Overfull Conjecture in Graph Theory

• Overfull Conjecture is a hypothesis stating that a graph’s chromatic index equals its maximum degree if and only if the graph contains no overfull subgraph.
• It has motivated diverse methods including randomized vertex partitioning, equitable Δ-colorings, and structural subgraph analysis to develop efficient edge-coloring algorithms.
• The conjecture bridges classical edge-coloring problems with modern algorithmic combinatorics, offering key insights into dense graph regimes and complexity boundaries.

The Overfull Conjecture is a central hypothesis in structural and algorithmic graph theory, positing a natural and sharp criterion for determining when the chromatic index of a graph equals its maximum degree. It unifies concepts from edge colorings, structural graph theory, and complexity, and has motivated diverse methodologies in extremal combinatorics, algebraic methods, and algorithmic graph theory.

1. Statement and Formulation

Let $G$ be a finite simple graph with maximum degree $\Delta(G)$ and chromatic index $\chi'(G)$, the minimum number of colors required for a proper edge-coloring. A subgraph $H \subseteq G$ is called \emph{overfull} (with respect to $G$) if

$|E(H)| > \Delta(G) \lfloor |V(H)| / 2 \rfloor$

since each color class in a proper $\Delta(G)$-edge-coloring is a matching of at most $\lfloor |V(H)|/2 \rfloor$ edges.

Chetwynd–Hilton Overfull Conjecture (1986):

If

$\Delta(G) > \frac{|V(G)|}{3},$

then

$$\chi'(G) = \Delta(G) \iff G \text{ contains no overfull subgraph.}$$

Equivalently, for graphs whose maximum degree exceeds one-third the order, the existence of an overfull subgraph is the only obstruction to being Class 1 (i.e., being $\Delta$-edge-colorable) (Plantholt et al., 2021).

2. Foundations and Connections

The Overfull Conjecture generalizes the classical 1-factorization conjecture: every regular $n$-vertex graph (even $n$), degree at least $n/2$, admits a 1-factorization (and hence $\chi'(G) = \Delta(G)$). If $G$ is regular with $\Delta \geq n/2$, it contains no overfull subgraph, reducing the Overfull Conjecture precisely to the 1-factorization case for dense regular graphs (Plantholt et al., 2021). The conjecture also connects to various edge-coloring conjectures, including Vizing's Average Degree Conjecture and the Just Overfull and Vertex-Splitting conjectures (Qi et al., 7 Dec 2025).

3. Confirmed Regimes and Algorithmic Implications

The Overfull Conjecture is confirmed under several density and structure constraints:

• High minimum degree: For every $0 < \varepsilon < 1$, there exists $n_0$ so that if $G$ is a simple graph on $2n \geq n_0$ vertices with $\delta(G) \geq (1+\varepsilon)n$, then $\chi'(G) = \Delta(G)$ iff $G$ contains no overfull subgraph; an explicit polynomial-time coloring algorithm is provided (Plantholt et al., 2021).
• Dense quasirandom graphs: If $G$ is lower-$(p,\varepsilon)$-regular with $\Delta(G) - \delta(G) \leq \varepsilon n$, $\chi'(G) = \Delta(G)$ exactly when there is no overfull subgraph, for both even and odd $n$ (Shan, 2021).
• Graphs with large maximum degree (no $\delta$ constraint): If $\Delta(G) \geq (1-\varepsilon)n$ for any $0 < \varepsilon \leq 1/14$ and $n \geq n_0(\varepsilon)$, then Overfull Conjecture holds (Shan, 2023). Similarly, for graphs of odd order and minimum degree at least $(1+\varepsilon)n$, the equivalence holds and a polynomial-time algorithm exists (Shan, 2022).
• Split-comparability graphs: The Overfull Conjecture is established for split-comparability graphs, with an explicit structural characterization (neighborhood-overfullness) as the precise obstruction (Cruz et al., 2017).

These results yield efficient (polynomial-time) algorithms for computing chromatic index and exhibiting edge-colorings in the above settings, contrasting with the NP-completeness of determining edge colorings in general (Plantholt et al., 2021, Shan, 2023).

4. Techniques and Methodologies

The most advanced proofs for high-density or structural classes proceed by a combination of:

• Randomized vertex partitioning (Chernoff bounds) to achieve regularization and balanced bipartition (Plantholt et al., 2021, Shan, 2021).
• Equitable $\Delta$-colorings using Vizing–Gupta and matching extension methods, key for constructing color classes with controlled deficiencies (Plantholt et al., 2021).
• Alternating path augmentations and Kempe-chain arguments to transform partial colorings into full matchings or 1-factors (Plantholt et al., 2021, Shan, 2021).
• Structural subgraph analysis: The introduction of forbidden subconfigurations (“kites,” “forks,” “short brooms”) which cannot appear in a $\Delta$-critical graph under certain degree constraints (Cao et al., 2021, Chen et al., 8 Dec 2025).
• Core-degree and elementary set methods: Especially for graphs with small minimal core-degree ($\delta(G_{\Delta})$), elementarity arguments and extended Vizing-fan constructions force overfullness (Cao et al., 2022).
• Decomposition and absorption: For dense regular/multigraph settings, edge decompositions facilitated by factorization theorems are essential (Plantholt et al., 2023).

