English

$C_{2k+1}$-coloring of bounded-diameter graphs

Combinatorics 2024-10-23 v3 Computational Complexity

Abstract

For a fixed graph HH, in the graph homomorphism problem, denoted by Hom(H)Hom(H), we are given a graph GG and we have to determine whether there exists an edge-preserving mapping φ:V(G)V(H)\varphi: V(G) \to V(H). Note that Hom(C3)Hom(C_3), where C3C_3 is the cycle of length 33, is equivalent to 33-Coloring. The question whether 33-Coloring is polynomial-time solvable on diameter-22 graphs is a well-known open problem. In this paper we study the Hom(C2k+1)Hom(C_{2k+1}) problem on bounded-diameter graphs for k2k\geq 2, so we consider all other odd cycles than C3C_3. We prove that for k2k\geq 2, the Hom(C2k+1)Hom(C_{2k+1}) problem is polynomial-time solvable on diameter-(k+1)(k+1) graphs -- note that such a result for k=1k=1 would be precisely a polynomial-time algorithm for 33-Coloring of diameter-22 graphs. Furthermore, we give subexponential-time algorithms for diameter-(k+2)(k+2) graphs. We complement these results with a lower bound for diameter-(2k+2)(2k+2) graphs -- in this class of graphs the Hom(C2k+1)Hom(C_{2k+1}) problem is NP-hard and cannot be solved in subexponential-time, unless the ETH fails. Finally, we consider another direction of generalizing 33-Coloring on diameter-22 graphs. We consider other target graphs HH than odd cycles but we restrict ourselves to diameter 22. We show that if HH is triangle-free, then Hom(H)Hom(H) is polynomial-time solvable on diameter-22 graphs.

Keywords

Cite

@article{arxiv.2403.06694,
  title  = {$C_{2k+1}$-coloring of bounded-diameter graphs},
  author = {Marta Piecyk},
  journal= {arXiv preprint arXiv:2403.06694},
  year   = {2024}
}

Comments

Wrong statement about diameter-(k+3) graphs removed

R2 v1 2026-06-28T15:15:43.718Z