Largest integral simplices with one interior integral point: solution of Hensley's conjecture and related results (Q2359925)

Given \(k\in\mathbb{N} \cup\{0\}\), we define: \[ \mathcal{P}^d(k) =\{P \subseteq \mathbb{R}^d \mid P \text{ polytope}, \dim(P) = d, \mathrm{vert}(P) \subseteq \mathbb{Z}^d \text{ and } |\mathrm{int}(P)\cap \mathbb{Z}^d|= k\} \] and \[ \mathcal{S}^d(k) = \{S \in \mathcal{P}^d(k) \mid S \text{ is a simplex}\}. \] In the work of \textit{D. Hensley} [Pac. J. Math. 105, 183--191 (1983; Zbl 0471.52006)] it was proved that \(\mathcal{P}^d(k)\) and \(\mathcal{S}^d(k)\) are both finite. The Sylvester sequence \((s_i)_{i\in\mathbb{N}}\) is defined recursively as \(s_1= 2\) and \(s_i= 1 +\prod_{j=1}^{i-1}s_j\) for \(i \geq 2\). In [loc. cit.], Hensley conjectured that the simplex \(\mathrm{conv}(\{0, s_1e_1,\ldots,s_de_d\}) \in \mathcal{S}^d(1)\) has maximal coefficient of asymmetry among all elements of \(\mathcal{S}^d(1)\), where \(e_i\) are the elements of the standard basis of \(\mathbb{R}^d\). The article under review is devoted to the proof of this conjecture. Another `modified' Hensley conjecture is also proved: the simplex \(S_1^d= \mathrm{conv}(\{0, s_1e_1,\ldots,s_{d-1}e_{d-1}, 2(s_d - 1)e_d\})\) has maximal volume in \(\mathcal{S}^d(1)\). There are other useful contributions of this interesting article. The authors establish a series of results about the size of simplices in \(\mathcal{P}^d(k)\) and \(\mathcal{S}^d(k)\) for \(k\geq1\), improving in many cases the known bounds. Several applications of their results are also accounted, they include: (1) bounds on the volume and Mahler volume of \(l\)-dimensional faces of \(S\in\mathcal{S}^d(1)\) and \(S^\ast\) (the dual simplex of \(S\)) which are used in toric geometry (see Section 2.6), (2) sharp upper bounds on the lattice diameter of polytopes in \(\mathcal{P}^d(k)\) for \(k\geq1\) which are maximal with respect to being lattice-free, such polyhedra are relevant in mixed-integer optimization. The tools used in the proofs of the presented results are `classical' in the geometry of numbers, such as Blichfeldt and Mahler theorems, which are included for easy reference. Some other difficulties are dealt with by using unit partitions from number theory and also some properties of the Sylvester sequence. The article is well written and almost self-contained (excluding the part about toric geometry).