Total Coloring Conjecture

• Total Coloring Conjecture is the hypothesis that any simple graph can be totally colored with at most Δ+2 colors, ensuring adjacent vertices, edges, and incident vertex–edge pairs receive distinct colors.
• Researchers have applied methods such as discharging, auxiliary graph transformations, and matching-based arguments to verify the conjecture for various graph classes, including planar, dense, and subcubic graphs.
• Despite many partial confirmations and asymptotic results, determining the exact total chromatic number remains NP-hard and unresolved for general graphs.

The Total Coloring Conjecture (TCC), independently due to Behzad and Vizing, asserts that every simple graph $G$ admits a proper coloring of $V(G)\cup E(G)$ with at most $\Delta(G)+2$ colors, where adjacent vertices, adjacent edges, and incident vertex–edge pairs receive distinct colors. The minimum such number is the total chromatic number, usually denoted $\chi''(G)$, although some papers write $\chi_T(G)$. Since every graph satisfies the lower bound $\chi''(G)\ge \Delta(G)+1$, the conjecture predicts the sharp two-point range $\chi''(G)\in\{\Delta(G)+1,\Delta(G)+2\}$ (Jayabalan et al., 2018).

1. Definitions, notation, and equivalent formulations

A total coloring of a simple graph $G$ assigns colors to both vertices and edges so that adjacent vertices receive different colors, adjacent edges receive different colors, and each edge receives a color distinct from the colors of its endpoints. The minimum number of colors in such a coloring is the total chromatic number $\chi''(G)$ (Jayabalan et al., 2018). In the notation of some recent dense-graph papers, the same parameter is written $\chi_T(G)$ (Dalal et al., 2024).

The conjectural bound is

$V(G)\cup E(G)$0

Combined with the basic lower bound $V(G)\cup E(G)$1, this yields the standard dichotomy between type-I graphs, for which $V(G)\cup E(G)$2, and type-II graphs, for which $V(G)\cup E(G)$3 (Jayabalan et al., 2018). This classification is central in the literature because many class-specific results actually determine whether a family is type I or type II, rather than merely proving the conjectured upper bound.

A particularly useful reformulation uses the total graph $V(G)\cup E(G)$4, whose vertex set is $V(G)\cup E(G)$5, with adjacency encoding either adjacency or incidence in $V(G)\cup E(G)$6. Under this construction,

$V(G)\cup E(G)$7

so total coloring becomes ordinary vertex coloring of $V(G)\cup E(G)$8 (Kaul et al., 2018, Basavaraju et al., 2021). This viewpoint links TCC to equitable coloring, list coloring, and graph-minor questions on total graphs.

For multigraphs, the Behzad–Vizing formulation becomes

$V(G)\cup E(G)$9

where $\Delta(G)+2$0 is the multiplicity. A stronger formulation noted in the literature is $\Delta(G)+2$1, where $\Delta(G)+2$2 is the chromatic index (Cao et al., 2021). The same paper also recalls Goldberg’s conjecture, which predicts

$\Delta(G)+2$3

and proves, assuming the Goldberg–Seymour conjecture, that

$\Delta(G)+2$4

for graphs in that regime (Cao et al., 2021).

2. Difficulty, exactness, and computational status

Total coloring is structurally harder than either vertex coloring or edge coloring in isolation because it imposes both constraints simultaneously. The survey literature explicitly records that finding a minimum total coloring is NP-hard in general and even NP-complete for some restricted classes such as cubic bipartite graphs (Jayabalan et al., 2018). This computational hardness persists even when the conjectural upper bound is known: the challenge often lies in determining whether the graph is type I or type II.

The subcubic case illustrates this distinction sharply. For graphs with $\Delta(G)+2$5, TCC specializes to $\Delta(G)+2$6, and the literature cited in the Halin-graph work states that this case was verified by Rosenfeld and independently by Vijayaditya; thus every subcubic graph is totally 5-colorable (Kardoš et al., 24 Mar 2026). Yet deciding whether only four colors suffice remains difficult, and the same paper notes that this decision problem is hard even for cubic bipartite graphs (Kardoš et al., 24 Mar 2026).

The conjecture remains unresolved in full generality. The survey emphasizes that it is not settled even for planar graphs, despite a long sequence of partial theorems, especially for graphs with forbidden local configurations or large maximum degree (Jayabalan et al., 2018). This persistent incompleteness is one reason total coloring continues to function both as a structural conjecture and as an algorithmic benchmark.

3. Established graph classes and exact theorems

A large part of the modern literature verifies TCC on restricted families, often by proving the stronger equality $\Delta(G)+2$7. For planar-type classes, the dominant techniques are discharging and reducible configurations; for product graphs, Cayley graphs, and dense families, constructive colorings, equitable decompositions, and matching-based arguments are more typical.

