Gorenstein polytopes and their stringy \(E\)-functions (Q1938061)

A Gorenstein polytope of index \(r\) is a lattice polytope whose \(r\)th dilate is a reflexive polytope. Gorenstein polytopes were originally introduced by \textit{V. V. Batyrev} and \textit{L. A. Borisov} [Greene, B. (ed.) et al., Mirror symmetry II. Cambridge, MA: International Press, AMS/IP Stud. Adv. Math. 1, 71--86 (1997; Zbl 0927.14019)] and played a crucial role in Batyrev's and Borisov's computation of the Hodge numbers of mirror-symmetric generic Calabi-Yau complete intersections. The stringy \(E\)-function of an arbitrary Gorenstein polytope is defined to be a certain rational function in two variables as was introduced by \textit{V. V. Batyrev} and \textit{B. Nill} [Contemporary Mathematics 452, 35--66 (2008; Zbl 1161.14037)]. Drawing a comparison with algebraic geometry, the authors of the paper under review show that the stringy \(E\)-function of a Gorenstein polytope \(P\) looks like the generating function for the stringy Hodge numbers of Calabi-Yau varieties of certain dimension. That dimension is called the Calabi-Yau dimension of \(P\). The main result is that if \(P\) is a Gorenstein polytope of Calabi-Yau dimension \(n\), then the stringy \(E\)-function of \(P\) is actually a polynomial; it is \(0\) if \(n < 0\). In order to prove this result, the authors first establish that the index of a Gorenstein polytope cannot exceed the sum of codegrees of any pair of corresponding faces, one in that polytope and the other in its dual. In addition, the authors study various kinds of joins of Gorenstein polytopes, introduce the notion of an irreducible Gorenstein polytope, and also study the relations between these notions and the decomposition of nef-partitions.