Almost primes of almost prime index
DefinitionAlmost primes of almost prime index#✓fully proved
For natural numbers \(k_1, \dots, k_m\) with \(m \ge 1\), the set \(N_{k_1 \cdots k_m}\) of natural numbers is defined by recursion on \(m\): \(N_{k_1}\) is the set of \(k_1\)-almost primes (the \(n \ne 0\) with exactly \(k_1\) prime factors counted with multiplicity), and for \(m \ge 2\), \(N_{k_1 \cdots k_m}\) is the set of the \(i\)th \(k_1\)-almost primes, in increasing order and counted from \(1\), for \(i \ge 1\) in \(N_{k_2 \cdots k_m}\). When there are fewer than \(i\) \(k_1\)-almost primes (only for \(k_1 = 0\), where \(1\) is the only one), the \(i\)th one is taken to be \(0\).
public def almostPrimeIndexSet (k : ℕ) : List ℕ → Set ℕ
| [] => {n | Nat.IsAlmostPrime k n}
| j :: L => {n | ∃ i ∈ almostPrimeIndexSet j L, 1 ≤ i ∧ n = Nat.nth (Nat.IsAlmostPrime k) (i - 1)}
Sources
- Paul Kinlaw, Megan Triplett, and William Tripp, Almost Primes of Almost Prime Index, INTEGERS 24 (2024), Article A99, Section 1 (definitions of
N_k,N_{j,k}, andN_{k_1 ⋯ k_m}).
Used by 15
TheoremAlmost prime index sets with one index#✓fully proved
For every natural number \(k\), the almost prime index set \(N_k\) with the single index \(k\) is the set of \(k\)-almost primes, the natural numbers \(n \ne 0\) with exactly \(k\) prime factors counted with multiplicity.
public theorem almostPrimeIndexSet_nil (k : ℕ) :
almostPrimeIndexSet k [] = {n | Nat.IsAlmostPrime k n}
public def almostPrimeIndexSet (k : ℕ) : List ℕ → Set ℕ
| [] => {n | Nat.IsAlmostPrime k n}
| j :: L => {n | ∃ i ∈ almostPrimeIndexSet j L, 1 ≤ i ∧ n = Nat.nth (Nat.IsAlmostPrime k) (i - 1)}
WantedExt/NumberTheory/AlmostPrimeIndexWanted.lean:60
Sources
- Paul Kinlaw, Megan Triplett, and William Tripp, Almost Primes of Almost Prime Index, INTEGERS 24 (2024), Article A99, Section 1 (definition of
N_k).
Uses 1
TheoremMembership in an almost prime index set#✓fully proved
For natural numbers \(k\), \(j\), \(n\) and \(\ell_1, \dots, \ell_r\) (with \(r \ge 0\)), \(n\) lies in the almost prime index set \(N_{k j \ell_1 \cdots \ell_r}\) if and only if there is an \(i \ge 1\) in \(N_{j \ell_1 \cdots \ell_r}\) such that \(n\) is the \(i\)th \(k\)-almost prime in increasing order, counted from \(1\) (taken to be \(0\) if there are fewer than \(i\) of them).
public theorem mem_almostPrimeIndexSet_cons (k j : ℕ) (L : List ℕ) (n : ℕ) :
n ∈ almostPrimeIndexSet k (j :: L) ↔
∃ i ∈ almostPrimeIndexSet j L, 1 ≤ i ∧ n = Nat.nth (Nat.IsAlmostPrime k) (i - 1)
public def almostPrimeIndexSet (k : ℕ) : List ℕ → Set ℕ
| [] => {n | Nat.IsAlmostPrime k n}
| j :: L => {n | ∃ i ∈ almostPrimeIndexSet j L, 1 ≤ i ∧ n = Nat.nth (Nat.IsAlmostPrime k) (i - 1)}
WantedExt/NumberTheory/AlmostPrimeIndexWanted.lean:60
Sources
- Paul Kinlaw, Megan Triplett, and William Tripp, Almost Primes of Almost Prime Index, INTEGERS 24 (2024), Article A99, Section 1 (definition of
N_{k_1 ⋯ k_m}).
Uses 1
TheoremDusart's bounds for the \(n\)th prime#●ready to prove
For \(n \ge 1\) let \(p_n\) denote the \(n\)th prime, so that \(p_1 = 2\), and let \(\log\) be the natural logarithm. For every natural number \(n \ge 2\),
and
and for every natural number \(n \ge 39017\),
public theorem_wanted nth_prime_dusart_bounds :
(∀ n : ℕ, 2 ≤ n →
(n : ℝ) * (Real.log n + Real.log (Real.log n) - 1) ≤ Nat.nth Nat.Prime (n - 1)) ∧
(∀ n : ℕ, 2 ≤ n →
(n : ℝ) * (Real.log n + Real.log (Real.log n) - 1 +
(Real.log (Real.log n) - 2.25) / Real.log n) ≤ Nat.nth Nat.Prime (n - 1)) ∧
(∀ n : ℕ, 39017 ≤ n →
(Nat.nth Nat.Prime (n - 1) : ℝ) ≤ n * (Real.log n + Real.log (Real.log n) - 0.9484))
Sources
- Paul Kinlaw, Megan Triplett, and William Tripp, Almost Primes of Almost Prime Index, INTEGERS 24 (2024), Article A99, Lemma 4 (quoting Dusart).
