Distinct multiples in an interval
DefinitionDistinct multiples of \(1, \dots, n\) in an interval#✓fully proved
For natural numbers \(n\), \(m\) and \(L\), the interval \((m, m + L]\) contains distinct multiples of \(1, \dots, n\) when there are pairwise distinct natural numbers \(a_1, \dots, a_n\) with \(m < a_i \le m + L\) and \(i \mid a_i\) for every \(1 \le i \le n\).
def HasDistinctMultiplesInInterval (n m L : ℕ) : Prop :=
∃ a : Fin n → ℕ, Function.Injective a ∧
∀ i : Fin n, m < a i ∧ a i ≤ m + L ∧ (i.val + 1) ∣ a i
Sources
- Wouter van Doorn, On the Length of an Interval That Contains Distinct Multiples of the First n Positive Integers, INTEGERS 26 (2026), Article A7, Section 1 (definition of
f(n, m), after Erdős and Pomerance).
Used by 6
DefinitionLeast interval length containing distinct multiples#✓fully proved
For natural numbers \(n\) and \(m\), \(f(n, m)\) is the least natural number \(L\) such that the interval \((m, m + L]\) contains distinct multiples of \(1, \dots, n\).
Such an \(L\) always exists, so the least one is well defined.
noncomputable def distinctMultiplesIntervalLength (n m : ℕ) : ℕ :=
sInf {L | HasDistinctMultiplesInInterval n m L}
def HasDistinctMultiplesInInterval (n m L : ℕ) : Prop :=
∃ a : Fin n → ℕ, Function.Injective a ∧
∀ i : Fin n, m < a i ∧ a i ≤ m + L ∧ (i.val + 1) ∣ a i
MathlibExt/NumberTheory/DistinctMultiplesInterval.lean:36
Sources
- Wouter van Doorn, On the Length of an Interval That Contains Distinct Multiples of the First n Positive Integers, INTEGERS 26 (2026), Article A7, Section 1 (definition of
f(n, m), after Erdős and Pomerance).
Uses 1 · Used by 6
TheoremLengthening the interval preserves distinct multiples#✓fully proved
For natural numbers \(n\), \(m\), \(L\) and \(L'\) with \(L \le L'\), if the interval \((m, m + L]\) contains distinct multiples of \(1, \dots, n\), then so does the interval \((m, m + L']\).
theorem HasDistinctMultiplesInInterval.mono {n m L L' : ℕ} (hLL' : L ≤ L')
(h : HasDistinctMultiplesInInterval n m L) : HasDistinctMultiplesInInterval n m L'
def HasDistinctMultiplesInInterval (n m L : ℕ) : Prop :=
∃ a : Fin n → ℕ, Function.Injective a ∧
∀ i : Fin n, m < a i ∧ a i ≤ m + L ∧ (i.val + 1) ∣ a i
MathlibExt/NumberTheory/DistinctMultiplesInterval.lean:36
Sources
- Wouter van Doorn, On the Length of an Interval That Contains Distinct Multiples of the First n Positive Integers, INTEGERS 26 (2026), Article A7, Section 1 (definition of
f(n, m)).
Uses 1 · Used by 1
TheoremSome interval contains distinct multiples#✓fully proved
For all natural numbers \(n\) and \(m\) there is a natural number \(L\) such that the interval \((m, m + L]\) contains distinct multiples of \(1, \dots, n\).
For example, \(L = n(m + 1)\) works, with \(a_i = i(m + 1)\) for \(1 \le i \le n\).
theorem exists_hasDistinctMultiplesInInterval (n m : ℕ) :
∃ L, HasDistinctMultiplesInInterval n m L
def HasDistinctMultiplesInInterval (n m L : ℕ) : Prop :=
∃ a : Fin n → ℕ, Function.Injective a ∧
∀ i : Fin n, m < a i ∧ a i ≤ m + L ∧ (i.val + 1) ∣ a i
MathlibExt/NumberTheory/DistinctMultiplesInterval.lean:36
Sources
- Wouter van Doorn, On the Length of an Interval That Contains Distinct Multiples of the First n Positive Integers, INTEGERS 26 (2026), Article A7, Section 1 (definition of
f(n, m)).
