Cells in the box and a hyperplane (Q6172690)

The well-known fact that a line can intersect at most \(2n-1\) cells of the \(n\times n\) chessboard is extended to higher dimensions, in an asymptotic way. Let \(Q_n^d\) be the \(n\times n\) cube in \({\mathbb R}^d\), divided in the chessboard manner into unit cubes, the cells. Let \(N^d(n)\) be the maximal number of interiors of cells that can be intersected by a hyperplane. It is proved that \(N^d(n)=V_dn^{d-1}(1+o(1))\), where \(V_d\) is a geometrically defined constant, satisfying \(\sqrt{d}\le V_d\le \sqrt{2d}\). The result has an extension to general convex bodies. The proof makes use of various estimates for numbers of lattice points contained in certain convex sets. Also, a result on boundary cells is proved (the novelty of which is surprising). A cell (now any unit cube \(\{(x_1,\dots,x_d)\in{\mathbb R}^d: z_i\le x_i\le z_i+1,\,i=1,\dots,d\}\) with \((z_1,\dots,z_d)\in{\mathbb Z}^d\)) is a boundary cell of the convex body \(K\) if it meets \(K\) and \({\mathbb R}^d\setminus K\). If \(K,L\) are convex bodies in \({\mathbb R}^d\) such that \(K\subset L\), then the number of boundary cells of \(K\) is at most the number of boundary cells of \(L\).