TheoremConvergence of reciprocal sums over almost primes of almost prime index#●ready to prove
A \(k\)-almost prime is a positive integer with exactly \(k\) prime factors counted with multiplicity. For positive integers \(k_1, \dots, k_m\) let \(N_{k_1}\) be the set of \(k_1\)-almost primes and, for \(m \ge 2\), let \(N_{k_1 k_2 \cdots k_m}\) be the set of \(k_1\)-almost primes whose index in the increasing enumeration of the \(k_1\)-almost primes (starting at \(1\)) lies in \(N_{k_2 \cdots k_m}\). Then for every \(m \ge 2\) and all positive integers \(k_1, \dots, k_m\) the series \(\sum_{n \in N_{k_1 k_2 \cdots k_m}} 1/n\) converges.
In particular (the case \(m = 2\)), for all positive integers \(j\) and \(k\) the sum of the reciprocals of the \(j\)-almost primes of \(k\)-almost prime index converges.
public theorem_wanted summable_inv_almostPrimeIndexSet {k : ℕ} {L : List ℕ} (hk : 1 ≤ k)
(hL : L ≠ []) (hpos : ∀ j ∈ L, 1 ≤ j) :
Summable fun n : almostPrimeIndexSet k L => (1 : ℝ) / n
public def almostPrimeIndexSet (k : ℕ) : List ℕ → Set ℕ
| [] => {n | Nat.IsAlmostPrime k n}
| j :: L => {n | ∃ i ∈ almostPrimeIndexSet j L, 1 ≤ i ∧ n = Nat.nth (Nat.IsAlmostPrime k) (i - 1)}
WantedExt/NumberTheory/AlmostPrimeIndexWanted.lean:60
Sources
- Paul Kinlaw, Megan Triplett, and William Tripp, Almost Primes of Almost Prime Index, INTEGERS 24 (2024), Article A99, Theorem 5.
Uses 1
TheoremReciprocal sum of the primes of semiprime index#●ready to prove
Let \(N_{1,2}\) be the set of primes \(p_i\) (indexed from \(p_1 = 2\)) whose index \(i\) is a semiprime, that is, a product of two not necessarily distinct primes. Then \(0.9910 < \sum_{n \in N_{1,2}} 1/n < 0.9915\).
public theorem_wanted tsum_inv_primes_of_semiprime_index_bounds :
0.9910 < ∑' n : almostPrimeIndexSet 1 [2], (1 : ℝ) / n ∧
∑' n : almostPrimeIndexSet 1 [2], (1 : ℝ) / n < 0.9915
public def almostPrimeIndexSet (k : ℕ) : List ℕ → Set ℕ
| [] => {n | Nat.IsAlmostPrime k n}
| j :: L => {n | ∃ i ∈ almostPrimeIndexSet j L, 1 ≤ i ∧ n = Nat.nth (Nat.IsAlmostPrime k) (i - 1)}
WantedExt/NumberTheory/AlmostPrimeIndexWanted.lean:60
Sources
- Paul Kinlaw, Megan Triplett, and William Tripp, Almost Primes of Almost Prime Index, INTEGERS 24 (2024), Article A99, Theorem 6.
Uses 1
TheoremCrişan–Erban expansion of the semiprime counting function#●ready to prove
A semiprime is a product of exactly two primes counted with multiplicity; for real \(x\) let \(N_2(x)\) be the number of semiprimes at most \(x\). Let \(\beta\) be the Meissel–Mertens constant, \(\gamma\) the Euler–Mascheroni constant and \(\log\) the natural logarithm. There are real constants \(c_0, c_1, c_2, \dots\) with \(c_0 = \beta\) and
the sum running over all primes \(p\), such that for every natural number \(N\), as \(x \to \infty\),
The paper calls the \(c_n\) explicit but displays only \(c_0\) and \(c_1\); the remaining constants are only asserted to exist.
public theorem_wanted semiprime_count_asymptotic_expansion :
∃ c : ℕ → ℝ, c 0 = meisselMertensConstant ∧
c 1 = meisselMertensConstant - 1 - Real.eulerMascheroniConstant -
∑' p : Nat.Primes, Real.log (p : ℕ) / (((p : ℕ) : ℝ) * (((p : ℕ) : ℝ) - 1)) ∧
∀ N : ℕ, (fun x : ℝ => (almostPrimeCounting 2 ⌊x⌋₊ : ℝ) -
x / Real.log x * ∑ n ∈ Finset.range N,
(n.factorial : ℝ) * (Real.log (Real.log x) + c n) / Real.log x ^ n) =O[atTop]
fun x => x * Real.log (Real.log x) / Real.log x ^ (N + 1)
noncomputable def meisselMertensConstant : ℝ :=
Filter.limUnder Filter.atTop (fun n : ℕ =>
(∑ p ∈ Finset.filter Nat.Prime (Finset.range (n + 1)),
1 / (p : ℝ)) - Real.log (Real.log n))
WantedExt/NumberTheory/PrimeCounting/DusartMertensReciprocalSumWanted.lean:27
noncomputable def almostPrimeCounting (k x : ℕ) : ℕ :=
((Finset.Icc 1 x).filter (fun n => Nat.IsAlmostPrime k n)).card
MathlibExt/NumberTheory/PrimeCounting/LandauAlmostPrimeAsymptotic.lean:55
Sources
- Paul Kinlaw, Megan Triplett, and William Tripp, Almost Primes of Almost Prime Index, INTEGERS 24 (2024), Article A99, Lemma 1 (quoting Crişan and Erban).
