List-Coloring and Chromatic-Choosability -- A Dynamic Survey

Abstract: List-coloring, introduced independently by Vizing and by Erdős, Rubin, and Taylor in the 1970s, generalizes ordinary vertex coloring by assigning to each vertex its own set of admissible colors. A graph is chromatic-choosable if its list chromatic number equals its chromatic number. The previous survey on list-coloring by D R Woodall (2001), emphasized defective choosability, the list-coloring conjectures, and different methods used for list-coloring. This survey reviews major developments on list-coloring and chromatic-choosability, with emphasis on graph classes for which equality is known, graph classes exhibiting a nontrivial gap, and the principal methods used to prove such results. The survey covers embedded graphs, perfect graphs, complete bipartite and multipartite graphs, claw-free graphs, line graphs, powers of graphs, graph products, and selected variants of list-coloring.

• The paper provides a dynamic survey outlining foundational results and new trends in list coloring and chromatic-choosability.
• It details various algebraic, combinatorial, and probabilistic methods that refine classical bounds such as Δ+1 and degeneracy.
• The survey discusses structural and algorithmic implications, including open problems in claw-free graphs and power-colorings.

Survey of List-Coloring and Chromatic-Choosability

Introduction and Background

List coloring generalizes classical vertex coloring by requiring vertices to be colored from assigned lists, with the list chromatic number $\chi_\ell(G)$ reflecting the minimal list size for which proper colorings from any assignment are guaranteed. The chromatic-choosability problem focuses on graph classes where $\chi_\ell(G)=\chi(G)$—the so-called chromatic-choosable graphs. Significant early advances, notably by Erdős, Rubin, and Taylor, established both foundational results and strong negative examples: the jump $\chi_\ell(G)-\chi(G)$ can be arbitrarily large (notably for complete bipartite graphs).

This survey provides a comprehensive and up-to-date treatment of key developments, conjectures, and techniques in this active subfield, with a systematic focus on classes where equality or separation between $\chi_\ell$ and $\chi$ is understood, as well as the algebraic, combinatorial, and probabilistic methods that underpin these results.

Fundamental Upper Bounds and Methods

Several classical and modern upper bounds on $\chi_\ell(G)$ are emphasized:

• Maximum Degree ($\Delta$) Bounds: The greedy and Brooks-type bounds for coloring extend to list-coloring: $\chi_\ell(G) \leq \Delta(G) + 1$, with improvements for special structures.
• Coloring Number (col$(G)$) and Degeneracy: The coloring number and degeneracy provide upper bounds, with col$(G) = 1 +$max$\chi_\ell(G)=\chi(G)$0, yielding $\chi_\ell(G)=\chi(G)$1.
• Alon-Tarsi Number (AT(G)): The Alon-Tarsi polynomial-orientation method gives potentially sharper bounds, exploiting nonvanishing coefficients in the graph polynomial or nontrivial differences in spanning orientation counts. It refines the hierarchy: $\chi_\ell(G)=\chi(G)$2.

Auxiliary bounds, such as those involving the graph's complement, matchings, or new coloring parameters (paintability, DP-coloring), further extend the toolkit.

Embedded Graphs: Planar and Toroidal

In planar graphs, Thomassen's proof that every planar graph is 5-choosable resolved a principal conjecture via an induction strengthened at rigid boundaries. The existence of non-4-choosable (even 3-colorable) planar graphs (Voigt, Mirzakhani) highlights the strictness of the list coloring generalization. Structural restrictions—such as forbidden cycles or specific girth—admit tighter results. For instance:

• Planar graphs free of certain small cycles are 4-choosable ($\chi_\ell(G)=\chi(G)$3-free: Lam et al., $\chi_\ell(G)=\chi(G)$4/$\chi_\ell(G)=\chi(G)$5-free: Fijavž et al.).
• Girth constraints (e.g., girth $\chi_\ell(G)=\chi(G)$6) yield 3-choosability (Thomassen).

Toroidal graphs, structurally more complex, have similar but strictly weaker results. For toroidal grids $\chi_\ell(G)=\chi(G)$7, recent work confirms 4-choosability for all $\chi_\ell(G)=\chi(G)$8.

Perfect Graphs and Hall-type Techniques

Perfect graphs, particularly chordal and interval graphs, are shown to be chromatic-choosable, often via elimination ordering or the application of Hall’s SDR theorem. This extends to certain claw-free perfect graphs and subclasses, with specialized arguments (e.g., acyclic orientations for interval graphs ensuring $\chi_\ell(G)=\chi(G)$9).

