Fact-checked by Grok 2 weeks ago

Ehrhart polynomial

In , the Ehrhart polynomial of a convex P \subset \mathbb{R}^d is the unique L_P(t) \in \mathbb{Q} of degree d such that for every positive integer k, L_P(k) equals the number of points in kP \cap \mathbb{Z}^d, where kP denotes the k-fold of P with respect to the . The leading coefficient of L_P(t) is the normalized volume of P, the second-highest coefficient is half the normalized surface area, and the constant term is always 1, reflecting the single lattice point at the when k = 0. These coefficients admit explicit formulas in terms of the intrinsic volumes of the faces of P, providing a bridge between discrete and continuous . Named after the French mathematician Eugène Ehrhart, who established the polynomial nature of this counting function in his seminal 1962 paper "Sur les polyèdres rationnels homothétiques à n dimensions," the theory generalizes on lattice polygons to arbitrary dimensions. For rational polytopes (with rational vertices), the counting function becomes a quasi-polynomial, periodic in the coefficients with a period dividing the of the denominators of the vertices. A cornerstone result is Ehrhart reciprocity, which asserts that L_P(-k) = (-1)^d times the number of interior lattice points in kP for positive integers k, linking values at positive and negative integers. The associated Ehrhart series \sum_{k=0}^\infty L_P(k) x^k = h^*(x) / (1 - x)^{d+1}, where h^*(x) is the h^*-polynomial with non-negative integer coefficients by Stanley's theorem, further encodes combinatorial invariants like the h^*-vector. Ehrhart polynomials find applications in for counting integer solutions to linear inequalities, in the study of toric varieties via their connections to h^*-polynomials, and in computational fields such as for optimizing loop transformations through exact lattice point counts. They also appear in voting theory to compute probabilities of election outcomes by enumerating lattice points in transportation polytopes.

Fundamentals

Definition

A lattice polytope P is a in \mathbb{R}^d whose vertices all belong to the \mathbb{Z}^d. The associated to such a P is defined for positive integers t by L_P(t) = |tP \cap \mathbb{Z}^d|, where tP = \{ t x \mid x \in P \} denotes the t-dilate of P. This function counts the lattice points lying inside or on the boundary of the dilated polytope. Eugène Ehrhart established that L_P(t) agrees with a in t of d = \dim(P). One proof of the polynomial nature of L_P(t) proceeds via generating functions: the Ehrhart series \sum_{t=0}^\infty L_P(t) z^t admits a rational generating function of the form h(z)/(1-z)^{d+1}, where h(z) is a polynomial with h(0) = 1, implying that the coefficients L_P(t) are polynomial in t for t \geq 0. The leading coefficient of this polynomial is the d-dimensional volume \vol(P) of P. For simple cases, consider the one-dimensional P = [0, a] with a \in \mathbb{Z}_{\geq 0}. Here, L_P(t) = at + 1, which is linear with leading a = \vol(P). In higher dimensions, the d- \Delta^d = \conv\{ \mathbf{0}, e_1, \dots, e_d \}, where e_i are the vectors, has Ehrhart polynomial L_{\Delta^d}(t) = \binom{t + d}{d}, a degree-d polynomial whose leading is $1/d! = \vol(\Delta^d).