Uses 2
TheoremExplicit estimate for the reciprocal sum of the semiprimes#●ready to prove
A semiprime is a product of exactly two primes counted with multiplicity; for real \(x\) let \(S(x)\) be the sum of \(1/n\) over the semiprimes \(n \le x\). Let \(\beta\) be the Meissel–Mertens constant, \(\gamma\) the Euler–Mascheroni constant and \(\log\) the natural logarithm, and put
the sums running over all primes \(p\). Then for every real \(x > 1\) the error
satisfies \(|E(x)| < (\log x)^{-3/2}\).
public theorem_wanted semiprime_reciprocal_sum_explicit_estimate {x : ℝ} (hx : 1 < x) :
|(∑' n : {n : ℕ // Nat.IsAlmostPrime 2 n ∧ (n : ℝ) ≤ x}, (1 : ℝ) / (n : ℕ)) -
((Real.log (Real.log x) + meisselMertensConstant) ^ 2 / 2 +
((∑' p : Nat.Primes, (1 : ℝ) / ((p : ℕ) : ℝ) ^ 2) - Real.pi ^ 2 / 6) / 2 +
(Real.eulerMascheroniConstant +
∑' p : Nat.Primes, Real.log (p : ℕ) / (((p : ℕ) : ℝ) * (((p : ℕ) : ℝ) - 1))) /
Real.log x)| <
Real.log x ^ (-(3 / 2 : ℝ))
noncomputable def meisselMertensConstant : ℝ :=
Filter.limUnder Filter.atTop (fun n : ℕ =>
(∑ p ∈ Finset.filter Nat.Prime (Finset.range (n + 1)),
1 / (p : ℝ)) - Real.log (Real.log n))
WantedExt/NumberTheory/PrimeCounting/DusartMertensReciprocalSumWanted.lean:27
Sources
- Paul Kinlaw, Megan Triplett, and William Tripp, Almost Primes of Almost Prime Index, INTEGERS 24 (2024), Article A99, Lemma 2 (quoting Bayless et al.).
Uses 1
TheoremErdős–Sárközy bound for almost primes#●ready to prove
There is a real constant \(C\) such that for every integer \(k \ge 1\) and every real \(t \ge 1\), the number \(\pi_{k+1}(t)\) of integers \(n\) with \(1 \le n \le t\) that have exactly \(k + 1\) prime factors, counted with multiplicity, satisfies \(\pi_{k+1}(t) \le C \cdot \frac{k^4}{2^k} \, t \log t\).
public theorem_wanted almostPrime_count_le_erdos_sarkozy :
∃ C : ℝ, ∀ k : ℕ, 1 ≤ k → ∀ t : ℝ, 1 ≤ t →
(almostPrimeCounting (k + 1) ⌊t⌋₊ : ℝ) ≤
C * ((k : ℝ) ^ 4 / 2 ^ k) * t * Real.log t
noncomputable def almostPrimeCounting (k x : ℕ) : ℕ :=
((Finset.Icc 1 x).filter (fun n => Nat.IsAlmostPrime k n)).card
MathlibExt/NumberTheory/PrimeCounting/LandauAlmostPrimeAsymptotic.lean:55
Sources
- Paul Kinlaw, Megan Triplett, and William Tripp, Almost Primes of Almost Prime Index, INTEGERS 24 (2024), Article A99, Lemma 3 (quoting Erdős and Sárközy).
Uses 1
TheoremLogarithmic expansions of the prime counting function#●ready to prove
Let \(\pi(x)\) be the number of primes \(p \le x\). As \(x \to \infty\),
and
The last expansion is the paper's form, whose error term \(O(\log \log x / (\log x)^3)\) sits inside the parentheses, multiplied out by the factor \(1/\log x\).
public theorem_wanted log_primeCounting_asymptotics :
((fun x : ℝ => Real.log (Nat.primeCounting ⌊x⌋₊) -
(Real.log x - Real.log (Real.log x) + 1 / Real.log x + 3 / (2 * Real.log x ^ 2)))
=O[atTop] fun x => 1 / Real.log x ^ 3) ∧
((fun x : ℝ => Real.log (Real.log (Nat.primeCounting ⌊x⌋₊)) -
(Real.log (Real.log x) - Real.log (Real.log x) / Real.log x -
Real.log (Real.log x) ^ 2 / (2 * Real.log x ^ 2) + 1 / Real.log x ^ 2))
=O[atTop] fun x => Real.log (Real.log x) ^ 3 / Real.log x ^ 3) ∧
((fun x : ℝ => 1 / Real.log (Nat.primeCounting ⌊x⌋₊) -
1 / Real.log x * (1 + Real.log (Real.log x) / Real.log x +
(Real.log (Real.log x) ^ 2 - 1) / Real.log x ^ 2 +
Real.log (Real.log x) ^ 3 / Real.log x ^ 3))
=O[atTop] fun x => Real.log (Real.log x) / Real.log x ^ 4)
Sources
- Paul Kinlaw, Megan Triplett, and William Tripp, Almost Primes of Almost Prime Index, INTEGERS 24 (2024), Article A99, Lemma 5.
