The hamburger theorem

TL;DR

This paper extends the ham sandwich theorem to multiple measures in higher dimensions, providing a new hyperplane partition result with applications to colored point set partitions.

Contribution

It generalizes the ham sandwich theorem to $d+1$ measures in $R^d$, establishing a hyperplane partition with specific measure bounds and a related point set partition result.

Findings

Existence of a hyperplane satisfying measure inequalities for $d+1$ measures.

Partitioning of colored point sets into rainbow simplices.

Application to combinatorial geometry problems.

Abstract

We generalize the ham sandwich theorem to $d+1$ measures in $\mathbb{R}^d$ as follows. Let $\mu_1,\mu_2, \dots, \mu_{d+1}$ be absolutely continuous finite Borel measures on $\mathbb{R}^d$. Let $\omega_i=\mu_i(\mathbb{R}^d)$ for $i\in [d+1]$, $\omega=\min\{\omega_i; i\in [d+1]\}$ and assume that $\sum_{j=1}^{d+1} \omega_j=1$. Assume that $\omega_i \le 1/d$ for every $i\in[d+1]$. Then there exists a hyperplane $h$ such that each open halfspace $H$ defined by $h$ satisfies $\mu_i(H) \le (\sum_{j=1}^{d+1} \mu_j(H))/d$ for every $i \in [d+1]$ and $\sum_{j=1}^{d+1} \mu_j(H) \ge \min(1/2, 1-d\omega) \ge 1/(d+1)$. As a consequence we obtain that every $(d+1)$-colored set of $nd$ points in $\mathbb{R}^d$ such that no color is used for more than $n$ points can be partitioned into $n$ disjoint rainbow $(d-1)$-dimensional simplices.