Reciprocity theorem

The Ehrhart reciprocity theorem provides a fundamental relation between the Ehrhart polynomial L_P(t) of a d-dimensional P and the counting function for lattice points in its interior. Specifically, for any positive integer t, L_P(-t) = (-1)^d L_{P^\circ}(t), where P^\circ denotes the relative interior of P, and L_{P^\circ}(t) counts the lattice points in the t-dilate of the interior. This signed evaluation at negative arguments reveals that the Ehrhart polynomial encodes both boundary and interior lattice point enumerations through . The theorem was proved by Eugène Ehrhart in 1962 as part of his foundational work on lattice point enumeration in rational polytopes. Its roots trace to Pick's theorem (1899), which equates the area of a lattice polygon to the number of interior plus half the boundary lattice points, serving as the d=2 case of reciprocity. Ehrhart's discovery extended this relation to higher dimensions, highlighting the polynomial's duality between the polytope and its interior. An important interpretation arises from expressing the Ehrhart polynomial via an alternating over intrinsic volumes: L_P(t) = \sum_{k=0}^d (-1)^{d-k} \vol_k(P) \binom{t + d - k}{d}, where \vol_k(P) denotes the k-dimensional intrinsic volume of P. This form connects directly to the h^*- of P, obtained by a from the standard monomial form, where the coefficients h_i^* are nonnegative integers reflecting the of the normal fan. Reciprocity implies that evaluating this at negative t yields the signed interior count, underscoring the theorem's role in decomposing lattice point data geometrically. A bijective proof of the proceeds via inclusion-exclusion on , then extends to general polytopes using . For a d- \Delta, consider the (t + d + 1)-dilate and apply inclusion-exclusion over its facets to relate and interior points; a geometric maps excess points to interior adjustments, yielding the signed relation. For general P, triangulate into and invoke inversion on the poset of faces to preserve the reciprocity. approaches alternatively use the toric ideal of the polytope's normal , where the denominator reflects contributions and numerator adjustments enforce the alternating sign via Serre duality. In low dimensions, the theorem simplifies computations. Consider the triangle P with vertices (0,0), (2,0), and (2,1) in \mathbb{R}^2, a lattice polytope of dimension d=2. The Ehrhart polynomial is L_P(t) = t^2 + 2t + 1, counting 4, 9, and 16 lattice points in P, $2P, and $3P, respectively. By reciprocity, L_P(-1) = (-1)^2 L_{P^\circ}(1) = 0, so zero interior points in P; similarly, L_P(-2) = 1 gives one interior point in $2P (e.g., at (3,1)). This matches direct enumeration and illustrates the theorem's utility for verifying interior counts without exhaustive listing.

Basic Examples

Integer polytopes

Integer polytopes, which have vertices at integer coordinates, provide straightforward examples where the Ehrhart polynomial exactly counts the lattice points in their dilates. Consider the unit square P = [0,1]^2 in \mathbb{R}^2, a 2-dimensional integer with 1. The Ehrhart polynomial of this polytope is L_P(t) = (t+1)^2, which enumerates the lattice points in the dilate tP = [0,t]^2. To verify this, enumerate the lattice points for small t. For t=1, the dilate is [0,1]^2, containing the four points (0,0), (0,1), (1,0), and (1,1). For t=2, [0,2]^2 includes all integer pairs (i,j) with $0 \leq i,j \leq 2, totaling nine points: the previous four plus (0,2), (1,2), (2,0), (2,1), and (2,2). For t=3, [0,3]^2 adds points with coordinates up to 3, yielding 16 points. These counts—4, 9, 16—match (t+1)^2 for t=1,2,3.
tLattice points in tPL_P(t)
144
299
31616
Another fundamental example is the standard d- \Delta_d = \operatorname{conv}\{0, e_1, \dots, e_d\} in \mathbb{R}^d, where e_i are the vectors; this is a d-dimensional with $1/d!. Its Ehrhart polynomial is L_{\Delta_d}(t) = \binom{t + d}{d}, counting non-negative solutions to x_1 + \dots + x_d \leq t. For the 2-dimensional case (d=2), enumerate explicitly. For t=1, \Delta_2 contains (0,0), (1,0), and (0,1), totaling 3 points. For t=2, add (2,0), (1,1), and (0,2), for 6 points. For t=3, include (3,0), (2,1), (1,2), and (0,3), reaching 10 points. These align with \binom{t+2}{2}: 3, 6, 10.
tLattice points in t\Delta_2L_{\Delta_2}(t)
133
266
31010
These examples illustrate key properties: the degree of L_P(t) equals the dimension of P, matching the growth rate of lattice points, while the leading coefficient equals the volume of P. For the unit square, degree 2 and leading coefficient 1 reflect its 2D nature and unit volume; for \Delta_d, degree d and leading coefficient $1/d! do the same.