TheoremExpansion of the reciprocal logarithm of the semiprime count#●ready to prove
Let \(\pi_2(x)\) be the number of integers \(n\) with \(1 \le n \le x\) that have exactly two prime factors counted with multiplicity, and let \(\beta\) be the Meissel–Mertens constant. Then, as \(x \to \infty\),
This is the paper's expansion, whose error term \(O(1/(\log x \, (\log \log x)^2))\) sits inside the parentheses, multiplied out by the factor \(1/\log x\).
public theorem_wanted inv_log_semiprime_count_asymptotic :
(fun x : ℝ => 1 / Real.log (almostPrimeCounting 2 ⌊x⌋₊) -
1 / Real.log x * (1 + Real.log (Real.log x) / Real.log x -
Real.log (Real.log (Real.log x)) / Real.log x -
meisselMertensConstant / (Real.log x * Real.log (Real.log x))))
=O[atTop] fun x => 1 / (Real.log x ^ 2 * Real.log (Real.log x) ^ 2)
noncomputable def almostPrimeCounting (k x : ℕ) : ℕ :=
((Finset.Icc 1 x).filter (fun n => Nat.IsAlmostPrime k n)).card
MathlibExt/NumberTheory/PrimeCounting/LandauAlmostPrimeAsymptotic.lean:55
noncomputable def meisselMertensConstant : ℝ :=
Filter.limUnder Filter.atTop (fun n : ℕ =>
(∑ p ∈ Finset.filter Nat.Prime (Finset.range (n + 1)),
1 / (p : ℝ)) - Real.log (Real.log n))
WantedExt/NumberTheory/PrimeCounting/DusartMertensReciprocalSumWanted.lean:27
Sources
- Paul Kinlaw, Megan Triplett, and William Tripp, Almost Primes of Almost Prime Index, INTEGERS 24 (2024), Article A99, Lemma 6.
Uses 2
TheoremCounting primes of semiprime index#●ready to prove
A semiprime is a product of exactly two primes counted with multiplicity. Let \(p_i\) denote the \(i\)th prime (\(p_1 = 2\)), and for real \(x\) let \(N_{1,2}(x)\) be the number of primes \(p_i \le x\) whose index \(i\) is a semiprime. Let \(\beta\) be the Meissel–Mertens constant, \(\gamma\) the Euler–Mascheroni constant, \(\log\) the natural logarithm, and
the sum running over all primes \(p\). Then, as \(x \to \infty\),
public theorem_wanted primes_of_semiprime_index_count_asymptotic :
(fun x : ℝ => (Set.ncard {n : ℕ | n ∈ almostPrimeIndexSet 1 [2] ∧ (n : ℝ) ≤ x} : ℝ) -
x / Real.log x ^ 2 * (Real.log (Real.log x) + meisselMertensConstant +
((Real.log (Real.log x) + meisselMertensConstant) * (Real.log (Real.log x) + 1) +
(meisselMertensConstant - 1 - Real.eulerMascheroniConstant -
∑' p : Nat.Primes, Real.log (p : ℕ) / (((p : ℕ) : ℝ) * (((p : ℕ) : ℝ) - 1)))) /
Real.log x))
=O[atTop] fun x => x * Real.log (Real.log x) ^ 3 / Real.log x ^ 4
public def almostPrimeIndexSet (k : ℕ) : List ℕ → Set ℕ
| [] => {n | Nat.IsAlmostPrime k n}
| j :: L => {n | ∃ i ∈ almostPrimeIndexSet j L, 1 ≤ i ∧ n = Nat.nth (Nat.IsAlmostPrime k) (i - 1)}
WantedExt/NumberTheory/AlmostPrimeIndexWanted.lean:60
noncomputable def meisselMertensConstant : ℝ :=
Filter.limUnder Filter.atTop (fun n : ℕ =>
(∑ p ∈ Finset.filter Nat.Prime (Finset.range (n + 1)),
1 / (p : ℝ)) - Real.log (Real.log n))
WantedExt/NumberTheory/PrimeCounting/DusartMertensReciprocalSumWanted.lean:27
Sources
- Paul Kinlaw, Megan Triplett, and William Tripp, Almost Primes of Almost Prime Index, INTEGERS 24 (2024), Article A99, Theorem 1.
