Stability for binary scalar products

TL;DR

This paper determines the maximum product of vertices and facets in 2-level polytopes, confirming a conjecture and extending stability results to characterize these polytopes beyond the cube and cross-polytope.

Contribution

It resolves a strong version of Bohn et al.'s conjecture by identifying the maximum product in 2-level polytopes not isomorphic to the cube or cross-polytope.

Findings

Maximum product achieved only by the cube or cross-polytope.

Established a sharp stability result for the upper bound.

Extended previous bounds to a broader class of 2-level polytopes.

Abstract

Bohn, Faenza, Fiorini, Fisikopoulos, Macchia, and Pashkovich (2015) conjectured that 2-level polytopes cannot simultaneously have many vertices and many facets, namely, that the maximum of the product of the number of vertices and facets is attained on the cube and cross-polytope. This was proved in a recent work by Kupavskii and Weltge. In this paper, we resolve a strong version of the conjecture by Bohn et al., and find the maximum possible product of the number of vertices and the number of facets in a 2-level polytope that is not affinely isomorphic to the cube or the cross-polytope. To do this, we get a sharp stability result of Kupavskii and Weltge's upper bound on $\left|\mathcal A\right|\cdot\left|\mathcal B\right|$ for $\mathcal A,\mathcal B \subseteq \mathbb R^d$ with a property that $\forall a \in \mathcal A, b \in \mathcal B$ the scalar product $\langle a, b\rangle \in\{0,1\}$.