Rational polytopes

A rational is a in \mathbb{R}^d whose vertices have rational coordinates. The denominator q of such a P is the smallest positive such that qP has vertices, making qP an integral . Unlike integral polytopes, where the Ehrhart function L_P(t) is a , for rational polytopes it is generally a quasi-polynomial whose period divides the denominator q. Consider the rational triangle P with vertices (0,0), (1/2,0), and (0,1/2). This has denominator q=2, as $2P is the integral unit with vertices (0,0), (1,0), and (0,1). The Ehrhart function L_P(t) counts the points in tP = \{(x,y) \in \mathbb{R}^2 \mid x \geq 0, y \geq 0, x + y \leq t/2\}, which consists of nonnegative pairs (x,y) satisfying x + y \leq t/2. This yields a quasi-polynomial of period 2: L_P(t) = \begin{cases} \dfrac{t^2 + 6t + 8}{8} & \text{if } t \text{ is even}, \\ \dfrac{t^2 + 4t + 3}{8} & \text{if } t \text{ is odd}. \end{cases} The following table lists the number of lattice points for small values of t, illustrating the periodic behavior and deviation from a single polynomial (e.g., the count remains constant from t=2 to t=3, unlike the steady growth expected from a ).
tLattice points in tPL_P(t)
1(0,0)
2(0,0), (1,0), (0,1)
3(0,0), (1,0), (0,1)
4(0,0), (1,0), (0,1), (2,0), (1,1), (0,2)6
When the denominator q=1, the polytope is integral and L_P(t) reduces to a true polynomial.

Ehrhart Quasi-Polynomials

Definition and properties

The Ehrhart quasi-polynomial of a rational polytope P \subseteq \mathbb{R}^d is defined in relation to its denominator q, the smallest positive integer such that all vertices of P lie in \frac{1}{q} \mathbb{Z}^d. For positive integers t, the Ehrhart function L_P(t) counts the number of lattice points in the dilation tP, i.e., L_P(t) = \#(tP \cap \mathbb{Z}^d). This function is a quasi-polynomial of degree d: L_P(t) = \sum_{r=0}^d c_r(t) \, t^r, where each coefficient function c_r: \mathbb{Z} \to \mathbb{Q} is periodic with period dividing q. The existence of L_P(t) as a quasi-polynomial follows from its behavior on arithmetic progressions tied to the integer dilates of the scaled polytope qP, which is an polytope with Ehrhart L_{qP}(s). Specifically, L_P(t) = L_{qP}(t/q), where the right-hand side extends the L_{qP} to rational arguments via the quasi-polynomial structure induced by the periodicity of period q. This transformation establishes that L_P(t) agrees with a on multiples of q and interpolates quasi-periodically elsewhere. A key property is the generalized reciprocity , which extends the case to rational polytopes: L_P(-t) = (-1)^d L_{\mathrm{int}\, P}(t), where \mathrm{int}\, P denotes the relative interior of P. This equates the quasi-polynomial evaluated at negative arguments (with a ) to the count of lattice points in the interior of the tP, providing a signed interpretation of interior points. The holds by of the Ehrhart series or via inclusion-exclusion on contributions. The periodicity theorem asserts that the minimal period of each c_r(t) divides q, the denominator of P, which is the least common multiple of the denominators of the coordinates of P's vertices (expressed in lowest terms). While the period always divides this LCM, the minimal period may be strictly smaller in some cases, but it achieves the full LCM for certain simplices and coefficients, such as the second-highest term.

Examples

A prominent example of an Ehrhart quasi-polynomial with period 3 is provided by the rational P in \mathbb{R}^2 with vertices at (0,0), (2/3,0), and (0,1/3). This has denominator 3, as the least common multiple of the denominators in its vertex coordinates is 3, and the of its Ehrhart quasi-polynomial is exactly 3. The volume of P is $1/9, so the leading coefficient of the quasi-polynomial is the constant $1/9. The full quasi-polynomial is L(P, t) = \frac{1}{9} t^2 + b(t) t + c(t), where the periodic functions with period 3 are b(t) = \begin{cases} 4/9 & \text{if } t \equiv 1 \pmod{3}, \\ 5/9 & \text{if } t \equiv 2 \pmod{3}, \\ 2/3 & \text{if } t \equiv 0 \pmod{3}, \end{cases} and c(t) = \begin{cases} 4/9 & \text{if } t \equiv 1 \pmod{3} \text{ or } t \equiv 2 \pmod{3}, \\ 1 & \text{if } t \equiv 0 \pmod{3}. \end{cases} To illustrate the periodicity, the number of lattice points in tP for t = 1,2,3 is 1, 2, and 4, respectively. These values repeat the pattern in coefficients when extended to higher t: for t=4 \equiv 1 \pmod{3}, there are 4 lattice points; for t=5 \equiv 2 \pmod{3}, 6 lattice points; and for t=6 \equiv 0 \pmod{3}, 9 lattice points. Substituting into the quasi-polynomial confirms the counts, such as for t=3: (1/9)(9) + (2/3)(3) + 1 = 1 + 2 + 1 = 4. Another example is the rational Q with base the unit square in the xy-plane with vertices (0,0,0), (1,0,0), (0,1,0), (1,1,0), and apex at (1/2, 1/2, 1/2). This has denominator 2, and its Ehrhart function has degree 3. The volume of Q is $1/6, so the leading coefficient is the constant $1/6. Computations show that, despite the denominator of 2, the Ehrhart function is actually a (period 1): L(Q, t) = \frac{1}{6} t^3 + t^2 + \frac{11}{6} t + 1. The following table lists the number of lattice points in tQ for t=1 to $6, demonstrating the polynomial behavior:
tLattice points in tQ
14
210
320
435
556
684
Fitting the to these values confirms the coefficients. This example highlights that the of the Ehrhart quasi- for a rational divides the denominator of the , often determined by the of the denominators in the coordinates, but can be strictly smaller (here, 1), and computations for small dilates suffice to identify and verify the structure.

