Ehrhart h*-Polynomial

• Ehrhart h*-polynomial is a central invariant defined as the numerator of the Ehrhart series, capturing refined combinatorial and geometric information of lattice polytopes.
• It provides explicit combinatorial interpretations via order polytopes, zonotopes, and matroid polytopes, linking permutation statistics like the swap statistic to polytope structure.
• Its algebraic properties, including symmetry, unimodality, and real-rootedness, have broad implications in enumerative combinatorics and toric geometry.

The Ehrhart $h^*$-polynomial is a central invariant in Ehrhart theory, encoding key combinatorial and geometric information about a lattice polytope through the numerators of its Ehrhart generating series. Recent work provides explicit combinatorial interpretations in various classes, with particular impact on the structure theory for order polytopes, zonotopes, matroid polytopes, and related families.

1. Definition and General Properties

Let $P\subset\mathbb{R}^n$ be a $d$-dimensional lattice polytope. The Ehrhart function $i_P(m) = |\mathbb{Z}^n \cap mP|$ is a polynomial in $m$ of degree $d$. The generating function, the Ehrhart series,

$$\mathrm{Ehr}_P(t) = \sum_{m\geq 0} i_P(m) t^m = \frac{h^*_P(t)}{(1-t)^{d+1}}$$

uniquely determines the numerator $h^*_P(t)\in\mathbb{Z}_{\geq 0}[t]$, called the Ehrhart $h^*$-polynomial (alternatively, the $\delta$-polynomial or $P\subset\mathbb{R}^n$0-vector) of $P\subset\mathbb{R}^n$1 (Coons et al., 2019). The degree satisfies $P\subset\mathbb{R}^n$2, and the sequence $P\subset\mathbb{R}^n$3 gathers refined data about the appearance of new lattice points in successive dilates of $P\subset\mathbb{R}^n$4 (Katz et al., 2014, Liu et al., 2018).

Normalization conditions include $P\subset\mathbb{R}^n$5 and $P\subset\mathbb{R}^n$6. Stanley's nonnegativity theorem ensures all coefficients $P\subset\mathbb{R}^n$7 (Katz et al., 2014, Liu et al., 2018).

2. Combinatorial Formulas: The Zig-Zag Poset and Swap Statistic

A paradigmatic case arises for the order polytope of the zig-zag poset $P\subset\mathbb{R}^n$8 (Coons et al., 2019). Here,

• The order polytope $P\subset\mathbb{R}^n$9 is defined by

$d$0

• Stanley's canonical triangulation for order polytopes is indexed by the linear extensions; for $d$1 these are the alternating permutations $d$2.

The $d$3-polynomial for $d$4 admits a closed combinatorial expansion via the swap statistic: $d$5 where $d$6 and

$d$7

with $d$8 (Coons et al., 2019).

Each $d$9 thus counts alternating permutations of length $i_P(m) = |\mathbb{Z}^n \cap mP|$0 with swap-value $i_P(m) = |\mathbb{Z}^n \cap mP|$1, making $i_P(m) = |\mathbb{Z}^n \cap mP|$2 the swap-distribution generating function on $i_P(m) = |\mathbb{Z}^n \cap mP|$3.

3. Shellings, Triangulations, and Permutation Statistics

Stanley's theory [BR07] applied to order polytopes provides, for any shelling, a rule that each maximal simplex contributes a monomial $i_P(m) = |\mathbb{Z}^n \cap mP|$4 to $i_P(m) = |\mathbb{Z}^n \cap mP|$5, with $i_P(m) = |\mathbb{Z}^n \cap mP|$6 the number of facets it is glued on. For $i_P(m) = |\mathbb{Z}^n \cap mP|$7, ordering the corresponding simplices $i_P(m) = |\mathbb{Z}^n \cap mP|$8 by nonincreasing inversion number of $i_P(m) = |\mathbb{Z}^n \cap mP|$9 gives a valid shelling.

Other classes (zonotopes (Beck et al., 2016), positroid polytopes (Jiang, 2024), type $m$0 hypersimplices (Abram et al., 4 Apr 2025)) provide remarkable formulas in terms of refined permutation statistics:

• For zonotopes, $m$1, a sum over independent sets $m$2 and refined Eulerian polynomials $m$3 (Beck et al., 2016).
• Type $m$4 hypersimplices admit formulas involving circular descents and big ascent statistics on the hyperoctahedral group; e.g., $m$5 (Abram et al., 4 Apr 2025).
• For positroid polytopes, the $m$6-polynomial can be expressed in terms of descents in a set of permutations parametrizing the triangulation (Jiang, 2024).

The emergence of these connections underscores the role of $m$7-polynomials as combinatorial statistics generating functions associated with shellable polyhedral subdivisions.

