English

Large Monochromatic Components in Colored Random Graphs

Combinatorics 2026-07-25 v1 Probability

Abstract

We study the size of the largest monochromatic connected component that must appear in any edge-coloring of a random graph. Let GG(n,p)G\sim G(n,p) with p1/np\gg 1/n and p=o(1)p=o(1), and write np=hehnp=he^h. We show that, with high probability, every 22-edge-coloring of GG contains a monochromatic connected component of order at least nΘ(neh)n-\Theta(ne^{-h}). Moreover, we construct colorings showing that this bound is best possible up to constant factors. We extend this result to three colors: for p1/np\gg 1/n and p=o(1)p=o(1), with high probability every 33-edge-coloring of GG contains a monochromatic connected component of size at least n2Θ(1/p)\frac{n}{2}-\Theta(1/p), and this estimate is again tight up to constant factors. In the bipartite setting GG(n,n,p)G\sim G(n,n,p), under the same assumptions on pp, we prove an analogous statement: with high probability, every 22-edge-coloring contains two monochromatic components whose union covers all but Θ(neh)\Theta(ne^{-h}) vertices, and this bound is asymptotically sharp. Our approach is elementary and is based on analyzing large connected structures across suitably balanced vertex partitions.

Cite

@article{arxiv.2607.23334,
  title  = {Large Monochromatic Components in Colored Random Graphs},
  author = {Xiao-Chuan Liu and Xu Yang},
  journal= {arXiv preprint arXiv:2607.23334},
  year   = {2026}
}

Comments

16 pages