English

An efficient asymmetric removal lemma and its limitations

Combinatorics 2025-02-19 v2

Abstract

The triangle removal states that if GG contains εn2\varepsilon n^2 edge-disjoint triangles, then GG contains δ(ε)n3\delta(\varepsilon)n^3 triangles. Unfortunately, there are no sensible bounds on the order of growth of δ(ε)\delta(\varepsilon), and at any rate, it is known that δ(ε)\delta(\varepsilon) is not polynomial in ε\varepsilon. Csaba recently obtained an asymmetric variant of the triangle removal, stating that if GG contains εn2\varepsilon n^2 edge-disjoint triangles, then GG contains 2poly(1/ε)n52^{-\mathrm{poly}(1/\varepsilon)}\cdot n^5 copies of C5C_5. To this end, he devised a new variant of Szemer\'edi's regularity lemma. We obtain the following results: - We first give a regularity-free proof of Csaba's theorem, which improves the number of copies of C5C_5 to the optimal number poly(ε)n5\mathrm{poly}(\varepsilon)\cdot n^5. - We say that HH is K3K_3-abundant if every graph containing εn2\varepsilon n^2 edge-disjoint triangles has poly(ε)nV(H)\mathrm{poly}(\varepsilon)\cdot n^{|V(H)|} copies of HH. It is easy to see that a K3K_3-abundant graph must be triangle-free and tripartite. Given our first result, it is natural to ask if all triangle-free tripartite graphs are K3K_3-abundant. Our second result is that assuming a well-known conjecture of Ruzsa in additive number theory, the answer to this question is negative. Our proofs use a mix of combinatorial, number-theoretic, probabilistic, and Ramsey-type arguments.

Keywords

Cite

@article{arxiv.2301.07693,
  title  = {An efficient asymmetric removal lemma and its limitations},
  author = {Lior Gishboliner and Asaf Shapira and Yuval Wigderson},
  journal= {arXiv preprint arXiv:2301.07693},
  year   = {2025}
}

Comments

20 pages

R2 v1 2026-06-28T08:14:45.960Z