Optimal logarithmic factor for complete minors in small-set expanders

Establish whether every $(\alpha,t)$-expander on n vertices contains a complete minor of order at least a constant multiple of $\sqrt{nt/\log t}$, improving the bound $\sqrt{nt/\log n}$ in Theorem 1.3.

Background

The paper proves that, for fixed α\alpha and sufficiently large parameters, an (α,t)(\alpha,t)-expander contains a complete minor of order Θα(nt/logn)\Theta_\alpha(\sqrt{nt/\log n}). The authors note that analogous results for graphs with strong edge-expansion properties achieve a denominator involving logd\log d, where d is the average degree. They leave unresolved whether the small-set-expansion hypothesis alone suffices to replace logn\log n by logt\log t, which would yield a stronger bound without imposing any assumption on the average degree.

References

It remains an interesting question whether the bound in Theorem \ref{thm:complete_minors} can be improved to $\sqrt{nt / \log t}$.

Minors in small-set expanders  (2503.06826 - Krivelevich et al., 10 Mar 2025) in Section 1, paragraph following Theorem 1.3