Universal cycles for k-subsets
DefinitionUniversal cycle for \(k\)-subsets#✓fully proved
For natural numbers \(n\) and \(k\), let \(N = \binom{n}{k}\). A sequence \(c_0, \dots, c_{N-1}\) of elements of \(\{0, \dots, n-1\}\), read cyclically, is a universal cycle for the \(k\)-subsets of \(\{0, \dots, n-1\}\) when, for every starting position \(i < N\), the \(k\) consecutive entries \(c_i, c_{i+1}, \dots, c_{i+k-1}\) (indices taken modulo \(N\)) are pairwise distinct, and every \(k\)-element subset of \(\{0, \dots, n-1\}\) is the set of these entries for exactly one starting position \(i\).
def IsUniversalCycle (n k : ℕ) (c : Fin (Nat.choose n k) → Fin n) : Prop :=
∀ hpos : 0 < Nat.choose n k,
(∀ i : Fin (Nat.choose n k),
((Finset.univ : Finset (Fin k)).image (fun j : Fin k =>
c ⟨(i.val + j.val) % Nat.choose n k,
Nat.mod_lt (i.val + j.val) hpos⟩)).card = k)
∧ ∀ S : Finset (Fin n), S.card = k →
∃! i : Fin (Nat.choose n k),
(Finset.univ : Finset (Fin k)).image (fun j : Fin k =>
c ⟨(i.val + j.val) % Nat.choose n k,
Nat.mod_lt (i.val + j.val) hpos⟩) = S
Sources
- Chung/Diaconis/Graham as cited in Cooper, Combinatorial Problems I Like, Words and Codes bullet 78 [AMR-030-0078].
Used by 1
DefinitionChung–Diaconis–Graham universal cycle conjecture#✓fully proved
The statement that for every natural number \(k\) there is a natural number \(N\) such that every \(n \ge N\) for which \(k\) divides \(\binom{n-1}{k-1}\) admits a universal cycle for the \(k\)-subsets of \(\{0, \dots, n-1\}\). Here \(n - 1\) and \(k - 1\) are truncated at \(0\), so for \(k = 0\) the divisibility condition reads \(0 \mid 1\) and never holds.
def conjecture : Prop :=
∀ k : ℕ, ∃ N : ℕ, ∀ n : ℕ, N ≤ n →
k ∣ Nat.choose (n - 1) (k - 1) →
∃ c : Fin (Nat.choose n k) → Fin n, IsUniversalCycle n k c
def IsUniversalCycle (n k : ℕ) (c : Fin (Nat.choose n k) → Fin n) : Prop :=
∀ hpos : 0 < Nat.choose n k,
(∀ i : Fin (Nat.choose n k),
((Finset.univ : Finset (Fin k)).image (fun j : Fin k =>
c ⟨(i.val + j.val) % Nat.choose n k,
Nat.mod_lt (i.val + j.val) hpos⟩)).card = k)
∧ ∀ S : Finset (Fin n), S.card = k →
∃! i : Fin (Nat.choose n k),
(Finset.univ : Finset (Fin k)).image (fun j : Fin k =>
c ⟨(i.val + j.val) % Nat.choose n k,
Nat.mod_lt (i.val + j.val) hpos⟩) = S
WantedExt/Combinatorics/UniversalCycleSubsetsWanted.lean:43
Sources
- Chung/Diaconis/Graham as cited in Cooper, Combinatorial Problems I Like, Words and Codes bullet 78 [AMR-030-0078].
Uses 1 · Used by 1
TheoremExistence of universal cycles for \(k\)-subsets#●ready to prove
For every natural number \(k\) there is a natural number \(N\) such that for every \(n \ge N\) for which \(k\) divides \(\binom{n-1}{k-1}\), there is a universal cycle for the \(k\)-subsets of \(\{0, \dots, n-1\}\). Here \(n - 1\) and \(k - 1\) are truncated at \(0\), so for \(k = 0\) the divisibility condition reads \(0 \mid 1\) and never holds.
This is the Chung–Diaconis–Graham conjecture. Glock, Joos, Kühn and Osthus (Theorem 1.1) proved it for every \(k \ge 2\) in the language of tight Euler tours of the complete \(k\)-uniform hypergraph; the case \(k = 1\) is elementary.
public theorem_wanted conjecture_holds : conjecture
def conjecture : Prop :=
∀ k : ℕ, ∃ N : ℕ, ∀ n : ℕ, N ≤ n →
k ∣ Nat.choose (n - 1) (k - 1) →
∃ c : Fin (Nat.choose n k) → Fin n, IsUniversalCycle n k c
WantedExt/Combinatorics/UniversalCycleSubsetsWanted.lean:59
Sources
- Stefan Glock, Felix Joos, Daniela Kühn, Deryk Osthus, Euler tours in hypergraphs, Combinatorica 40 (2020), 679-690; arXiv:1808.07720, https://arxiv.org/abs/1808.07720.
Uses 1
Loading dependency explorer…
The dependency explorer needs JavaScript and its graph data.