Indexed metadata

Sharp bounds for Duke--Erdős--Rödl problems on cycle-connected subgraphs

Guorong Gao, Run Zhao

Source record

Source: arXiv

Published: Oct 25, 2026

arXiv: 2610.34307

Open original source ↗

Source abstract

For k≥2k\ge2, a graph is C2kC_{2k}-connected if every two distinct edges lie together on an even cycle of length at most 2k2k in that graph. Let fk(n,m)f_k(n,m) be the largest number of edges in a C2kC_{2k}-connected subgraph that every nn-vertex graph with mm edges must contain. In this paper, we study two cycle-connected problems raised by Duke, Erdős and Rödl. First, Duke, Erdős and Rödl showed that f3(n,⌊n2−β⌋)=Θ(n2−3β)f_3(n,\lfloor n^{2-β}\rfloor)=Θ(n^{2-3β}) for every fixed 0<β<1/20< β<1/2 and ask for the case of β≥1/2β\ge 1/2. We solve this problem in a stronger form by determining that the optimal range of ββ is (0,2/3](0,2/3]. Second, we prove that f2(n,(n2)−s)=(1/(2e)+o(1))n3/sf_2(n,\binom n2-s)=(1/(2e)+o(1))n^3/s whenever sn→∞\frac{s}{n}\to\infty and s=o(n2)s=o(n^2), determining the leading constant in an estimate of Duke, Erdős, and Rödl. The first result uses a finite-field construction and a count of three-edge paths. The second uses the known structure of C4C_4-connected graphs and estimates for complete bipartite subgraphs in a random graph.

Evidence graph

No public relationships recorded yet.

Integrity note: This page is a factual metadata record created by deterministic ingestion. It is not a claim that the work moves a mathematical frontier or has been independently verified.