Given a complete graph whose edges are colored by 3 colors, a compatible 3-coloring is a coloring of nodes such that no edge of the graph has the same color as its end-points.

The best algorithm I know is quasi-polynomial time. The problem is not known to be NP-complete. What is known about the status of this problem ?

  • 2
    $\begingroup$ Hi turkistany it would be very helpful if you were to provide more context for your questions, and this one in general. Specifically, what's a 'compatible' 3-coloring, and what's the history of the problem that you're already aware of ? $\endgroup$ Aug 26, 2010 at 19:30
  • 1
    $\begingroup$ The phrase "no edge of the graph has the same color as its end-points" might be somewhat ambiguous. What is meant is that for each edge, at least one of its endpoints has a different color than the edge. $\endgroup$ Aug 26, 2010 at 20:31

2 Answers 2


Apparently, this paper by Cygan, Pilipczuk, Pilipczuk and Wojtaszczyk (2010+) gives a faster (polynomial time) algorithm: http://arxiv.org/abs/1004.5010

It was accepted to SODA 2011.


The fastest known algorithm has running time $n^{O(\log n/ \log \log n)}$ [1].
Your question is reposed as an open problem in [2, page 4].

[1]: T. Feder, P. Hell, D. Král, and J. Sgall. Two algorithms for general list matrix partitions. SODA 2005, pages 870–876.

[2]: E. D. Demaine, M. Hajiaghayi, D. Marx. Open Problems -- Parameterized complexity and approximation algorithms. Dagstuhl Seminar 09511 (2009). Available here.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.