Wiki
Wiki

Research notes on every problem, and a library of the papers behind them. Built from the open erdos repository.

Updated


Claim. P. Erdős and Z. Tuza, Rainbow subgraphs in edge-colorings of complete graphs, in Quo vadis, graph theory?, Ann. Discrete Math. 55, North-Holland (1993), 81--88. For a graph FF with ee edges they call an edge-coloring of KnK_n with precisely ee colors, every vertex meeting at least dd edges of each color, an (e,d)(e,d)-coloring, and set d(n,F)=∞d(n,F)=\infty when some (e,⌊(n−1)/e⌋)(e,\lfloor(n-1)/e\rfloor)-coloring has no rainbow FF, and otherwise let d(n,F)d(n,F) be the least dd for which every (e,d)(e,d)-coloring contains a rainbow FF (p. 81). They prove:

  • Theorem 2 (p. 82): the exact rainbow-triangle threshold for every number k≥3k\ge3 of colors, and in particular "$d(n,K_3)=2\lfloor(\lfloor n/2\rfloor-1)/4\rfloor=2\lfloor(n-2)/8\rfloor+1$" (as printed; the middle expression lacks the +1+1 of the theorem's first sentence at k=3k=3);
  • Theorem 3 (p. 83): "⌊n/6⌋≤d(n,C4)≤(1/4−c)n\lfloor n/6\rfloor \leq d(n,C_4) \leq (1/4-c)n for some positive constant cc";
  • Proposition 1 (p. 83): d(n,F)≤e−1d(n,F)\le e-1 for a tree and d(n,F)≤2e−2d(n,F)\le2e-2 for a forest with ee edges, improved to e−2e-2 and 2e−32e-3 for large nn.

Covers. For Problem 811, each bound lies below (n−1)/e(n-1)/e for large admissible nn, so every balanced coloring contains a rainbow copy: K3K_3 (Theorem 2), C4C_4 (Theorem 3, since (1/4−c)n<(n−1)/4(1/4-c)n<(n-1)/4 for large nn) and every forest (Proposition 1) are in the answer set. The paper's own summary (p. 81) names the trees, K3K_3 and C4C_4 as the only graphs for which the authors can prove the requirements of their Problems 1 and 2. Nothing is claimed for any other graph.

Depends on. Nothing in this wiki; the results are the paper's own.

Standing. Claimed. Crossref types the paper as a book chapter of Annals of Discrete Mathematics 55, and no refereeing of the volume is documented, so refereed is not listed; the site credits the paper for the dC4(n)d_{C_4}(n) bounds, but it labels the problem OPEN, so no reviewed evidence is listed either. This corpus checked the statements clause by clause and did not reconstruct the proofs.

Dating. Crossref gives only the year 1993, so the day and month in the page's name are placeholders.