A typical proof involves reducing the graph to a near-regular or nearly bipartite structure, then applying matching theory (König, Hall, Dirac), plus parity and elementarity constraints on missing colors.

5. Partial Progress and Improved Thresholds

Key advances towards relaxing degree constraints have occurred:

• Linear trade-off results: Any $\Delta$-critical $G$ is overfull if $$\Delta(G) - \tfrac{7}{4}\delta(G) \geq \frac{3n-17}{4}$$; more recently, this was improved to $$\Delta(G) - \tfrac{5}{3}\delta(G) \geq \frac{2n-7}{3}$$, approaching the conjecture's linear threshold (Cao et al., 2021, Qi et al., 7 Dec 2025).
• Short-broom and fork analysis: For $\Delta$-critical $G$, if $\Delta \geq \tfrac{2n+5\delta-12}{3}$, then $G$ is overfull (Chen et al., 8 Dec 2025).
• Core degree results: If $G$ is critical, $\Delta(G) \geq \frac{2}{3}n + \frac{3k}{2}$, and $\delta(G_{\Delta}) \leq k$, then $G$ is overfull, showing the power of structural parameters beyond just global degrees (Cao et al., 2022).
• Graphs with special substructures: For $\Delta$-critical graphs with a vertex of degree 2 and $\Delta \geq 0.75n$, overfullness is guaranteed, strengthening the conjecture for graphs near this extremal configuration (Cao et al., 2020).

Recent work systematically reduces the allowed minimum degree and linear gaps, primarily via increasingly sophisticated local structure analysis, forbidden induced subgraphs, and combinatorial counting (Qi et al., 7 Dec 2025, Chen et al., 8 Dec 2025).

6. Extensions, Generalizations, and Open Problems

• Multigraph Overfull Conjecture: Generalized for multigraphs with maximum multiplicity $r$, where the threshold is $\Delta > \frac{1}{3}r|V(G)|$. Asymptotic results confirm the conjecture in several regimes for large even order, via decomposition to simple graphs and regularization lemmas (Plantholt et al., 2023).
• Algorithmic complexity: For graphs in the conjecture's regime ($\Delta > n/3$), verifying overfullness (and hence determining the chromatic index) is in P due to Seymour's matching polytope results. Thus, a full resolution would produce efficient chromatic index algorithms for all graphs above the threshold (Shan, 2023).
• Sparseness and tight bounds: For $\Delta$ just above $n/3$ or when $\delta$ is small, the conjecture remains open. Constructing corresponding extremal families that are not overfull yet have $\Delta$-chromatic index, or identifying finer forbidden configurations, is the principal challenge (Chen et al., 8 Dec 2025).
• Special graph classes: Complete solutions exist for split-comparability graphs, due to their explicit neighborhood-overfull characterization (Cruz et al., 2017). Similar full characterizations for broader classes (e.g., general chordal, comparability graphs) remain to be explored.
• Tighter combinatorial bounds and forbidden configurations: Continuous reduction of the gap between current partial results and the conjectured $\Delta > n/3$ remains a central focus. The development of new structural combinatorial gadgets (short brooms, branches, forks, kites) has been especially fruitful (Qi et al., 7 Dec 2025, Chen et al., 8 Dec 2025).

7. Summary Table of Key Results

Regime / Condition	Main Result	Reference
$\delta(G) > (1+\epsilon)n/2$ (large min degree)	Overfull Conjecture holds, polytime algorithm	(Plantholt et al., 2021, Shan, 2022)
$\Delta(G) \geq (1 - \epsilon)n$	Overfull Conjecture holds	(Shan, 2023)
Dense quasirandom graphs	Overfull Conjecture holds, polytime algorithm	(Shan, 2021)
$\Delta(G), \delta(G)$ linear bounds (e.g., $\Delta(G) - (7/4)\delta(G) \geq (3n-17)/4$)	Overfullness guaranteed	(Cao et al., 2021)
Core-degree condition ($\delta(G_\Delta) \leq k$, $\Delta(G) \geq (2/3)n + (3k/2)$)	Overfullness guaranteed	(Cao et al., 2022)
Split-comparability graphs	Overfull Conjecture holds	(Cruz et al., 2017)
Multigraphs, high degree/multiplicity	Overfull Conjecture holds or polytime coloring	(Plantholt et al., 2023)

The Overfull Conjecture synthesizes deep structural edges in extremal graph theory with algorithmic edge-coloring, providing a natural dividing line between tractable and intractable chromatic index computation in dense graphs. Continuing progress is driven by refined combinatorial structures (multi-fans, short brooms, branches), density-based decomposition, and enhanced local-global degree interplay. The general case, especially in sparseness or with irregular degree distributions, remains unresolved and is a subject of ongoing research.