Ehrhart polynomials of convex polytopes with small volumes (Q616376): Difference between revisions

This paper deals with \(\delta\)-vectors of integral convex polytopes. Namely, let \(P\subset\mathbb R^d\) be an integral convex polytope, i.e. a convex polytope any of whose vertices has integer coordinates. Given a positive integer \(n\), set \(i(P,n)=| nP\cap\mathbb Z^N|\). This function is a polynomial in \(n\), it is called the Ehrhart polynomial of \(P\) [cf. \textit{E.~Ehrhart}, Polynômes arithmétiques et méthode des polyèdres en combinatoire. International Series of Numerical Mathematics. Vol. 35. Basel-Stuttgart: Birkhäuser Verlag (1977; Zbl 0337.10019)]. Then the \(\delta\)-vector of \(P\), \(\delta (P)=(\delta_0,\dots , \delta_d)\), is defined by the formula \[ (1-\lambda)^{d+1}\left[1+\sum_{n=1}^\infty i(P,n)\lambda^n\right] = \sum_{j=0}^d\delta_j\lambda^j, \] see [\textit{T.~Hibi}, Algebraic combinatorics on convex polytopes. Glebe: Carslaw Publications (1992; Zbl 0772.52008); \textit{A.~Stapledon}, Trans. Am. Math. Soc. 361, No.~10, 5615--5626 (2009; Zbl 1181.52024)]. The main result of the paper is the following classification of \(\delta\)-vectors with \(\sum\limits_{j=0}^d\delta_j\leq 3\). Theorem 1. Let \(d\geq 3\). Given a finite sequence \((\delta_0, \dots, \delta_d)\) of nonnegative integers, where \(\delta_0=1\) and \(\delta_1\geq \delta_d\), which satisfies \(\sum_{j=0}^d\delta_j\leq 3\), there exists an integral convex polytope \(P\subset\mathbb R^d\) whose \(\delta\)-vector coincides with \((\delta_0, \dots, \delta_d)\) if and only if \((\delta_0, \dots, \delta_d)\) satisfies the inequalities \[ \delta_{d-1}+\delta_{d-2}+\dots+\delta_{d-j}\leq \delta_{2}+\delta_{3}+ \dots+\delta_{j+1} \] for all \(1\leq j\leq [(d-1)/2]\) and \[ \delta_{0}+\delta_{1}+\dots+\delta_j\leq \delta_{s}+\delta_{s-1}+\dots+\delta_{s-j} \] for all \(0\leq j\leq [s/2]\), where \(s=\max\{j, \delta_j\not= 0\}\). Moreover the authors present an example which shows that Theorem 1 is no longer true for the case \(\sum_{j=0}^d\delta_j=4\). Notice that the possible \(\delta\)-vectors for the case of \(d\leq 2\) are discussed in [\textit{P. R.~Scott}, Bull. Aust. Math. Soc. 15, 395--399 (1976; Zbl 0333.52002)].