Uses 2
TheoremCounting semiprimes of prime index#●ready to prove
A semiprime is a product of exactly two primes counted with multiplicity. Let \(s_i\) denote the \(i\)th semiprime (\(s_1 = 4\)), and for real \(x\) let \(N_{2,1}(x)\) be the number of semiprimes \(s_i \le x\) whose index \(i\) is prime. Let \(\beta\) be the Meissel–Mertens constant and \(\log\) the natural logarithm. Then, as \(x \to \infty\),
public theorem_wanted semiprimes_of_prime_index_count_asymptotic :
(fun x : ℝ => (Set.ncard {n : ℕ | n ∈ almostPrimeIndexSet 2 [1] ∧ (n : ℝ) ≤ x} : ℝ) -
x / Real.log x ^ 2 * (Real.log (Real.log x) + meisselMertensConstant +
Real.log (Real.log x) * (Real.log (Real.log x) - Real.log (Real.log (Real.log x))) /
Real.log x))
=O[atTop] fun x => x * Real.log (Real.log x) / Real.log x ^ 3
public def almostPrimeIndexSet (k : ℕ) : List ℕ → Set ℕ
| [] => {n | Nat.IsAlmostPrime k n}
| j :: L => {n | ∃ i ∈ almostPrimeIndexSet j L, 1 ≤ i ∧ n = Nat.nth (Nat.IsAlmostPrime k) (i - 1)}
WantedExt/NumberTheory/AlmostPrimeIndexWanted.lean:60
noncomputable def meisselMertensConstant : ℝ :=
Filter.limUnder Filter.atTop (fun n : ℕ =>
(∑ p ∈ Finset.filter Nat.Prime (Finset.range (n + 1)),
1 / (p : ℝ)) - Real.log (Real.log n))
WantedExt/NumberTheory/PrimeCounting/DusartMertensReciprocalSumWanted.lean:27
Sources
- Paul Kinlaw, Megan Triplett, and William Tripp, Almost Primes of Almost Prime Index, INTEGERS 24 (2024), Article A99, Theorem 2.
Uses 2
TheoremPrimes of semiprime index eventually outnumber semiprimes of prime index#●ready to prove
A semiprime is a product of exactly two primes counted with multiplicity. Let \(p_i\) denote the \(i\)th prime (\(p_1 = 2\)) and \(s_i\) the \(i\)th semiprime (\(s_1 = 4\)). For real \(x\) let \(N_{1,2}(x)\) be the number of primes \(p_i \le x\) whose index \(i\) is a semiprime, and \(N_{2,1}(x)\) the number of semiprimes \(s_i \le x\) whose index \(i\) is prime. Then \(N_{2,1}(x) < N_{1,2}(x)\) for all sufficiently large \(x\), and, with \(\log\) the natural logarithm, as \(x \to \infty\)
This is the form \(\frac{x \log \log x \, \log \log \log x}{\log^3 x} (1 + O(1 / \log \log \log x))\) with the main factor multiplied out.
public theorem_wanted primes_of_semiprime_index_count_sub_semiprimes_of_prime_index :
(∀ᶠ x : ℝ in atTop,
Set.ncard {n : ℕ | n ∈ almostPrimeIndexSet 2 [1] ∧ (n : ℝ) ≤ x} <
Set.ncard {n : ℕ | n ∈ almostPrimeIndexSet 1 [2] ∧ (n : ℝ) ≤ x}) ∧
(fun x : ℝ => (Set.ncard {n : ℕ | n ∈ almostPrimeIndexSet 1 [2] ∧ (n : ℝ) ≤ x} : ℝ) -
Set.ncard {n : ℕ | n ∈ almostPrimeIndexSet 2 [1] ∧ (n : ℝ) ≤ x} -
x * Real.log (Real.log x) * Real.log (Real.log (Real.log x)) / Real.log x ^ 3)
=O[atTop] fun x => x * Real.log (Real.log x) / Real.log x ^ 3
public def almostPrimeIndexSet (k : ℕ) : List ℕ → Set ℕ
| [] => {n | Nat.IsAlmostPrime k n}
| j :: L => {n | ∃ i ∈ almostPrimeIndexSet j L, 1 ≤ i ∧ n = Nat.nth (Nat.IsAlmostPrime k) (i - 1)}
WantedExt/NumberTheory/AlmostPrimeIndexWanted.lean:60
Sources
- Paul Kinlaw, Megan Triplett, and William Tripp, Almost Primes of Almost Prime Index, INTEGERS 24 (2024), Article A99, Corollary 1.
Uses 1
TheoremCounting almost primes of almost prime index#●ready to prove
Let \(j, k \ge 1\) be integers and, for real \(x\), let \(N_{jk}(x)\) be the number of \(j\)-almost primes \(n \le x\) whose index in the increasing enumeration of the \(j\)-almost primes, counted from \(1\), is a \(k\)-almost prime. Writing \(M(x) = \frac{x \, (\log \log x)^{j+k-2}}{(\log x)^2 \, (j-1)! \, (k-1)!}\), we have \(N_{jk}(x) - M(x) = O(M(x) / \log \log x)\) as \(x \to \infty\).