Coefficient Analysis

Geometric interpretations

The coefficients of the Ehrhart polynomial L_P(t) = \sum_{k=0}^d c_k t^k for a d-dimensional P carry significant geometric meaning, primarily relating the discrete lattice-point counts to continuous measures of P and its faces. The leading coefficient c_d equals the Euclidean of P divided by d!, or equivalently, the normalized of P (defined as d! times the Euclidean ) divided by d!. This normalization ensures that the standard d- has normalized 1, highlighting the combinatorial structure underlying the polytope's geometry. The second-highest coefficient c_{d-1} is given by \frac{1}{2} \sum vol(F), where the sum is over all facets F of P and vol(F) denotes the (d-1)-dimensional volume of F. Equivalently, in normalized terms, c_{d-1} = \frac{1}{2(d-1)!} \sum \tilde{vol}(F), where \tilde{vol}(F) is the normalized volume of the facet F. This term captures half the "surface volume" of P, providing a measure of its complexity. For lower-degree coefficients c_k with k < d-1, the interpretations become more intricate but generally involve explicit formulas tied to the volumes of the k-dimensional faces of P, incorporating contributions from the lattice structure via inclusion-exclusion principles over the face . These coefficients encode higher-order geometric invariants, such as angular deficits or edge lengths adjusted by lattice widths, though explicit closed forms are rare beyond low dimensions. The Ehrhart–Macdonald reciprocity further enriches these interpretations by relating the coefficients to interior geometry: L_P(-t) = (-1)^d L_{P^\circ}(t), where P^\circ is the interior of P, implying that alternating sums of coefficients yield signed volumes of interior face structures. This duality connects boundary-dominated terms (for positive t) to interior volumes with sign alternations. In the two-dimensional case, these interpretations specialize to , where for a polygon P, L_P(t) = A t^2 + \frac{B}{2} t + 1, with A the area (so c_2 = A), B the number of points (so c_1 = B/2), and the constant term 1 corresponding to the origin plus interior adjustments; states A = I + \frac{B}{2} - 1, with I the number of interior points, directly linking coefficients to area and .