4. Structural Properties: Symmetry, Unimodality, and Real-Rootedness

Many families exhibit deep algebraic properties:

• For zonotopes, $m$8 is real-rooted, hence unimodal (Beck et al., 2016).
• For $m$9, the $d$0-vector is symmetric and unimodal as a consequence of Gorenstein properties and the existence of regular unimodular triangulations (Coons et al., 2019).
• In general, for polytopes with suitable Gorenstein and triangulation properties, the $d$1-polynomial is palindromic.

A key open question in (Coons et al., 2019) remains: to provide a direct combinatorial involution accounting for the symmetry/unimodality of the swap distribution. For many other classes (e.g., matroid polytopes, $d$2-lecture hall simplices), $d$3-real-rootedness is often observed or conjectured, directly implying unimodality (Liu et al., 2018).

5. Examples and Explicit Calculations

For $d$4:

• The alternating permutations $d$5
• One computes swap values as: $d$6, $d$7, $d$8, $d$9, $$\mathrm{Ehr}_P(t) = \sum_{m\geq 0} i_P(m) t^m = \frac{h^*_P(t)}{(1-t)^{d+1}}$$0, and thus

$$\mathrm{Ehr}_P(t) = \sum_{m\geq 0} i_P(m) t^m = \frac{h^*_P(t)}{(1-t)^{d+1}}$$1

For $$\mathrm{Ehr}_P(t) = \sum_{m\geq 0} i_P(m) t^m = \frac{h^*_P(t)}{(1-t)^{d+1}}$$2, $$\mathrm{Ehr}_P(t) = \sum_{m\geq 0} i_P(m) t^m = \frac{h^*_P(t)}{(1-t)^{d+1}}$$3 and $$\mathrm{Ehr}_P(t) = \sum_{m\geq 0} i_P(m) t^m = \frac{h^*_P(t)}{(1-t)^{d+1}}$$4 (Coons et al., 2019).

These explicit enumerations illustrate the correspondence between $$\mathrm{Ehr}_P(t) = \sum_{m\geq 0} i_P(m) t^m = \frac{h^*_P(t)}{(1-t)^{d+1}}$$5-coefficients and underlying combinatorics.

6. Connections, Generalizations, and Open Problems

The theory of Ehrhart $$\mathrm{Ehr}_P(t) = \sum_{m\geq 0} i_P(m) t^m = \frac{h^*_P(t)}{(1-t)^{d+1}}$$6-polynomials interfaces with diverse domains:

• The $$\mathrm{Ehr}_P(t) = \sum_{m\geq 0} i_P(m) t^m = \frac{h^*_P(t)}{(1-t)^{d+1}}$$7-polynomial for $$\mathrm{Ehr}_P(t) = \sum_{m\geq 0} i_P(m) t^m = \frac{h^*_P(t)}{(1-t)^{d+1}}$$8 coincides with that of the CFN–MC phylogenetic polytope, connecting algebraic statistics and phylogenetic inference (Coons et al., 2019).
• For regular Gorenstein order polytopes, Bruns–Römer guarantees symmetry/unimodality of $$\mathrm{Ehr}_P(t) = \sum_{m\geq 0} i_P(m) t^m = \frac{h^*_P(t)}{(1-t)^{d+1}}$$9 (Coons et al., 2019).
• Open problems include finding explicit combinatorial proofs of these properties and characterizing the image of $h^*_P(t)\in\mathbb{Z}_{\geq 0}[t]$0-vectors arising from specific families.

Known results do not in general guarantee that all $h^*_P(t)\in\mathbb{Z}_{\geq 0}[t]$1-vectors are unimodal or real-rooted. The appearance of $h^*_P(t)\in\mathbb{Z}_{\geq 0}[t]$2-polynomials as permutation statistics generating functions provides a powerful tool for explicit enumeration and structural study, but the full set of constraints and achievable forms remains an active topic (Coons et al., 2019, Liu et al., 2018).

7. Significance and Broader Implications

The Ehrhart $h^*_P(t)\in\mathbb{Z}_{\geq 0}[t]$3-polynomial encodes subtle geometric and combinatorial features of polytopes, often serving as a unifying framework for disparate counting problems. Its interplay with triangulations, shellings, permutation statistics, and polyhedral geometry has led to deep insights into the structure of polytopes and the distribution of lattice points. In the poset and matroid context, $h^*_P(t)\in\mathbb{Z}_{\geq 0}[t]$4-polynomials serve as generating series for statistics of linear extensions, descents, and related enumerative features. Their algebraic properties, such as symmetry, unimodality, and real-rootedness, yield connections to toric geometry, commutative algebra, and the theory of permutation statistics.

Continued investigation into $h^*_P(t)\in\mathbb{Z}_{\geq 0}[t]$5-polynomials and their combinatorial interpretations promises to further elucidate the combinatorial and geometric landscape of lattice polytopes and their applications across algebra, statistics, and geometry (Coons et al., 2019, Jiang, 2024, Beck et al., 2016, Abram et al., 4 Apr 2025).