The paper's clause "the same estimate holds for \(N_{kj}\)" is the instance with \(j\) and \(k\) swapped, since the main term is symmetric.
public theorem_wanted almostPrimeIndexSet_pair_count_asymptotic {j k : ℕ} (hj : 1 ≤ j)
(hk : 1 ≤ k) :
(fun x : ℝ => (Set.ncard {n : ℕ | n ∈ almostPrimeIndexSet j [k] ∧ (n : ℝ) ≤ x} : ℝ) -
x / Real.log x ^ 2 * Real.log (Real.log x) ^ (j + k - 2) /
((j - 1).factorial * (k - 1).factorial))
=O[atTop] fun x => x / Real.log x ^ 2 * Real.log (Real.log x) ^ (j + k - 2) /
((j - 1).factorial * (k - 1).factorial) / Real.log (Real.log x)
public def almostPrimeIndexSet (k : ℕ) : List ℕ → Set ℕ
| [] => {n | Nat.IsAlmostPrime k n}
| j :: L => {n | ∃ i ∈ almostPrimeIndexSet j L, 1 ≤ i ∧ n = Nat.nth (Nat.IsAlmostPrime k) (i - 1)}
WantedExt/NumberTheory/AlmostPrimeIndexWanted.lean:60
Sources
- Paul Kinlaw, Megan Triplett, and William Tripp, Almost Primes of Almost Prime Index, INTEGERS 24 (2024), Article A99, Theorem 3.
Uses 1
TheoremCounting almost primes of iterated almost prime index#●ready to prove
Let \(m \ge 1\), let \(k_1, \dots, k_m\) be positive integers, let \(N_{k_1 \cdots k_m}\) be the almost prime index set (for \(m = 1\) the \(k_1\)-almost primes, and for \(m \ge 2\) the \(k_1\)-almost primes whose index in increasing order, counted from \(1\), lies in \(N_{k_2 \cdots k_m}\)), and for real \(x\) let \(N_{k_1 \cdots k_m}(x)\) be the number of its elements \(n \le x\). Writing \(M(x) = \frac{x}{(\log x)^m} \prod_{i=1}^m \frac{(\log \log x)^{k_i - 1}}{(k_i - 1)!}\), we have \(N_{k_1 \cdots k_m}(x) - M(x) = O(M(x) / \log \log x)\) as \(x \to \infty\).
The paper's displayed exponent \(k_1 + \cdots + k_{m-1} k_m - m\) is a typo for \(k_1 + \cdots + k_{m-1} + k_m - m\), as its equivalent product form shows.
public theorem_wanted almostPrimeIndexSet_count_asymptotic {k : ℕ} {ks : List ℕ}
(hpos : ∀ i ∈ k :: ks, 1 ≤ i) :
(fun x : ℝ => (Set.ncard {n : ℕ | n ∈ almostPrimeIndexSet k ks ∧ (n : ℝ) ≤ x} : ℝ) -
x / Real.log x ^ (ks.length + 1) *
((k :: ks).map fun i => Real.log (Real.log x) ^ (i - 1) / (i - 1).factorial).prod)
=O[atTop] fun x => x / Real.log x ^ (ks.length + 1) *
((k :: ks).map fun i => Real.log (Real.log x) ^ (i - 1) / (i - 1).factorial).prod /
Real.log (Real.log x)
public def almostPrimeIndexSet (k : ℕ) : List ℕ → Set ℕ
| [] => {n | Nat.IsAlmostPrime k n}
| j :: L => {n | ∃ i ∈ almostPrimeIndexSet j L, 1 ≤ i ∧ n = Nat.nth (Nat.IsAlmostPrime k) (i - 1)}
WantedExt/NumberTheory/AlmostPrimeIndexWanted.lean:60
Sources
- Paul Kinlaw, Megan Triplett, and William Tripp, Almost Primes of Almost Prime Index, INTEGERS 24 (2024), Article A99, Theorem 4.
Uses 1
TheoremSize of the nth almost prime of almost prime index#●ready to prove
Let \(j, k \ge 1\) be integers and, for \(n \ge 1\), let \(p_{j,k,n}\) be the \(n\)th element, in increasing order, of the set \(N_{jk}\) of \(j\)-almost primes whose index in the increasing enumeration of the \(j\)-almost primes, counted from \(1\), is a \(k\)-almost prime. Writing \(P(n) = \frac{n (\log n)^2 \, (j-1)! \, (k-1)!}{(\log \log n)^{j+k-2}}\), we have \(p_{j,k,n} - P(n) = O(P(n) / \log \log n)\) as \(n \to \infty\) through the integers.
public theorem_wanted nth_almostPrimeIndexSet_pair_asymptotic {j k : ℕ} (hj : 1 ≤ j)
(hk : 1 ≤ k) :
(fun n : ℕ => (Nat.nth (· ∈ almostPrimeIndexSet j [k]) (n - 1) : ℝ) -
n * Real.log n ^ 2 * ((j - 1).factorial * (k - 1).factorial) /
Real.log (Real.log n) ^ (j + k - 2))
=O[atTop] fun n : ℕ => n * Real.log n ^ 2 * ((j - 1).factorial * (k - 1).factorial) /
Real.log (Real.log n) ^ (j + k - 2) / Real.log (Real.log n)
public def almostPrimeIndexSet (k : ℕ) : List ℕ → Set ℕ
| [] => {n | Nat.IsAlmostPrime k n}
| j :: L => {n | ∃ i ∈ almostPrimeIndexSet j L, 1 ≤ i ∧ n = Nat.nth (Nat.IsAlmostPrime k) (i - 1)}
WantedExt/NumberTheory/AlmostPrimeIndexWanted.lean:60
Sources
- Paul Kinlaw, Megan Triplett, and William Tripp, Almost Primes of Almost Prime Index, INTEGERS 24 (2024), Article A99, Corollary 2.