Bounds and inequalities

The Betke–Kneser theorem provides a fundamental characterization of the coefficients of the Ehrhart polynomial for a d-dimensional P. It states that the space of all real-valued, translation-invariant, additive, and unimodular valuations on the set of polytopes in \mathbb{R}^d is (d+1)-dimensional, with the coefficients c_k of the Ehrhart polynomial i_P(t) = \sum_{k=0}^d c_k t^k forming a basis for this space. This means any such valuation Z can be uniquely expressed as Z(P) = \sum_{k=0}^d \lambda_k c_k for some constants \lambda_k \in \mathbb{R}. The theorem implies that the Ehrhart coefficients are the "primitive" valuations in this context, capturing essential geometric invariants like (c_d = \mathrm{Vol}(P)) and lower-dimensional measures. A proof sketch of the Betke–Kneser theorem relies on the existence of unimodular triangulations of lattice polytopes, which decompose P into standard unimodular simplices whose Ehrhart polynomials are known explicitly (e.g., \binom{t+d}{d} for the standard d-simplex). By additivity, the valuation on P reduces to a linear combination over these simplices, and unimodularity ensures the coefficients align with the basis spanned by the c_k. Alternatively, Fourier–Motzkin elimination can be applied to the system of linear equations arising from valuation properties on simplices and their faces, yielding the basis dimension and form. Equality holds in the basis representation when the valuation is a multiple of a single c_k, such as the volume functional itself. Building on this framework, the Betke–McMullen inequalities provide explicit on the non-leading c_r for $1 \leq r \leq d-1. Specifically, for the c_{d-1}, which geometrically interprets as half the normalized surface area of P, the inequality is c_{d-1} \leq \binom{d}{2} c_d + (d-1)! = \frac{d(d-1)}{2} \mathrm{Vol}(P) + (d-1)!, with equality achieved for certain polytopes like the product of a and unimodular simplices in lower dimensions. More generally, for c_r, c_r \leq (-1)^{d-r} s(d,r) c_d + (-1)^{d-r-1} s(d,r+1) (d-1)!, where s(d,r) are the signed of the first kind; these bounds are optimal in the sense that they are tight for specific families of polytopes, such as crosspolytopes or simplices. Lower bounds follow from analogous constructions, such as c_r \geq a(d,r) \mathrm{Vol}(P) + b(d,r) for explicit functions a and b derived from minimal-volume examples. Broader inequalities for non-leading coefficients c_k (k < d) arise from properties of forward differences of the Ehrhart polynomial. In particular, \Delta^k i_P(0) \leq \binom{d}{k} d! c_d, where \Delta is the forward difference operator, and since c_k = \frac{1}{k!} \sum_{j=0}^k (-1)^{k-j} \binom{k}{j} i_P(j), this implies |c_k| \leq \binom{d}{k} \mathrm{Vol}(P) + lower-order terms bounded by binomial coefficients times factorial volumes of faces. These hold with equality for the unit cube [0,1]^d, whose Ehrhart polynomial is (t+1)^d = \sum_{k=0}^d \binom{d}{k} t^k, yielding c_{d-1} = d and \mathrm{Vol}(P) = 1, saturating the leading term in the bound. For cubes of side length m, scaling gives tight instances where the inequality matches the surface-volume ratio d m^{d-1}. Scott's theorem offers sharp bounds relating the number of interior lattice points to the volume, via Ehrhart reciprocity which gives the number of interior points i = (-1)^d L_P(-1). For a 2-dimensional polygon equivalent to a unimodular with i \geq 1 interior points, the normalized satisfies \mathrm{Vol}(P) \leq 4(i+1) in the non-special case, implying i \geq \frac{\mathrm{Vol}(P)}{4} - 1; this extends to higher-dimensional simplices unimodularly equivalent to products involving such triangles, providing a lower bound on interior points i \geq \frac{d! \mathrm{Vol}(P)}{4} - 1 in low-degree cases. Equality occurs for the of 9 with one interior point in the special case, and analogs like the 3-simplex with 9 and one interior point. These bounds are tight for simplices and highlight how unimodularity constrains interior point counts relative to .

Ehrhart Series

Generating functions

The Ehrhart series of an integer polytope P of dimension d is defined as the E_P(u) = \sum_{t=0}^\infty L_P(t) u^t, where L_P(t) denotes the number of lattice points in the tP. This series is a expressible as E_P(u) = \frac{h^*(u)}{(1-u)^{d+1}}, where h^*(u) is the h^*- of P. The h^*-polynomial is h^*(u) = \sum_{k=0}^d h_k^* u^k, with non-negative integer coefficients satisfying h_0^* = 1. The evaluation h^*(1) (the sum of its coefficients) equals the normalized of P. The degree of h^*(u) is at most d. For reflexive polytopes, the h^*-polynomial is palindromic, satisfying h_k^* = h_{d-k}^* for $0 \leq k \leq d. The coefficients of the h^*-polynomial can be extracted from those of the Ehrhart polynomial via of the E_P(u), which isolates the numerator after accounting for the poles at the roots of unity in the denominator. For example, consider the standard d- \Delta_d = \operatorname{conv}\{ \mathbf{0}, \mathbf{e}_1, \dots, \mathbf{e}_d \}, whose Ehrhart polynomial is L_{\Delta_d}(t) = \binom{t+d}{d}. The corresponding Ehrhart series is E_{\Delta_d}(u) = \frac{1}{(1-u)^{d+1}}, so h^*(u) = 1 with h^*(1) = 1 = normalized volume of \Delta_d.

