ConceptComplete

Perfect Graphs and Advanced Coloring

Perfect graphs form a rich class where the chromatic number equals the clique number for every induced subgraph. The Strong Perfect Graph Theorem, proved in 2006, gives a complete characterization.


Perfect Graphs

Definition5.6Perfect Graph

A graph GG is perfect if ฯ‡(H)=ฯ‰(H)\chi(H) = \omega(H) for every induced subgraph HH of GG. Equivalently (by the Perfect Graph Theorem of Lovรกsz, 1972): GG is perfect if and only if its complement Gห‰\bar{G} is perfect. Examples: bipartite graphs, chordal graphs, comparability graphs, co-comparability graphs.

Definition5.7Strong Perfect Graph Theorem

A graph is perfect if and only if it contains no odd hole (C2k+1C_{2k+1} with kโ‰ฅ2k \geq 2) and no odd antihole (C2k+1โ€พ\overline{C_{2k+1}} with kโ‰ฅ2k \geq 2) as an induced subgraph. This was conjectured by Berge (1961) and proved by Chudnovsky, Robertson, Seymour, and Thomas (2006). The proof is over 150 pages and uses a structural decomposition theory for Berge graphs.


Fractional and Circular Coloring

ExampleFractional Chromatic Number

The fractional chromatic number ฯ‡f(G)\chi_f(G) is the linear programming relaxation of ฯ‡(G)\chi(G): minimize โˆ‘IxI\sum_I x_I over independent sets II subject to โˆ‘Iโˆ‹vxIโ‰ฅ1\sum_{I \ni v} x_I \geq 1 for all vv, xIโ‰ฅ0x_I \geq 0. Always ฯ‰(G)โ‰คฯ‡f(G)โ‰คฯ‡(G)\omega(G) \leq \chi_f(G) \leq \chi(G), and ฯ‡f=ฯ‰\chi_f = \omega for perfect graphs. For the Kneser graph: ฯ‡f(K(n,k))=n/k\chi_f(K(n,k)) = n/k, demonstrating that ฯ‡f\chi_f and ฯ‡\chi can differ significantly.

RemarkCircular Chromatic Number

The circular chromatic number ฯ‡c(G)\chi_c(G) generalizes ฯ‡(G)\chi(G): it is the infimum of p/qp/q over integers pโ‰ฅ2qp \geq 2q such that GG has a (p,q)(p,q)-coloring (vertices get arcs of length qq on a circle of circumference pp, with adjacent arcs non-overlapping). Always ฯ‡(G)โˆ’1<ฯ‡c(G)โ‰คฯ‡(G)\chi(G) - 1 < \chi_c(G) \leq \chi(G), and ฯ‡(G)=โŒˆฯ‡c(G)โŒ‰\chi(G) = \lceil \chi_c(G) \rceil. The circular chromatic number provides a finer measure of colorability.


Coloring Sparse Graphs

Definition5.8Coloring Planar and Sparse Graphs

The Four Color Theorem: every planar graph is 4-colorable. The Five Color Theorem (simpler proof): every planar graph is 5-colorable (by induction using Euler's formula: ฮด(G)โ‰ค5\delta(G) \leq 5). For graphs on surfaces: the Heawood bound ฯ‡โ‰คโŒŠ(7+1+48g)/2โŒ‹\chi \leq \lfloor(7 + \sqrt{1 + 48g})/2\rfloor for genus gโ‰ฅ1g \geq 1 is tight for all gโ‰ฅ1g \geq 1 (Ringel-Youngs theorem). Graphs with large girth have small chromatic number: ฯ‡(G)=O(ฮ”/logโกฮ”)\chi(G) = O(\Delta/\log\Delta) for triangle-free graphs (Johansson, 1996).

ExampleThomassen's Five-Choosability Theorem

Thomassen (1994) proved every planar graph is 5-choosable (list chromatic number โ‰ค5\leq 5) with an elegant inductive proof. Whether every planar graph is 4-choosable is false: Voigt (1993) constructed a planar graph that is not 4-choosable. This shows list coloring is strictly harder than ordinary coloring for planar graphs.