeight (π : ℕ → ℕ) : ℕ → ℕ :=
additiveOfPrimeValues fun p => seqHeightIndex π (Nat.count Nat.Prime p + 1)
public noncomputable def additiveOfPrimeValues (v : ℕ → ℕ) (m : ℕ) : ℕ :=
m.factorization.sum fun p a => a * v p
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:88
public noncomputable def seqHeightIndex (π : ℕ → ℕ) (i : ℕ) : ℕ :=
sInf {k | i ≤ profilePartialSum π k}
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:438
Sources
- Hartosh Singh Bal, Completely Additive Height Functions: Profile Laws, Matula Bounds, and Inverse Growth, INTEGERS 26 (2026), Article A116, Definition 2 (canonical sequential realization, Section 7.2).
Uses 2 · Used by 5
TheoremInverting a polynomial profile#●ready to prove
Let \(\pi : \mathbb{N} \to \mathbb{N}\) be a sequence with cumulative profile \(\Pi_k = \sum_{j=1}^{k} \pi_j\), let \(k(i)\) be its block index (the least \(k\) with \(i \le \Pi_k\), and \(0\) if there is none), let \(H_{\mathrm{seq}}\) be its canonical sequential realization, and write \(p_i\) for the \(i\)-th prime, indexed from one. Let \(C > 0\) and \(\alpha > 0\) be real numbers with \(\Pi_k \sim C k^\alpha\) as \(k \to \infty\). Then \(k(i) \sim (i/C)^{1/\alpha}\) as \(i \to \infty\), and \(H_{\mathrm{seq}}(p_i) = k(i)\) for every \(i \ge 1\).
The paper also assumes that \(\Pi_k\) is eventually strictly increasing; that hypothesis is dropped here, since \(\Pi\) is nondecreasing, which suffices.
public theorem_wanted seqHeightIndex_isEquivalent {π : ℕ → ℕ} {C α : ℝ} (hC : 0 < C)
(hα : 0 < α)
(hcum : Asymptotics.IsEquivalent atTop (fun k : ℕ => (profilePartialSum π k : ℝ))
(fun k : ℕ => C * (k : ℝ) ^ α)) :
Asymptotics.IsEquivalent atTop (fun i : ℕ => (seqHeightIndex π i : ℝ))
(fun i : ℕ => ((i : ℝ) / C) ^ (1 / α)) ∧
∀ i, 1 ≤ i → canonicalSeqHeight π (nthPrimeOne i) = seqHeightIndex π i
public def profilePartialSum (π : ℕ → ℕ) (k : ℕ) : ℕ :=
∑ j ∈ Finset.Icc 1 k, π j
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:420
public noncomputable def seqHeightIndex (π : ℕ → ℕ) (i : ℕ) : ℕ :=
sInf {k | i ≤ profilePartialSum π k}
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:438
public noncomputable def canonicalSeqHeight (π : ℕ → ℕ) : ℕ → ℕ :=
additiveOfPrimeValues fun p => seqHeightIndex π (Nat.count Nat.Prime p + 1)
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:458
public noncomputable def nthPrimeOne (i : ℕ) : ℕ :=
Nat.nth Nat.Prime (i - 1)
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:71
Sources
- Hartosh Singh Bal, Completely Additive Height Functions: Profile Laws, Matula Bounds, and Inverse Growth, INTEGERS 26 (2026), Article A116, Lemma 4.
Uses 4
TheoremA logarithmic-integral sum#●ready to prove
For every real \(\beta > 0\),
as \(m \to \infty\) through the natural numbers.
public theorem_wanted sum_rpow_div_log_isEquivalent {β : ℝ} (hβ : 0 < β) :
Asymptotics.IsEquivalent atTop
(fun m : ℕ => ∑ i ∈ Finset.Icc 2 m, (i : ℝ) ^ (β - 1) / Real.log i)
(fun m : ℕ => (m : ℝ) ^ β / (β * Real.log m))
Sources
- Hartosh Singh Bal, Completely Additive Height Functions: Profile Laws, Matula Bounds, and Inverse Growth, INTEGERS 26 (2026), Article A116, Lemma 5.
TheoremPrime-harmonic growth of the canonical sequential realization#●ready to prove
Let \((\pi_k)_{k \ge 1}\) be a sequence of nonnegative integers with partial sums \(\Pi_k = \sum_{1 \le j \le k} \pi_j\), let \(H_{\mathrm{seq}}\) be its canonical sequential realization, and suppose that \(C > 0\) and \(\alpha > 0\) are real numbers with \(\Pi_k \sim C k^\alpha\) as \(k \to \infty\). Write \(S(x) = \sum_{p \le x} H_{\mathrm{seq}}(p)/p\), the sum over primes \(p \le x\), and let \(m(x)\) be the number of primes \(p \le x\). Then, as \(x \to \infty\) over the reals, \(S(x) \sim \frac{\alpha}{C^{1/\alpha}} \cdot \frac{m(x)^{1/\alpha}}{\log m(x)}\) and \(S(x) = \Theta\bigl(x^{1/\alpha} / (\log x)^{1+1/\alpha}\bigr)\), where \(f = \Theta(g)\) means \(f = O(g)\) and \(g = O(f)\).