Rational polytope extensions

For a rational polytope P \subseteq \mathbb{R}^d with denominator q (the of the denominators of its vertex coordinates), the Ehrhart series is defined as the E_P(u) = \sum_{t=0}^\infty L_P(t) u^t, where L_P(t) denotes the number of integer lattice points in the dilation tP. This series extends the integer case by accounting for the fractional vertices, resulting in a whose coefficients reflect the quasi-polynomial nature of L_P(t). The Ehrhart series encodes the periodic behavior of lattice point counts through its denominator. The Ehrhart series E_P(u) takes the form of a rational function h^*(P; u) / (1 - u^q)^{d+1}, where h^*(P; u) is a in \mathbb{Z} of degree less than q(d+1) (with coefficients; a refined version has nonnegative coefficients). More generally, it can be expressed with a denominator that factors into a product \prod (1 - \zeta u)^{m_\zeta} over roots of unity \zeta, where the exponents m_\zeta capture the multiplicity of poles; this structure arises from the periodicity inherent in rational dilations. The periodicity in the series manifests through cyclotomic factors tied to the denominator q, as the quasi-polynomial coefficients of L_P(t) have periods dividing q, leading to a that isolates contributions from each cyclic component. An analogue to the h^*-polynomial for integer polytopes is the generalized h^*-quasi-polynomial for rational polytopes, denoted rh^*(P; t), which incorporates periodic coefficients to encode the Ehrhart series numerator while preserving nonnegativity and combinatorial interpretations. This refinement allows the series to be rewritten as E_P(u) = rh^*(P; u) / \prod_{\zeta} (1 - \zeta u)^{m_\zeta}, emphasizing the quasi-polynomial periodicity. A concrete example is the rational triangle T with vertices (1/2, 1/2), (1, 1/2), and (1/2, 1), which has denominator 2 and thus period 2. The Ehrhart is L_T(t) = (1/8)t^2 - 1/8 for odd t and L_T(t) = (1/8)t^2 + (3/4)t + 1 for even t. The corresponding Ehrhart series is E_T(u) = h^*(T; u) / (1 - u^2)^{3}, where h^*(T; u) is a of degree less than 6; its reveals terms aligned with the roots of unity of order 2, such as poles at u = \pm 1. Key properties of E_P(u) for rational polytopes include its rationality, with poles exclusively at roots of unity (reflecting the periodic fluctuations in L_P(t)), and an order of d+1 at the principal pole u=1, analogous to the integer case but modulated by the denominator q. For Gorenstein* rational polytopes, the numerator rh^*(P; u) exhibits palindromicity, providing symmetry in the periodic coefficients.

Advanced Applications

Toric varieties

A normal X_\Sigma associated to a polytope P \subset \mathbb{R}^d is constructed from the normal fan \Sigma_P of P, which consists of cones dual to the faces of P. The fan \Sigma_P lies in the real span of the N_\mathbb{R}, and the X_\Sigma is obtained by gluing affine toric varieties U_\sigma = \operatorname{Spec}(\mathbb{C}[S_\sigma]) for each cone \sigma \in \Sigma_P, where S_\sigma is the dual semigroup to \sigma. This construction endows X_\Sigma with an action of the algebraic (\mathbb{C}^*)^d, which acts on the dense open orbit corresponding to the zero-dimensional cone. The polytope P determines an ample Cartier D_P on X_\Sigma, whose support is the toric . The Ehrhart polynomial of P connects directly to the geometry of X_\Sigma through . Specifically, Danilov's theorem states that the number of points in the m-dilate mP equals the \chi(X_\Sigma, \mathcal{O}(mD_P)) of the associated to mD_P. The for these counts, the Ehrhart series, has numerator the h^*-polynomial of P, which coincides with the h-polynomial of X_\Sigma. This h-polynomial is the Poincaré series \sum_i \dim H^{2i}(X_\Sigma, \mathbb{Q}) t^i, reflecting the even-degree dimensions determined by the fan structure via the Danilov-Jurkiewicz theorem. For example, consider the standard n-simplex \Delta^n = \operatorname{conv}(0, e_1, \dots, e_n) in \mathbb{R}^n. Its normal fan yields the toric variety X_\Sigma = \mathbb{P}^n, the , with the torus action inherited from the standard coordinates. The Ehrhart polynomial of \Delta^n is \binom{m+n}{n}, and the h^*-polynomial is 1, matching the fact that \dim H^{2i}(\mathbb{P}^n, \mathbb{Q}) = 1 for i = 0, \dots, n and zero otherwise, with the binomial coefficients arising as the Hilbert polynomial of \mathcal{O}(m) on \mathbb{P}^n. Toric geometry provides tools to establish key properties of Ehrhart polynomials, such as of the h^*-vector. The hard Lefschetz theorem applied to the intersection of X_\Sigma implies that the toric h-vector—equivalent to the h^*-vector of P—is symmetric and unimodal, meaning the coefficients increase to a peak and then decrease symmetrically. This cohomological approach, building on Danilov's framework, yields algebraic proofs of combinatorial conjectures without relying on direct manipulations.