Uses 1 · Used by 1
TheoremThe least interval length is admissible#✓fully proved
For natural numbers \(n\) and \(m\), let \(f(n, m)\) be the least natural number \(L\) such that the interval \((m, m + L]\) contains distinct multiples of \(1, \dots, n\). Then the interval \((m, m + f(n, m)]\) contains distinct multiples of \(1, \dots, n\).
theorem hasDistinctMultiplesInInterval_distinctMultiplesIntervalLength (n m : ℕ) :
HasDistinctMultiplesInInterval n m (distinctMultiplesIntervalLength n m)
def HasDistinctMultiplesInInterval (n m L : ℕ) : Prop :=
∃ a : Fin n → ℕ, Function.Injective a ∧
∀ i : Fin n, m < a i ∧ a i ≤ m + L ∧ (i.val + 1) ∣ a i
MathlibExt/NumberTheory/DistinctMultiplesInterval.lean:36
noncomputable def distinctMultiplesIntervalLength (n m : ℕ) : ℕ :=
sInf {L | HasDistinctMultiplesInInterval n m L}
MathlibExt/NumberTheory/DistinctMultiplesInterval.lean:46
Sources
- Wouter van Doorn, On the Length of an Interval That Contains Distinct Multiples of the First n Positive Integers, INTEGERS 26 (2026), Article A7, Section 1 (definition of
f(n, m)).
Uses 3 · Used by 1
TheoremCharacterization of the least interval length#✓fully proved
For natural numbers \(n\) and \(m\), let \(f(n, m)\) be the least natural number \(K\) such that the interval \((m, m + K]\) contains distinct multiples of \(1, \dots, n\). Then for every natural number \(L\), \(f(n, m) \le L\) if and only if the interval \((m, m + L]\) contains distinct multiples of \(1, \dots, n\).
theorem distinctMultiplesIntervalLength_le_iff {n m L : ℕ} :
distinctMultiplesIntervalLength n m ≤ L ↔ HasDistinctMultiplesInInterval n m L
noncomputable def distinctMultiplesIntervalLength (n m : ℕ) : ℕ :=
sInf {L | HasDistinctMultiplesInInterval n m L}
MathlibExt/NumberTheory/DistinctMultiplesInterval.lean:46
def HasDistinctMultiplesInInterval (n m L : ℕ) : Prop :=
∃ a : Fin n → ℕ, Function.Injective a ∧
∀ i : Fin n, m < a i ∧ a i ≤ m + L ∧ (i.val + 1) ∣ a i
MathlibExt/NumberTheory/DistinctMultiplesInterval.lean:36
Sources
- Wouter van Doorn, On the Length of an Interval That Contains Distinct Multiples of the First n Positive Integers, INTEGERS 26 (2026), Article A7, Section 1 (definition of
f(n, m)).
TheoremThe least interval length for \(n = 1\)#✓fully proved
For natural numbers \(n\) and \(m\), let \(f(n, m)\) be the least natural number \(L\) such that the interval \((m, m + L]\) contains distinct multiples of \(1, \dots, n\). Then \(f(1, m) = 1\) for every natural number \(m\): the interval \((m, m + 1]\) contains the multiple \(m + 1\) of \(1\), and the empty interval \((m, m]\) contains nothing.
theorem distinctMultiplesIntervalLength_one (m : ℕ) :
distinctMultiplesIntervalLength 1 m = 1
noncomputable def distinctMultiplesIntervalLength (n m : ℕ) : ℕ :=
sInf {L | HasDistinctMultiplesInInterval n m L}
MathlibExt/NumberTheory/DistinctMultiplesInterval.lean:46
Sources
- Wouter van Doorn, On the Length of an Interval That Contains Distinct Multiples of the First n Positive Integers, INTEGERS 26 (2026), Article A7, Section 1 (definition of
f(n, m)).
