Tensor valuations on lattice polytopes (Q2404631): Difference between revisions

A \textit{valuation} is a function \(f\) from some class of polyhedral objects into an abelian group that satisfies inclusion-exclusion: \(f(P) + f(Q) = f(P \cup Q) + f(P \cap Q)\). The paper under review forms a starting point of the theory of tensor valuations on lattice polytopes. (A \textit{lattice polytope} in \(\mathbb R^d\) is the convex hull of finitely many points in~\(\mathbb Z^d\).) The main object is a tensor version of the lattice-point counting function \(|P \cap \mathbb Z^d|\). More precisely, the authors define the \textit{discrete moment tensor of rank \(r\)} for a lattice polytope \(P\) as \[ L^r(P) \;:= \;{1 \over { r! }} \sum_{ \mathbf x \in P \cap \mathbb Z^d } \mathbf x^r \] where \(\mathbf x^r\) denotes the \(r\)-fold symmetric tensor product of \(\mathbf x\). The first main theorem is a tensor analogue of Ehrhart's celebrated theorem that \(|kP \cap \mathbb Z^d|\) is a polynomial in the positive integer variable \(k\) [\textit{E. Ehrhart}, C. R. Acad. Sci., Paris 254, 616--618 (1962; Zbl 0100.27601)]: the authors prove that \[ L^r(kP) \;= \;\sum_{ j=1 }^{ n+r } L_j^r(P) \, k^j \] where each \(L_j^r(P)\) is an \(\text{SL}_d(\mathbb Z)\)-equivariant, translation covariant, \(j\)-homogeneous valuation. The paper also gives a tensor version of the Ehrhart-Macdonald reciprocity theorem relating the lattice-point enumerator of \(P\) to that of the (relative) interior of \(P\) [\textit{I. G. Macdonald}, J. Lond. Math. Soc., II. Ser. 4, 181--192 (1971; Zbl 0216.45205)], as well as a complete classification of \(\text{SL}_d(\mathbb Z)\)-equivariant, translation covariant tensor valuations of rank \(\leq 8\); every such valuation is a linear combination of Ehrhart tensors, but this does not hold for rank~9.