Generalizations

Ehrhart theory has been extended from classical polytopes to manifolds equipped with a into , known as the Ehrhart-Macdonald reciprocity. In this setting, a manifold is a triangulated manifold where each has integer vertices, and the Ehrhart polynomial counts points in the of the manifold, generalizing the case by incorporating boundary and interior contributions through a signed count. Specifically, for a compact oriented manifold M of d, the Ehrhart polynomial L_M(t) satisfies the L_{M^\circ}(-t) = (-1)^d L_M(t), where M^\circ is the interior, extending Ehrhart's original reciprocity for polytopes. Valuated polytopes generalize Ehrhart theory by incorporating valuations—additive functions on lattice polytopes that assign values to points, such as weights reflecting geometric or combinatorial properties. A valuation \phi on polytopes is -point if it commutes with integer translations, and McMullen's ensures that for such \phi, the function \phi(nP) is a of degree at most \dim P. The Betke-Kneser characterizes these valuations as spanning a of \dim P + 1, with basis elements corresponding to the coefficients of the Ehrhart , allowing weighted counts like \sum_{x \in nP \cap \mathbb{Z}^d} w(x) where w is a weight function on points. This framework unifies discrete volume and Ehrhart polynomials, enabling applications to tensor-valued invariants. Local h*-polynomials extend Ehrhart theory to non-convex polytopes and singular settings via polyhedral subdivisions. For a polytope P with a subdivision S, the local h*-polynomial h^*(P, S; u, v) refines the classical h*-polynomial by encoding contributions from faces and links in the subdivision, with coefficients that are nonnegative integers as proven by Karu. In non-convex cases, where P may have holes or singularities, S decomposes P into convex pieces, and the h^*(P, S; t) = h^*(P, S; t, 1) captures the Ehrhart series through relations like Ehr(P, x) = h^*(P, S; x, 1)/(1-x)^{\dim P + 1}, providing a combinatorial robust to triangulation choices. This is particularly useful for polytopes or singular toric varieties, where classical Ehrhart fail. Post-2000 developments have introduced quasi-polynomials in weighted Ehrhart theory for and weighted counts, broadening the scope beyond dilations. In weighted Ehrhart theory, the counting function \sum_{x \in nP \cap \mathbb{Z}^d} w(x) yields a quasi-polynomial when weights w are polynomials, with nonnegative h*-coefficients under suitable conditions, generalizing Stanley's nonnegativity . For , such as toric stacks, the weighted delta-vector coefficients correspond to dimensions of orbifold groups, linking combinatorial counts to ; for example, the weighted Ehrhart series of a relates to the orbifold Poincaré polynomial via a change-of-variables . These quasi-polynomials satisfy reciprocity laws analogous to the classical case, with applications to motivic integration and stringy invariants. An illustrative example arises in graphs and posets through order polytopes, where the Ehrhart polynomial counts order ideals combinatorially. For a finite poset P with p elements, the order polytope O(P) is the set of points in [0,1]^p corresponding to order-preserving maps, and its Ehrhart polynomial Ehr(O(P), m) equals the order polynomial \Omega_P(m+1), which enumerates the number of order ideals in the product poset P \times [m+1]. For a chain poset of length k, Ehr(O(I_k), x) = 1/(1-x)^{k+1}, reflecting binomial coefficients for ideals. This connects Ehrhart theory to poset enumeration, with applications to distributive lattices and Young diagrams.

