ε-flexibility of d-degenerate graphs with (d+1)-lists
Problem 1 · arXiv:1612.08698
Status partial high confidence
Problem 1 remains open in full generality: no paper has proved epsilon-flexibility for all d-degenerate graphs with (d+1)-lists for every d >= 2. Substantial partial progress exists: the d=2 case has been confirmed for planar graphs of girth >= 6 (lists of size 3) and for all graphs of maximum average degree < 3; the d=3 case has been confirmed for triangle-free planar graphs (lists of size 4); and epsilon-flexibility has been established for graphs of treewidth 2, treedepth k, and bipartite d-degenerate graphs. The general conjecture for arbitrary d-degenerate graphs with (d+1)-lists remains open for d >= 2.
Cited literature (5)
-
Proves that triangle-free planar graphs (which are 3-degenerate) with lists of size at least four are weighted epsilon-flexible for some constant epsilon > 0, resolving the conjecture for this class.
-
Proves that planar graphs of girth at least six (which are 2-degenerate) with lists of size three are weighted epsilon-flexible, resolving the d=2 case for this subclass.
-
Proves epsilon-flexibility for graphs of treewidth 2 (which are (1/3)-flexibly 3-choosable), treedepth k graphs, and characterizes which maximum-degree-Delta graphs are epsilon-flexibly Delta-choosable.
-
Extends flexibility results to bipartite d-degenerate graphs and introduces the list flexibility number, without resolving the full conjecture for arbitrary d-degenerate graphs.
-
Proves that every graph with maximum average degree less than 3 is epsilon-flexibly 3-choosable, resolving the d=2 case of the conjecture for a broad class of 2-degenerate graphs (including planar graphs of girth >= 6).
Reviewer notes. The conjecture splits into two sub-questions (epsilon-flexible vs. weighted epsilon-flexible); partial results address both variants. The d=1 case (forests with 2-lists) was settled in the original paper. For d=2, several major subclasses are now resolved including maximum average degree < 3 (Bi-Bradshaw 2023). For d=3, triangle-free planar graphs are settled. The full conjecture for arbitrary d-degenerate graphs remains open for d >= 2.
Context
Theorem 2 shows that lists of size $d+2$ suffice for weighted $\varepsilon$-flexibility of $d$-degenerate graphs. The question is whether one extra color (size $d+1$, matching choosability) also guarantees flexibility. For $d=0$ trivial; for $d=1$ (forests with lists of size 2) weighted $1/2$-flexibility is easy; the first open case is $d=2$.
Also stated in
- List coloring with requests (2018-11-17)
- List coloring with requests (2018-11-17)
- List coloring with requests (2018-11-17)
- List coloring with requests (2018-11-17)
- List coloring with requests (2018-11-17)
Notes. PDF source; math is simple enough to read cleanly.
Source paper
List coloring with requests
Zdeněk Dvořák, Sergey Norin, Luke Postle · 2018-11-17
https://arxiv.org/abs/1612.08698
PDF source