Uses 1
TheoremSize of the nth element of an iterated almost prime index set#●ready to prove
Let \(m \ge 1\), let \(k_1, \dots, k_m\) be positive integers, and for \(n \ge 1\) let \(p_{k_1, \dots, k_m, n}\) be the \(n\)th element, in increasing order, of the almost prime index set \(N_{k_1 \cdots k_m}\). Writing \(P(n) = n (\log n)^m \prod_{i=1}^m \frac{(k_i - 1)!}{(\log \log n)^{k_i - 1}}\), we have \(p_{k_1, \dots, k_m, n} - P(n) = O(P(n) / \log \log n)\) as \(n \to \infty\) through the integers.
public theorem_wanted nth_almostPrimeIndexSet_asymptotic {k : ℕ} {ks : List ℕ}
(hpos : ∀ i ∈ k :: ks, 1 ≤ i) :
(fun n : ℕ => (Nat.nth (· ∈ almostPrimeIndexSet k ks) (n - 1) : ℝ) -
n * Real.log n ^ (ks.length + 1) *
((k :: ks).map fun i => ((i - 1).factorial : ℝ) / Real.log (Real.log n) ^ (i - 1)).prod)
=O[atTop] fun n : ℕ => n * Real.log n ^ (ks.length + 1) *
((k :: ks).map fun i => ((i - 1).factorial : ℝ) / Real.log (Real.log n) ^ (i - 1)).prod /
Real.log (Real.log n)
public def almostPrimeIndexSet (k : ℕ) : List ℕ → Set ℕ
| [] => {n | Nat.IsAlmostPrime k n}
| j :: L => {n | ∃ i ∈ almostPrimeIndexSet j L, 1 ≤ i ∧ n = Nat.nth (Nat.IsAlmostPrime k) (i - 1)}
WantedExt/NumberTheory/AlmostPrimeIndexWanted.lean:60
Sources
- Paul Kinlaw, Megan Triplett, and William Tripp, Almost Primes of Almost Prime Index, INTEGERS 24 (2024), Article A99, Corollary 3.
Uses 1
TheoremReciprocal sum of the primes of \(k\)-almost prime index#●ready to prove
For \(k \ge 0\) let \(N_{1k}\) be the set of primes \(p_i\) (indexed from \(p_1 = 2\)) whose index \(i\) is a \(k\)-almost prime, that is, has exactly \(k\) prime factors counted with multiplicity. Then \(\sum_{n \in N_{1k}} 1/n = 1 + O(e^{-k/2})\) as \(k \to \infty\).
public theorem_wanted tsum_inv_primes_of_almostPrime_index_sub_one_isBigO :
(fun k : ℕ => (∑' n : almostPrimeIndexSet 1 [k], (1 : ℝ) / n) - 1)
=O[atTop] fun k : ℕ => Real.exp (-(k : ℝ) / 2)
public def almostPrimeIndexSet (k : ℕ) : List ℕ → Set ℕ
| [] => {n | Nat.IsAlmostPrime k n}
| j :: L => {n | ∃ i ∈ almostPrimeIndexSet j L, 1 ≤ i ∧ n = Nat.nth (Nat.IsAlmostPrime k) (i - 1)}
WantedExt/NumberTheory/AlmostPrimeIndexWanted.lean:60
Sources
- Paul Kinlaw, Megan Triplett, and William Tripp, Almost Primes of Almost Prime Index, INTEGERS 24 (2024), Article A99, Theorem 7.
Uses 1
TheoremLower bound for reciprocal sums over almost primes of almost prime index#●ready to prove
A \(k\)-almost prime is a positive integer with exactly \(k\) prime factors counted with multiplicity. For positive integers \(j\) and \(k\) let \(N_{jk}\) be the set of \(j\)-almost primes whose index in the increasing enumeration of the \(j\)-almost primes (starting at \(1\)) is a \(k\)-almost prime. Fix an integer \(j \ge 1\). There is a real constant \(c > 0\) such that
for all sufficiently large \(k\). Moreover, if \(j \ge 2\), then \(\sum_{n \in N_{jk}} 1/n \to \infty\) as \(k \to \infty\).
