Abstract

A classic theorem of Erdős and Pósa (1965) states that every graph has either kk vertex-disjoint cycles or a set of O(klog⁡k)O(k \log k) vertices meeting all its cycles. While the standard proof revolves around finding a large `frame' in the graph (a subdivision of a large cubic graph), an alternative way of proving this theorem is to use a ball packing argument of Kühn and Osthus (2003) and Diestel and Rempel (2005). In this paper, we argue that the latter approach is particularly well suited for studying edge variants of the Erdős-Pósa theorem. As an illustration, we give a short proof of a theorem of Bruhn, Heinlein, and Joos (2019), that cycles of length at least ℓ\ell have the so-called edge-Erdős-Pósa property. More precisely, we show that every graph GG either contains kk edge-disjoint cycles of length at least ℓ\ell or an edge set FF of size O(kℓ⋅log⁡(kℓ))O(k\ell \cdot \log (k\ell)) such that G−FG-F has no cycle of length at least ℓ\ell. For fixed ℓ\ell, this improves on the previously best known bound of O(k2log⁡k+kℓ)O(k^2 \log k +k\ell).

Explore this paper’s authors, topics and related work.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.

Reviews 0

Write a reviewWrite

No reviews yet.

Discussion 0

Add a commentWrite

No comments yet.