Tree Independence Number IV. Even-hole-free Graphs
Maria Chudnovsky, Peter Gartland, Sepehr Hajebi, Daniel Lokshtanov, Sophie Spirkl
TL;DR
This work proves that every $n$-vertex even-hole-free graph has tree independence number bounded by $c\log^{10} n$ for some constant $c>0$, implying that a tree decomposition with bags of independence number at most $c\log^{10} n$ exists. The authors develop a suite of structural tools—pyramids, wheels, loaded pyramids, star cutsets, tree-strip systems, and a central-bag method—to derive polylogarithmic clique-separator bounds that translate into balanced separators. By combining local-to-global separation results with a hub-based decomposition and a central-bag framework, they obtain a quasi-polynomial-time algorithmic regime for Maximum Weight Independent Set and several related problems on even-hole-free graphs, tightening the link between structural graph theory and algorithmic tractability. The approach builds on and extends prior work on logarithmic treewidth for $K_t$-free even-hole-free graphs, generalizing the methodology to the independence-number width and CMSO$_2$-based algorithmic consequences. Overall, the paper advances our understanding of how forbidding even holes constrains graph structure to yield powerful algorithmic consequences in quasi-polynomial time.
Abstract
We prove that the tree independence number of every even-hole-free graph is at most polylogarithmic in its number of vertices. More explicitly, we prove that there exists a constant c>0 such that for every integer n>1 every n-vertex even-hole-free graph has a tree decomposition where each bag has stability (independence) number at most c log^10 n. This implies that the Maximum Weight Independent Set problem, as well as several other natural algorithmic problems that are known to be NP-hard in general, can be solved in quasi-polynomial time if the input graph is even-hole-free.
