/-
Copyright 2026 The Formal Conjectures Authors.
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
https://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
-/
module
public import FormalConjecturesUtilThe linear arboricity conjecture
References:
[AEH81] Akiyama, J., Exoo, G., and Harary, F. (1981). "Covering and packing in graphs. IV. Linear arboricity." Networks 11, pp. 69--72.
[EP84] Enomoto, H. and Péroche, B. (1984). "The linear arboricity of some regular graphs." J. Graph Theory 8, pp. 309--324.
[Al88] Alon, N. (1988). "The linear arboricity of graphs." Israel J. Math. 62, pp. 311--325.
@[expose] public sectionnamespace LinearArboricityConjecturevariable {V : Type*}A linear forest is a graph whose connected components are paths: an acyclic graph in which every vertex has at most two neighbours.
def IsLinearForest (F : SimpleGraph V) : Prop :=
F.IsAcyclic ∧ ∀ v, (F.neighborSet v).ncard ≤ 2
G decomposes into k linear forests: there are k linear-forest subgraphs of G
such that every edge of G lies in exactly one of them.
def HasLinearForestDecomposition (G : SimpleGraph V) (k : ℕ) : Prop :=
∃ D : Fin k → SimpleGraph V, (∀ i, IsLinearForest (D i) ∧ D i ≤ G) ∧
∀ u v, G.Adj u v ↔ ∃! i, (D i).Adj u v
The linear arboricity la(G): the least number of linear forests into which the edges
of G can be partitioned.
noncomputable def linearArboricity (G : SimpleGraph V) : ℕ :=
sInf {k | HasLinearForestDecomposition G k}The empty graph is a linear forest.
right V:Type u_1v:Vh:⊥.neighborSet v = ∅⊢ ∅.ncard ≤ 2
simp All goals completed! 🐙The linear arboricity conjecture (Akiyama–Exoo–Harary [AEH81]).
The edges of every graph of maximum degree Δ can be partitioned into at most ⌈(Δ + 1) / 2⌉
linear forests.
@[category research open, AMS 5]
theorem linear_arboricity_conjecture {V : Type} [Fintype V] (G : SimpleGraph V)
[DecidableRel G.Adj] :
linearArboricity G ≤ (G.maxDegree + 2) / 2 := by V:Typeinst✝¹:Fintype VG:SimpleGraph Vinst✝:DecidableRel G.Adj⊢ linearArboricity G ≤ (G.maxDegree + 2) / 2
sorry All goals completed! 🐙The linear arboricity conjecture for regular graphs.
Every nonempty Δ-regular graph with 0 < Δ has linear arboricity exactly
⌈(Δ + 1) / 2⌉. Since a nonempty Δ-regular
graph is easily seen to have linear arboricity strictly greater than Δ / 2, this form of
the conjecture is equivalent to the general upper bound; see [AEH81].
@[category research open, AMS 5]
theorem linear_arboricity_conjecture.variants.regular {V : Type} [Fintype V] [Nonempty V]
(G : SimpleGraph V) [DecidableRel G.Adj] (Δ : ℕ) (hΔ : 0 < Δ)
(hreg : ∀ v, G.degree v = Δ) :
linearArboricity G = (Δ + 2) / 2 := by V:Typeinst✝²:Fintype Vinst✝¹:Nonempty VG:SimpleGraph Vinst✝:DecidableRel G.AdjΔ:ℕhΔ:0 < Δhreg:∀ (v : V), G.degree v = Δ⊢ linearArboricity G = (Δ + 2) / 2
sorry All goals completed! 🐙Cubic graphs have linear arboricity two (Akiyama–Exoo–Harary [AEH81]).
The edges of every nonempty 3-regular graph can be partitioned into two linear forests, and one
linear forest is not enough.
@[category research solved, AMS 5]
theorem linear_arboricity_conjecture.variants.cubic {V : Type} [Fintype V] [Nonempty V]
(G : SimpleGraph V) [DecidableRel G.Adj] (hreg : ∀ v, G.degree v = 3) :
linearArboricity G = 2 := by V:Typeinst✝²:Fintype Vinst✝¹:Nonempty VG:SimpleGraph Vinst✝:DecidableRel G.Adjhreg:∀ (v : V), G.degree v = 3⊢ linearArboricity G = 2
sorry All goals completed! 🐙Alon (1988): the linear arboricity conjecture holds asymptotically.
For every ε > 0 there is a Δ₀ such that every graph of maximum degree Δ ≥ Δ₀ has linear
arboricity at most (1 / 2 + ε) Δ.
Reference: [Al88].
@[category research solved, AMS 5]
theorem linear_arboricity_conjecture.variants.alon (ε : ℝ) (hε : 0 < ε) :
∃ Δ₀ : ℕ, ∀ Δ : ℕ, Δ₀ ≤ Δ → ∀ (V : Type) [Fintype V] (G : SimpleGraph V),
∀ [DecidableRel G.Adj], G.maxDegree ≤ Δ →
(linearArboricity G : ℝ) ≤ (1 / 2 + ε) * Δ := by ε:ℝhε:0 < ε⊢ ∃ Δ₀,
∀ (Δ : ℕ),
Δ₀ ≤ Δ →
∀ (V : Type) [inst : Fintype V] (G : SimpleGraph V) [inst_1 : DecidableRel G.Adj],
G.maxDegree ≤ Δ → ↑(linearArboricity G) ≤ (1 / 2 + ε) * ↑Δ
sorry All goals completed! 🐙The empty graph has linear arboricity zero.
The empty partition witnesses la(⊥) = 0.
@[category test, AMS 5]
theorem linear_arboricity_conjecture.variants.bot :
linearArboricity (⊥ : SimpleGraph V) = 0 := by V:Type u_1⊢ linearArboricity ⊥ = 0
have h : HasLinearForestDecomposition (⊥ : SimpleGraph V) 0 :=
⟨fun i => i.elim0, fun i => i.elim0,
fun u v => iff_of_false (fun h => h) fun ⟨i, _⟩ => i.elim0⟩ V:Type u_1h:HasLinearForestDecomposition ⊥ 0⊢ linearArboricity ⊥ = 0
exact Nat.le_zero.mp (Nat.sInf_le h) All goals completed! 🐙end LinearArboricityConjecture