The survey records that TCC is known for planar graphs with $\Delta(G)+2$8, that every planar graph with $\Delta(G)+2$9 is 9-total-colorable, that TCC was verified for planar graphs with $\chi''(G)$0, and that planar graphs with $\chi''(G)$1 are type I (Jayabalan et al., 2018). More recent work narrows the unresolved planar frontier further by handling specific $\chi''(G)$2 configurations (Su et al., 17 Jul 2025).

Graph class	Conclusion	Source
Pseudo-outerplanar graphs	TCC holds for all pseudo-outerplanar graphs; if $\chi''(G)$3, then $\chi''(G)$4	(Zhang et al., 2011)
1-toroidal graphs with $\chi''(G)$5 and no adjacent triangles	$\chi''(G)$6	(Wang, 2012)
Planar graphs with $\chi''(G)$7 and no mushroom, tent, or cone	Totally 8-colorable	(Su et al., 17 Jul 2025)
Coronas of cubic graphs	$\chi''(G)$8	(Furmańczyk et al., 2015)
Selected Cayley graph families	Exact type-I and type-II results; perfect Cayley graphs under stated hypotheses satisfy TCC	(S et al., 2020)
Cubic Halin graphs	All cubic Halin graphs except $\chi''(G)$9 satisfy $\chi_T(G)$0	(Kardoš et al., 24 Mar 2026)
Subcubic Halin graphs	All subcubic Halin graphs except $\chi_T(G)$1 satisfy $\chi_T(G)$2	(Kardoš et al., 24 Mar 2026)

These results are methodologically diverse. For pseudo-outerplanar graphs, the proof proceeds by induction on $\chi_T(G)$3 combined with a structural lemma that forces one of four reducible configurations (Zhang et al., 2011). For 1-toroidal graphs, the proof uses a minimal-counterexample setup, passage to an associated embedded graph, and a detailed discharging argument with semi-fans (Wang, 2012). For cubic and subcubic Halin graphs, the problem is reduced to finite palette computations on Halin tripoles, yielding a linear-time algorithm and finite exception lists (Kardoš et al., 24 Mar 2026).

4. Dense graphs and asymptotic confirmations

Recent progress has concentrated on dense graphs, where global degree hypotheses permit auxiliary-graph reductions and controlled edge-coloring constructions. A 2024 result proves the general bound

$\chi_T(G)$4

for every graph $\chi_T(G)$5. As an immediate corollary, if $\chi_T(G)$6, then

$\chi_T(G)$7

The same paper proves a sharp asymptotic theorem for regular graphs: for every $\chi_T(G)$8, there exists $\chi_T(G)$9 such that every $\chi''(G)\ge \Delta(G)+1$0-regular graph on $\chi''(G)\ge \Delta(G)+1$1 vertices with

$\chi''(G)\ge \Delta(G)+1$2

satisfies

$\chi''(G)\ge \Delta(G)+1$3

thereby confirming TCC for sufficiently large dense regular graphs (Dalal et al., 2024).

The proof architecture in that paper combines equitable $\chi''(G)\ge \Delta(G)+1$4-vertex-coloring via Hajnal–Szemerédi, auxiliary hypergraph edge-coloring, matching arguments in the complement, and multifan or alternating-path recolorings in the style of Vizing’s theorem (Dalal et al., 2024). The dense regular theorem was notable because it lowered the density threshold from earlier results on much denser regular graphs to just above $\chi''(G)\ge \Delta(G)+1$5.

A 2025 paper strengthens this line by replacing the regularity assumption with a minimum-degree hypothesis. It proves that for every $\chi''(G)\ge \Delta(G)+1$6, there exists $\chi''(G)\ge \Delta(G)+1$7 such that every graph $\chi''(G)\ge \Delta(G)+1$8 on $\chi''(G)\ge \Delta(G)+1$9 vertices with

$\chi''(G)\in\{\Delta(G)+1,\Delta(G)+2\}$0

satisfies

$\chi''(G)\in\{\Delta(G)+1,\Delta(G)+2\}$1

Thus every sufficiently large graph with minimum degree slightly more than $\chi''(G)\in\{\Delta(G)+1,\Delta(G)+2\}$2 satisfies TCC (Henderschedt et al., 8 Jul 2025). The paper explicitly states that this extends prior work in two ways: it replaces a maximum-degree-based hypothesis by a minimum-degree hypothesis, and it removes the requirement that the graph be regular (Henderschedt et al., 8 Jul 2025).

