Open problem, Combinatorics and discrete mathematics, posed 1943
Hadwiger's conjecture (graph minors)
Open
Every loopless graph with no minor is -colorable; equivalently, every graph with chromatic number contains the complete graph as a minor.
As of 2026, Hadwiger's conjecture is proved for and open for all . Even the linear Hadwiger conjecture—that for every -minor-free graph —remains open. After Kostochka (1982) and Thomason (1984) showed that degeneracy gives , Norin, Postle, and Song (2019/2023) broke the degeneracy barrier, and Delcourt and Postle (2021) established the current best general bound .
Best known results
- Hadwiger's conjecture holds exactly for all (Robertson, Seymour, and Thomas, 1993).
- Every -minor-free graph has chromatic number (Delcourt and Postle, 2021, improving Norin–Postle–Song).
Tools and where they stop
| Tool | Achieved | Where it stops |
|---|---|---|
| Graph minor structure theory and apex reductions | Classifies minimal -chromatic -minor-free graphs as apex graphs over planar graphs, reducing to the four-color theorem. | For , -minor-free graphs can be built from higher-genus surfaces and vortices that are no longer controlled by planar coloring theorems. |
| Dense subgraph linking and chromatic-separability decompositions | Extracts vertex-disjoint dense highly connected subgraphs in high-chromatic graphs and links them into a minor, yielding . | Recursive density-increment steps lose a factor when balancing small-graph bounds against the number of parts needed to build . |
Open questions
- Is every graph with no minor -colorable ()?
- Does there exist an absolute constant such that every -minor-free graph satisfies (the linear Hadwiger conjecture)?
References
- Neil Robertson, Paul Seymour, Robin Thomas (1993). Hadwiger's conjecture for K6-free graphs · DOI:10.1007/BF01202354
- Sergey Norin, Luke Postle, Zi-Xia Song (2023). Breaking the degeneracy barrier for coloring graphs with no Kt minor · DOI:10.1016/j.aim.2023.108959 · arXiv:1910.09378