public theorem_wanted canonicalSeq_primeHarmonicSum {π : ℕ → ℕ} {C α : ℝ} (hC : 0 < C)
(hα : 0 < α)
(hcum : Asymptotics.IsEquivalent atTop (fun k : ℕ => (profilePartialSum π k : ℝ))
(fun k : ℕ => C * (k : ℝ) ^ α)) :
Asymptotics.IsEquivalent atTop
(fun x : ℝ => ∑ p ∈ (Finset.Iic ⌊x⌋₊).filter Nat.Prime,
(canonicalSeqHeight π p : ℝ) / p)
(fun x : ℝ => α / C ^ (1 / α) *
((Nat.primeCounting ⌊x⌋₊ : ℝ) ^ (1 / α) / Real.log (Nat.primeCounting ⌊x⌋₊))) ∧
(fun x : ℝ => ∑ p ∈ (Finset.Iic ⌊x⌋₊).filter Nat.Prime,
(canonicalSeqHeight π p : ℝ) / p) =Θ[atTop]
(fun x : ℝ => x ^ (1 / α) / Real.log x ^ (1 + 1 / α))
public def profilePartialSum (π : ℕ → ℕ) (k : ℕ) : ℕ :=
∑ j ∈ Finset.Icc 1 k, π j
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:420
public noncomputable def canonicalSeqHeight (π : ℕ → ℕ) : ℕ → ℕ :=
additiveOfPrimeValues fun p => seqHeightIndex π (Nat.count Nat.Prime p + 1)
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:458
Sources
- Hartosh Singh Bal, Completely Additive Height Functions: Profile Laws, Matula Bounds, and Inverse Growth, INTEGERS 26 (2026), Article A116, Theorem 5.
Uses 2
TheoremAverage order of the canonical sequential realization#●ready to prove
Let \((\pi_k)_{k \ge 1}\) be a sequence of nonnegative integers with partial sums \(\Pi_k = \sum_{1 \le j \le k} \pi_j\), let \(H_{\mathrm{seq}}\) be its canonical sequential realization, and suppose that \(C > 0\) and \(\alpha > 0\) are real numbers with \(\Pi_k \sim C k^\alpha\) as \(k \to \infty\). Then, as \(x \to \infty\) over the reals, each of the three functions \(x \sum_{p \le x} H_{\mathrm{seq}}(p)/p\), \(\sum_{p \le x} H_{\mathrm{seq}}(p) \log x / \log p\) (both sums over primes \(p \le x\)) and \(\sum_{1 \le n \le x} H_{\mathrm{seq}}(n)\) is \(\Theta\bigl(x^{1+1/\alpha} / (\log x)^{1+1/\alpha}\bigr)\), where \(f = \Theta(g)\) means \(f = O(g)\) and \(g = O(f)\).
public theorem_wanted canonicalSeq_averageOrder {π : ℕ → ℕ} {C α : ℝ} (hC : 0 < C)
(hα : 0 < α)
(hcum : Asymptotics.IsEquivalent atTop (fun k : ℕ => (profilePartialSum π k : ℝ))
(fun k : ℕ => C * (k : ℝ) ^ α)) :
(fun x : ℝ => x * ∑ p ∈ (Finset.Iic ⌊x⌋₊).filter Nat.Prime,
(canonicalSeqHeight π p : ℝ) / p) =Θ[atTop]
(fun x : ℝ => x ^ (1 + 1 / α) / Real.log x ^ (1 + 1 / α)) ∧
(fun x : ℝ => ∑ p ∈ (Finset.Iic ⌊x⌋₊).filter Nat.Prime,
(canonicalSeqHeight π p : ℝ) * (Real.log x / Real.log p)) =Θ[atTop]
(fun x : ℝ => x ^ (1 + 1 / α) / Real.log x ^ (1 + 1 / α)) ∧
(fun x : ℝ => ∑ n ∈ Finset.Icc 1 ⌊x⌋₊, (canonicalSeqHeight π n : ℝ)) =Θ[atTop]
(fun x : ℝ => x ^ (1 + 1 / α) / Real.log x ^ (1 + 1 / α))
public def profilePartialSum (π : ℕ → ℕ) (k : ℕ) : ℕ :=
∑ j ∈ Finset.Icc 1 k, π j
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:420
public noncomputable def canonicalSeqHeight (π : ℕ → ℕ) : ℕ → ℕ :=
additiveOfPrimeValues fun p => seqHeightIndex π (Nat.count Nat.Prime p + 1)
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:458
Sources
- Hartosh Singh Bal, Completely Additive Height Functions: Profile Laws, Matula Bounds, and Inverse Growth, INTEGERS 26 (2026), Article A116, Corollary 2, statement read from the journal PDF (the corpus record carries the prose of Section 7.6 that follows it).