Its main reduction introduces a matching $\chi''(G)\in\{\Delta(G)+1,\Delta(G)+2\}$3 in the complement $\chi''(G)\in\{\Delta(G)+1,\Delta(G)+2\}$4 and an auxiliary graph $\chi''(G)\in\{\Delta(G)+1,\Delta(G)+2\}$5 obtained by adding a new vertex $\chi''(G)\in\{\Delta(G)+1,\Delta(G)+2\}$6, the edges of $\chi''(G)\in\{\Delta(G)+1,\Delta(G)+2\}$7, and all edges from $\chi''(G)\in\{\Delta(G)+1,\Delta(G)+2\}$8 to vertices not covered by $\chi''(G)\in\{\Delta(G)+1,\Delta(G)+2\}$9. A “good edge-coloring” of $G$0, in which all edges of $G$1 receive distinct colors, immediately yields a total coloring of $G$2 by transferring those edge colors to the corresponding vertices (Henderschedt et al., 8 Jul 2025). The coloring construction uses equitable colorings, decompositions into bipartite subgraphs, multigraph edge-coloring lemmas, and Kempe-change-style recolorings (Henderschedt et al., 8 Jul 2025).

5. Total graphs, weakenings, and strengthened variants

Because $G$3, many refinements of TCC are naturally phrased on total graphs. One such direction is equitable total coloring. Fu’s conjecture, in total-graph language, asserts that $G$4 should be equitably $G$5-colorable whenever

$G$6

A list-coloring analogue asks for equitable $G$7-choosability under

$G$8

This conjecture was proved for all graphs with $G$9, using the fact that total graphs in that range decompose into squares of paths, squares of cycles, and $\chi''(G)$0 components (Kaul et al., 2018).

A second direction is the weak TCC

$\chi''(G)$1

This weakening is strong enough to imply a graph-minor statement: if a class $\chi''(G)$2 is closed under taking subgraphs and weak TCC holds for every graph in $\chi''(G)$3, then Hadwiger’s conjecture holds for the class of total graphs $\chi''(G)$4 (Basavaraju et al., 2021). The same paper proves that every 5-colorable graph satisfies weak TCC,

$\chi''(G)$5

thereby extending earlier easy cases for 4-colorable graphs (Basavaraju et al., 2021).

A third direction strengthens properness by requiring adjacent vertices to be distinguishable by their local color sets. The AVD total coloring conjecture asks for

$\chi''(G)$6

This conjecture has been verified for several broad classes in recent work, including 3-degenerate graphs (Behera et al., 5 Aug 2025), many corona products (Furmańczyk et al., 2022), and several families of central graphs, subdivision graphs, and joins (Banerjee, 2 Jul 2026). These results are not equivalent to TCC, but they show that the $\chi''(G)$7 threshold is natural in strengthened total-coloring settings.

6. Present status, bibliographic tension, and unresolved directions

The dominant status signal in the recent literature is that TCC remains open in full generality. The survey states explicitly that it is not settled even for planar graphs (Jayabalan et al., 2018), and the 2025 planar paper identifies the remaining open $\chi''(G)$8 frontier through three special substructures—mushroom, tent, and cone—whose exclusion suffices for total 8-colorability (Su et al., 17 Jul 2025). Likewise, the 2025 dense minimum-degree paper presents its theorem as a partial confirmation for a large asymptotic class rather than a resolution of the general conjecture (Henderschedt et al., 8 Jul 2025).

This also clarifies a common misunderstanding: TCC does not claim that every graph is type I. Type-II graphs are explicitly allowed, and many exact results are formulated precisely as classifications into $\chi''(G)$9 versus $\chi_T(G)$0 behavior (Jayabalan et al., 2018). Another recurrent misunderstanding is that dense-graph confirmations solve the universal problem; the recent dense results require either regularity plus $\chi_T(G)$1 (Dalal et al., 2024) or minimum degree $\chi_T(G)$2 for sufficiently large graphs (Henderschedt et al., 8 Jul 2025).

There is also a bibliographic tension in the corpus. A 2020 preprint explicitly claims “A proof of the Total Coloring Conjecture” and states that it settles the conjecture by an algebraic method over $\chi_T(G)$3 (Murthy, 2020). However, subsequent papers in 2021, 2025, and 2026 continue to treat TCC as open in general and to develop partial confirmations for specific classes or asymptotic regimes (Basavaraju et al., 2021, Henderschedt et al., 8 Jul 2025, Su et al., 17 Jul 2025). The record represented here therefore documents an unresolved conjectural landscape together with a claimed proof that has not displaced the open-problem framing in later literature.

From the standpoint of current research directions represented in these papers, three fronts stand out. The first is the planar $\chi_T(G)$4 case, where progressively weaker forbidden-configuration hypotheses have been established but the full case remains open (Jayabalan et al., 2018, Su et al., 17 Jul 2025). The second is the asymptotic dense regime, where the transition from large maximum degree to large minimum degree has already been achieved (Dalal et al., 2024, Henderschedt et al., 8 Jul 2025). The third is the interface with related theories—equitable, list, and adjacent-vertex-distinguishing total coloring—where the total-graph viewpoint continues to generate sharper conjectures and transferable methods (Kaul et al., 2018, Behera et al., 5 Aug 2025).