public theorem_wanted tsum_inv_almostPrimeIndexSet_pair_lower_bound {j : ℕ} (hj : 1 ≤ j) :
(∃ c : ℝ, 0 < c ∧ ∀ᶠ k : ℕ in atTop,
c * (k : ℝ) ^ (j - 1) ≤ ∑' n : almostPrimeIndexSet j [k], (1 : ℝ) / n) ∧
(2 ≤ j → Tendsto (fun k : ℕ => ∑' n : almostPrimeIndexSet j [k], (1 : ℝ) / n) atTop atTop)
public def almostPrimeIndexSet (k : ℕ) : List ℕ → Set ℕ
| [] => {n | Nat.IsAlmostPrime k n}
| j :: L => {n | ∃ i ∈ almostPrimeIndexSet j L, 1 ≤ i ∧ n = Nat.nth (Nat.IsAlmostPrime k) (i - 1)}
WantedExt/NumberTheory/AlmostPrimeIndexWanted.lean:60
Sources
- Paul Kinlaw, Megan Triplett, and William Tripp, Almost Primes of Almost Prime Index, INTEGERS 24 (2024), Article A99, Theorem 8.
Uses 1
TheoremExponential decay of reciprocal sums over triple almost prime index sets#●ready to prove
A \(k\)-almost prime is a positive integer with exactly \(k\) prime factors counted with multiplicity. For positive integers \(a\) and \(b\) let \(N_{ab}\) be the set of \(a\)-almost primes whose index in the increasing enumeration of the \(a\)-almost primes (starting at \(1\)) is a \(b\)-almost prime, and for positive integers \(k_1, k_2, k\) let \(N_{k_1 k_2 k}\) be the set of \(k_1\)-almost primes whose index in the increasing enumeration of the \(k_1\)-almost primes lies in \(N_{k_2 k}\). Fix positive integers \(k_1\) and \(k_2\). Then, as \(k \to \infty\),
This is the case \(m = 2\) of the paper's bound \(\ll (e^{k/2})^{-(m-1)}\) for the reciprocal sum of \(N_{k_1 \cdots k_m k}\), the only case its proof supports. For \(m = 3\) and \(k_1 = k_2 = k_3 = 1\), with \(p_n\) the \(n\)th prime, the single term \(1/p_{p_{p_{2^k}}}\) is \(\gg 2^{-k} k^{-3}\) because \(p_n < n (\log n + \log \log n)\) for \(n \ge 6\), and \(e^{k} 2^{-k} k^{-3} \to \infty\), so the printed bound fails for \(m = 3\).
public theorem_wanted tsum_inv_almostPrimeIndexSet_triple_isBigO {k₁ k₂ : ℕ} (hk₁ : 1 ≤ k₁)
(hk₂ : 1 ≤ k₂) :
(fun k : ℕ => ∑' n : almostPrimeIndexSet k₁ [k₂, k], (1 : ℝ) / n)
=O[atTop] fun k : ℕ => Real.exp (-(k : ℝ) / 2)
public def almostPrimeIndexSet (k : ℕ) : List ℕ → Set ℕ
| [] => {n | Nat.IsAlmostPrime k n}
| j :: L => {n | ∃ i ∈ almostPrimeIndexSet j L, 1 ≤ i ∧ n = Nat.nth (Nat.IsAlmostPrime k) (i - 1)}
WantedExt/NumberTheory/AlmostPrimeIndexWanted.lean:60
Sources
- Paul Kinlaw, Megan Triplett, and William Tripp, Almost Primes of Almost Prime Index, INTEGERS 24 (2024), Article A99, Theorem 9, case
m = 2.
Uses 1
TheoremReciprocal sums over iterated almost prime index sets tend to zero#●ready to prove
A \(k\)-almost prime is a positive integer with exactly \(k\) prime factors counted with multiplicity. For positive integers \(k_1, \dots, k_m\) let \(N_{k_1}\) be the set of \(k_1\)-almost primes and, for \(m \ge 2\), let \(N_{k_1 k_2 \cdots k_m}\) be the set of \(k_1\)-almost primes whose index in the increasing enumeration of the \(k_1\)-almost primes (starting at \(1\)) lies in \(N_{k_2 \cdots k_m}\). Fix \(m \ge 2\) and positive integers \(k_1, \dots, k_m\). Then
public theorem_wanted tendsto_tsum_inv_almostPrimeIndexSet_append_zero {k : ℕ} {ks : List ℕ}
(hks : ks ≠ []) (hpos : ∀ i ∈ k :: ks, 1 ≤ i) :
Tendsto (fun k' : ℕ => ∑' n : almostPrimeIndexSet k (ks ++ [k']), (1 : ℝ) / n) atTop (𝓝 0)
public def almostPrimeIndexSet (k : ℕ) : List ℕ → Set ℕ
| [] => {n | Nat.IsAlmostPrime k n}
| j :: L => {n | ∃ i ∈ almostPrimeIndexSet j L, 1 ≤ i ∧ n = Nat.nth (Nat.IsAlmostPrime k) (i - 1)}
WantedExt/NumberTheory/AlmostPrimeIndexWanted.lean:60
Sources
- Paul Kinlaw, Megan Triplett, and William Tripp, Almost Primes of Almost Prime Index, INTEGERS 24 (2024), Article A99, Theorem 9, "in particular" clause.
Uses 1
Loading dependency explorer…
The dependency explorer needs JavaScript and its graph data.