Uses 2
TheoremAn inequality between diagonal and shifted interval lengths#●ready to prove
For natural numbers \(n\) and \(m\), let \(f(n, m)\) be the least natural number \(L\) such that the interval \((m, m + L]\) contains distinct multiples of \(1, \dots, n\). Then for all positive integers \(k\) and \(n\), \(kn + f(kn, kn) \le k^2 n + f(n, k^2 n)\).
public theorem_wanted distinctMultiplesIntervalLength_mul_le {k n : ℕ} (hk : 0 < k)
(hn : 0 < n) :
k * n + distinctMultiplesIntervalLength (k * n) (k * n) ≤
k ^ 2 * n + distinctMultiplesIntervalLength n (k ^ 2 * n)
noncomputable def distinctMultiplesIntervalLength (n m : ℕ) : ℕ :=
sInf {L | HasDistinctMultiplesInInterval n m L}
MathlibExt/NumberTheory/DistinctMultiplesInterval.lean:46
Sources
- Wouter van Doorn, On the Length of an Interval That Contains Distinct Multiples of the First n Positive Integers, INTEGERS 26 (2026), Article A7, Lemma 2.
Uses 1
TheoremErdős–Pomerance bounds for the diagonal interval length#●ready to prove
For natural numbers \(n\) and \(m\), let \(f(n, m)\) be the least natural number \(L\) such that the interval \((m, m + L]\) contains distinct multiples of \(1, \dots, n\). For every real \(\varepsilon > 0\) and all sufficiently large natural numbers \(n\),
public theorem_wanted distinctMultiplesIntervalLength_diag_bounds (ε : ℝ) (hε : 0 < ε) :
∀ᶠ n : ℕ in Filter.atTop,
(2 / Real.sqrt (Real.exp 1) - ε) * n * Real.sqrt (Real.log n / Real.log (Real.log n)) <
distinctMultiplesIntervalLength n n ∧
(distinctMultiplesIntervalLength n n : ℝ) < (2 + ε) * n * Real.sqrt (Real.log n)
noncomputable def distinctMultiplesIntervalLength (n m : ℕ) : ℕ :=
sInf {L | HasDistinctMultiplesInInterval n m L}
MathlibExt/NumberTheory/DistinctMultiplesInterval.lean:46
Sources
- the constants are read from the paper text (arXiv:2601.16972, same paper, Lemma 3), since the record's are garbled; they agree with the record's
√2e > 1.21remark (2/√e ≈ 1.213). - Wouter van Doorn, On the Length of an Interval That Contains Distinct Multiples of the First n Positive Integers, INTEGERS 26 (2026), Article A7, Lemma 3, quoting Erdős and Pomerance.
Uses 1
TheoremShifting the interval can greatly increase the least length#●ready to prove
For natural numbers \(n\) and \(m\), let \(f(n, m)\) be the least natural number \(L\) such that the interval \((m, m + L]\) contains distinct multiples of \(1, \dots, n\). For all sufficiently large natural numbers \(n\) there is a positive integer \(m\) with
Since \(f(n, m)\) is periodic in \(m\) with period \(\operatorname{lcm}(1, \dots, n)\), this says that \(\max_{m \ge 1} f(n, m) - f(n, n) > 0.36\, n \log n / \log \log n\) for all large \(n\), the form stated in the source.
public theorem_wanted exists_distinctMultiplesIntervalLength_sub_gt :
∀ᶠ n : ℕ in Filter.atTop, ∃ m : ℕ, 0 < m ∧
0.36 * (n : ℝ) * Real.log n / Real.log (Real.log n) <
(distinctMultiplesIntervalLength n m : ℝ) - distinctMultiplesIntervalLength n n
noncomputable def distinctMultiplesIntervalLength (n m : ℕ) : ℕ :=
sInf {L | HasDistinctMultiplesInInterval n m L}
MathlibExt/NumberTheory/DistinctMultiplesInterval.lean:46
Sources
- Wouter van Doorn, On the Length of an Interval That Contains Distinct Multiples of the First n Positive Integers, INTEGERS 26 (2026), Article A7, Theorem 1.