Uses 2
TheoremGrowth for the ordinary partition profile#●ready to prove
Let \(H_{\mathrm{seq}}\) be the canonical sequential realization of the constant profile \(\pi_k = 1\) for all \(k \ge 1\) (so \(H_{\mathrm{seq}}(p_i) = i\) for the \(i\)-th prime \(p_i\)). Then, as \(x \to \infty\) over the reals, \(\sum_{p \le x} H_{\mathrm{seq}}(p)/p = \Theta\bigl(x / (\log x)^2\bigr)\), the sum running over primes \(p \le x\), and \(\sum_{1 \le n \le x} H_{\mathrm{seq}}(n) = \Theta\bigl(x^2 / (\log x)^2\bigr)\), where \(f = \Theta(g)\) means \(f = O(g)\) and \(g = O(f)\).
public theorem_wanted canonicalSeq_partitionProfile_growth :
(fun x : ℝ => ∑ p ∈ (Finset.Iic ⌊x⌋₊).filter Nat.Prime,
(canonicalSeqHeight (fun _ => 1) p : ℝ) / p) =Θ[atTop]
(fun x : ℝ => x / Real.log x ^ 2) ∧
(fun x : ℝ => ∑ n ∈ Finset.Icc 1 ⌊x⌋₊, (canonicalSeqHeight (fun _ => 1) n : ℝ)) =Θ[atTop]
(fun x : ℝ => x ^ 2 / Real.log x ^ 2)
public noncomputable def canonicalSeqHeight (π : ℕ → ℕ) : ℕ → ℕ :=
additiveOfPrimeValues fun p => seqHeightIndex π (Nat.count Nat.Prime p + 1)
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:458
Sources
- Hartosh Singh Bal, Completely Additive Height Functions: Profile Laws, Matula Bounds, and Inverse Growth, INTEGERS 26 (2026), Article A116, Corollary 3.
Uses 1
TheoremGrowth for the plane partition profile#●ready to prove
Let \(H_{\mathrm{seq}}\) be the canonical sequential realization of the profile \(\pi_k = k\) for \(k \ge 1\), let \(S(x) = \sum_{p \le x} H_{\mathrm{seq}}(p)/p\) (the sum over primes \(p \le x\)) and \(F(x) = \sum_{1 \le n \le x} H_{\mathrm{seq}}(n)\), and let \(m(x)\) be the number of primes \(p \le x\). Then, as \(x \to \infty\) over the reals, \(S(x) \sim 2\sqrt{2}\, \sqrt{m(x)} / \log m(x)\), \(S(x) = \Theta\bigl(x^{1/2} / (\log x)^{3/2}\bigr)\), \(F(x) = \Theta\bigl(x\, S(x)\bigr)\) and \(F(x) = \Theta\bigl(x^{3/2} / (\log x)^{3/2}\bigr)\), where \(f = \Theta(g)\) means \(f = O(g)\) and \(g = O(f)\).
Here the partial sums are \(k(k+1)/2 \sim k^2/2\), so this is the case \(\alpha = 2\), \(C = 1/2\) of the general growth laws.
public theorem_wanted canonicalSeq_planePartitionProfile_growth :
Asymptotics.IsEquivalent atTop
(fun x : ℝ => ∑ p ∈ (Finset.Iic ⌊x⌋₊).filter Nat.Prime,
(canonicalSeqHeight (fun k => k) p : ℝ) / p)
(fun x : ℝ => 2 * Real.sqrt 2 *
(Real.sqrt (Nat.primeCounting ⌊x⌋₊) / Real.log (Nat.primeCounting ⌊x⌋₊))) ∧
(fun x : ℝ => ∑ p ∈ (Finset.Iic ⌊x⌋₊).filter Nat.Prime,
(canonicalSeqHeight (fun k => k) p : ℝ) / p) =Θ[atTop]
(fun x : ℝ => x ^ (1 / 2 : ℝ) / Real.log x ^ (3 / 2 : ℝ)) ∧
(fun x : ℝ => ∑ n ∈ Finset.Icc 1 ⌊x⌋₊, (canonicalSeqHeight (fun k => k) n : ℝ)) =Θ[atTop]
(fun x : ℝ => x * ∑ p ∈ (Finset.Iic ⌊x⌋₊).filter Nat.Prime,
(canonicalSeqHeight (fun k => k) p : ℝ) / p) ∧
(fun x : ℝ => ∑ n ∈ Finset.Icc 1 ⌊x⌋₊, (canonicalSeqHeight (fun k => k) n : ℝ)) =Θ[atTop]
(fun x : ℝ => x ^ (3 / 2 : ℝ) / Real.log x ^ (3 / 2 : ℝ))
public noncomputable def canonicalSeqHeight (π : ℕ → ℕ) : ℕ → ℕ :=
additiveOfPrimeValues fun p => seqHeightIndex π (Nat.count Nat.Prime p + 1)
WantedExt/NumberTheory/CompletelyAdditiveHeightWanted.lean:458
Sources
- Hartosh Singh Bal, Completely Additive Height Functions: Profile Laws, Matula Bounds, and Inverse Growth, INTEGERS 26 (2026), Article A116, Corollary 4.
Uses 1
Loading dependency explorer…
The dependency explorer needs JavaScript and its graph data.