A generalisation of Sylvester's problem to higher dimensions (Q2410134)

In the paper under review, the authors consider a higher dimensional analogue of the famous orchard problem (or Sylvester's problem). Let \(S\) be a set of points in \(d\)-space with the property that any \(d\) points of \(S\) span a hyperplane and not all the points of \(S\) are contained in a hyperplane. Let \(e_{d}(n)\) denote the minimal number of hyperplanes meeting \(S\) in precisely \(d\) points, minimising over all such sets of points \(S\) with \(|S| = n\). For any set \(S\) of points in \(d\)-space we say that a hyperplane \(\pi\) is an ordinary hyperplane if \(|\pi \cap S | = d\). The very first general result can be formulated as follows. Theorem. For \(d \geq 3\), one has \[ e_{d}(n) \geq \bigg\lceil \frac{n}{d} \bigg\lceil \frac{n-1}{d-1} \bigg\lceil \frac{n-2}{d-2} \cdots \bigg\lceil \frac{(n-d+3)}{3}e_{2}(n-d+2)\bigg\rceil \cdots \bigg\rceil \bigg\rceil \bigg\rceil. \] Using some bound on \(e_{2}(n)\), for instance due to \textit{J. Csima} and \textit{E. T. Sawyer} [Discrete Comput. Geom. 9, No. 2, 187--202 (1993; Zbl 0771.52003)], one can obtain an explicit bound, namely \[ e_{d}(n) \geq \bigg\lceil \frac{n}{d} \bigg\lceil \frac{n-1}{d-1} \bigg\lceil \frac{n-2}{d-2} \cdots \bigg\lceil \frac{(n-d+3)}{3}\bigg\lceil \frac{6(n-d+2)}{13} \bigg\rceil\bigg\rceil \cdots \bigg\rceil \bigg\rceil \bigg\rceil \] if only \(n \neq d+5\). The main result of the paper is summed up as Theorem 2.5, where the authors present a list of possible values of \(e_{d}(n)\) for \(n\) sufficiently large, especially for \(d\geq 4\) there exists a universal constant \(c\) for which \[ \frac{3}{d!} n^{d-1} - \frac{c}{d!}n^{d-2} \leq e_{d}(n) \leq { n - 1 \choose d-1 }. \] In the last section, the authors formulate some conjectures. Let me recall the following one, the most interesting from my point of view. Conjecture. For all \(n\), \(e_{3}(n) \geq \frac{1}{4}n^2 - n\).