Uses 1
DefinitionErdős's matching function for multiples in intervals#✓fully proved
For a natural number \(m\), \(f(m)\) is the largest natural number \(r\) such that for every natural number \(N\), every set \(A \subseteq \{1, \dots, N\}\) with exactly \(m\) elements and every real \(x \ge 1\), there are distinct \(a_1, \dots, a_r \in A\) and distinct natural numbers \(b_1, \dots, b_r\) in the open interval \((x, x + 2N)\) with \(a_i \mid b_i\) for every \(i\).
noncomputable def f (m : ℕ) : ℕ :=
sSup {r : ℕ | ∀ N : ℕ, ∀ A ⊆ Finset.Icc 1 N, A.card = m → ∀ x : ℝ, 1 ≤ x →
∃ a b : Fin r → ℕ, Function.Injective a ∧ Function.Injective b ∧
(∀ i, a i ∈ A) ∧ (∀ i, (b i : ℝ) ∈ Set.Ioo x (x + 2 * (N : ℝ))) ∧
(∀ i, a i ∣ b i)}
Used by 1
DefinitionStatement of Erdős Problem 650#✓fully proved
Let \(f\) be Erdős's matching function for multiples in intervals. The statement is the conjunction of four claims: for every natural number \(m\), \(f(m) = \min(m, \lceil 2\sqrt{m}\,\rceil)\); it is not the case that \(f(m) \le \sqrt{m}\) for every natural number \(m\); \(\sqrt{m} \le f(m)\) for every natural number \(m\); and \(f(m^2) \le 2m\) for every natural number \(m\).
def conjecture : Prop :=
(∀ m : ℕ, f m = min m ⌈2 * Real.sqrt m⌉₊) ∧
(¬ ∀ m : ℕ, (f m : ℝ) ≤ Real.sqrt m) ∧
(∀ m : ℕ, Real.sqrt m ≤ (f m : ℝ)) ∧
(∀ m : ℕ, f (m ^ 2) ≤ 2 * m)
noncomputable def f (m : ℕ) : ℕ :=
sSup {r : ℕ | ∀ N : ℕ, ∀ A ⊆ Finset.Icc 1 N, A.card = m → ∀ x : ℝ, 1 ≤ x →
∃ a b : Fin r → ℕ, Function.Injective a ∧ Function.Injective b ∧
(∀ i, a i ∈ A) ∧ (∀ i, (b i : ℝ) ∈ Set.Ioo x (x + 2 * (N : ℝ))) ∧
(∀ i, a i ∣ b i)}
WantedExt/NumberTheory/ErdosProblem650Wanted.lean:35
Uses 1 · Used by 1
TheoremErdős Problem 650: matching integers to distinct multiples#●ready to prove
For a natural number \(m\), let \(f(m)\) be the largest natural number \(r\) such that for every natural number \(N\), every set \(A \subseteq \{1, \dots, N\}\) with exactly \(m\) elements and every real \(x \ge 1\), there are distinct \(a_1, \dots, a_r \in A\) and distinct natural numbers \(b_1, \dots, b_r\) in the open interval \((x, x + 2N)\) with \(a_i \mid b_i\) for every \(i\). Then \(f(m) = \min(m, \lceil 2\sqrt{m}\,\rceil)\) for every natural number \(m\); it is not the case that \(f(m) \le \sqrt{m}\) for every natural number \(m\); \(\sqrt{m} \le f(m)\) for every natural number \(m\); and \(f(m^2) \le 2m\) for every natural number \(m\).
Solved by Erdős–Surányi [ErSu59], Erdős–Selfridge [Er78], [Er86c] and van Doorn–Li–Tang [VLT26].
public theorem_wanted conjecture_holds : conjecture
def conjecture : Prop :=
(∀ m : ℕ, f m = min m ⌈2 * Real.sqrt m⌉₊) ∧
(¬ ∀ m : ℕ, (f m : ℝ) ≤ Real.sqrt m) ∧
(∀ m : ℕ, Real.sqrt m ≤ (f m : ℝ)) ∧
(∀ m : ℕ, f (m ^ 2) ≤ 2 * m)
WantedExt/NumberTheory/ErdosProblem650Wanted.lean:42
Sources
- [VLT26] W. van Doorn, Y. Li and Q. Tang, Optimal bounds for an Erdős problem on matching integers to distinct multiples, arXiv:2603.28636 (2026), https://arxiv.org/abs/2603.28636; W. van Doorn, Lean formalization of Erdős Problem 650 (Aristotle-assisted), https://github.com/Woett/Lean-files/blob/c100ed4a4429e987a9555ee428a1aa524a46e16d/ErdosProblem650.lean.
Uses 1
Loading dependency explorer…
The dependency explorer needs JavaScript and its graph data.