Complete Bipartite and Multipartite Graphs

The detailed structure of $\chi_\ell(G)-\chi(G)$0 is charted for small $\chi_\ell(G)-\chi(G)$1, with precise thresholds for the transition from $\chi_\ell(G)-\chi(G)$2- to $\chi_\ell(G)-\chi(G)$3-choosability. Noteworthy is the sharp gap between $\chi_\ell(G)-\chi(G)$4 and $\chi_\ell(G)-\chi(G)$5 for large, balanced bipartite graphs: $\chi_\ell(G)-\chi(G)$6 for large $\chi_\ell(G)-\chi(G)$7.

The Ohba Conjecture (now a theorem due to Noel et al.), positing chromatic-choosability for graphs with at most $\chi_\ell(G)-\chi(G)$8 vertices, is a cornerstone result. Explicit extremal constructions (Zhu et al.) show the bound is tight.

Claw-Free Graphs and Line Graphs

Claw-free graphs present an ongoing challenge. While some bounds exist (e.g., $\chi_\ell(G)-\chi(G)$9: Chudnovsky and Seymour), the full List Coloring Conjecture (LCC) for claw-free graphs is unresolved. Partial positive results appear for subclasses—including elementary and peculiar graphs, as well as for claw-free perfect graphs with small clique number.

For line graphs, the List Edge Coloring Conjecture (LECC) is open in full generality, but proven for key classes: bipartite graphs (Galvin’s kernel method), 2-connected regular planar graphs, and others. The equivalence of edge coloring (in the original graph) and list coloring (in the line graph) underlies the interplay between these parameters.

Squares and Powers of Graphs

The List Square Coloring Conjecture (LSCC)—and its implications for total colorings—has been disproven by explicit algebraic counterexamples for both general and bipartite graphs (Kim & Park). Recent research focuses on upper bounds for $\chi_\ell$0 in planar graphs, with girth and degree constraints yielding best-known results. For subcubic planar graphs, improvements have drastically reduced the bound for $\chi_\ell$1 when girth is increased, and new work addresses forbidden cycle conditions.

Generic powers of graphs do not admit a universal $\chi_\ell$2 such that $\chi_\ell$3 is chromatic-choosable (Kim et al.), and constructions achieve $\chi_\ell$4 unbounded over various graph families.

Graph Products and Operations

List-coloring under graph operations is systematically studied:

• Join: Ohba provides conditions under which $\chi_\ell$5 is chromatic-choosable.
• Cartesian Product: Nontrivial upper and lower bounds (Borowiecki & Jozef), with explicit values for certain products (e.g., $\chi_\ell$6).
• Lexicographic Product: Recent work shows sublinear bounds in terms of $\chi_\ell$7 and $\chi_\ell$8 plus logarithmic factors, though sharpness and universality remain open.

Uniform inflations and lexicographic products equate in the context of chromatic-choosability for powers of cycles and related structures.

Regular Graphs and Variants

Specific families of regular graphs are addressed, with the Alon-Tarsi method yielding $\chi_\ell$9 for well-structured decompositions (e.g., certain 4-regular graphs).

Key generalizations—paintability (on-line list-coloring), DP-coloring (correspondence coloring), $\chi$0-choosability, fractional coloring—are discussed, with the survey noting disproved conjectures (e.g., Weak $\chi$1-Choosability Conjecture: Dvořák et al.) and significant separations between $\chi$2 and these stronger parameters.

Implications and Future Directions

The results mapped here reveal several deep and persistent phenomena:

• Sharp separation between chromatic number and list chromatic number in natural classes, as well as extreme instances with arbitrarily large jump.
• Structural influences such as girth, forbidden subgraphs, and degeneracy, which can restore chromatic-choosability or sharply reduce bounds on $\chi$3.
• Algebraic and combinatorial techniques (Alon-Tarsi, kernel-perfect orientations, Hall-type arguments) are central; new polynomials and orientation frameworks continue to be developed for challenging graph classes.

Major conjectures remain open in claw-free graphs, powers of graphs, and graph operations, with recent counterexamples refocusing speculation on parameter bounds and structural necessary/sufficient conditions. Theoretical advances (e.g., the Ohba conjecture resolution) have immediate algorithmic and complexity implications, as complexity for general list coloring remains hard, but is tractable under well-understood structural regimes.

Promising future research includes refining bounds for multipartite graphs; characterizing chromatic-choosability in broader structured classes (e.g., claw-free graphs); investigating algebraic bounds for powers and products; and extending variants such as DP-coloring and paintability to wider families of graphs.

Conclusion

This dynamic survey demonstrates both the breadth and the technical depth of the literature on list-coloring and chromatic-choosability. Techniques blend probabilistic, algebraic, and combinatorial approaches. Many boundaries in the theory are now sharply drawn by construction, while new structural and algorithmic frontiers remain open. The interplay between global structure and local constraints in coloring remains a central theme, positioning list-coloring as a touchstone in graph theory and combinatorics (2606.31702).