References

  1. [1]
    [PDF] Coefficients and roots of Ehrhart polynomials - MIT Mathematics
    The Ehrhart polynomial of a convex lattice polytope counts integer points in integral dilates of the polytope. We present new linear inequalities satisfied by ...
  2. [2]
    Ehrhart polynomials - SpringerLink
    Eugène Ehrhart, Sur les polyèdres rationnels homothétiques à n dimensions, C. R. Acad. Sci. Paris 254 (1962), 616–618. MATH MathSciNet Google Scholar.
  3. [3]
    [PDF] Ehrhart Theory for Lattice Polytopes - Mathematics
    The final major theorem about Ehrhart series and polynomials that we will men- tion is a reciprocity theorem. Definition 2.28. For a d-dimensional polytope P in ...
  4. [4]
    [PDF] Ehrhart Polynomials - matthias beck
    § In applications, the volume of the polytope represented by a linear system measures some fundamental data of this system (“average”). Ehrhart Polynomials.
  5. [5]
    EHRHART POLYNOMIALS FOR PRECISE PROGRAM ANALYSIS
    Ph. Clauss, Counting Solutions to Linear and Nonlinear Constraints through Ehrhart polynomials: Applications to Analyze and Transform Scientific Programs, ...
  6. [6]
    On Ehrhart polynomials and probability calculations in voting theory
    Aug 9, 2025 · The purpose of this paper is threefold. Firstly, we want to do justice to Eugène Ehrhart, who, more than forty years ago, discovered the ...
  7. [7]
    [PDF] Computing the Continuous Discretely - matthias beck
    Jun 19, 2020 · Matthias Beck and Dennis Pixton, The Ehrhart polynomial of the Birkhoff polytope,. Discrete Comput. Geom. 30 (2003), no. 4, 623–637, arXiv ...
  8. [8]
    [0801.4432] A bijective proof for a theorem of Ehrhart - arXiv
    Jan 29, 2008 · The proof involves a geometric bijection, inclusion-exclusion, and recurrence relations, and we also prove Ehrhart reciprocity using these ...
  9. [9]
    [PDF] 1 The Ehrhart Polynomial
    To prove the power of this theorem, we can easily derive the classical Pick's theorem as a special case. Theorem 1.0.2 (Pick). Let n = 2 and P ⊂ R2 be a lattice ...
  10. [10]
    [PDF] VOLUMES AND EHRHART POLYNOMIALS OF CONVEX ...
    Now all ars an park, and there will be one empty spa e. A is a parking fun tion if and only if the empty spa e is n + 1. If A = (a1;:::;an) leads to.Missing: vol_k | Show results with:vol_k
  11. [11]
    [1006.5612] Rational Ehrhart quasi-polynomials - arXiv
    Jun 29, 2010 · Ehrhart's famous theorem states that the number of integral points in a rational polytope is a quasi-polynomial in the integral dilation factor.
  12. [12]
    Rational Ehrhart quasi-polynomials - ScienceDirect.com
    Ehrhart. Sur les polyèdres rationnels homothétiques à n dimensions. C. R. Acad. Sci. Paris Ser. A, 254 (1962), pp. 616-618. Google Scholar. [10]. C. Haase, T.B. ...<|control11|><|separator|>
  13. [13]
  14. [14]
  15. [15]
    [PDF] Maximal periods of (Ehrhart) quasi-polynomials
    Prominent examples of quasi- polynomials appear in Ehrhart's theory as integer-point counting functions for rational polytopes, and McMullen gives upper bounds ...
  16. [16]
    [PDF] Valuations on Lattice Polytopes
    The basic real valued valuations, the coefficients of the Ehrhart polynomial, are introduced and their characterization by Betke & Kneser is discussed. More ...
  17. [17]
    [PDF] Ehrhart Theory and Unimodular Decompositions of Lattice Polytopes
    Ehrhart theory studies the behaviour of lattice points contained in dilates of lattice poly- topes. We provide an introduction to Ehrhart theory.
  18. [18]
  19. [19]
    [PDF] Decompositions of rational convex polytopes - MIT Mathematics
    DECOMPOSITIONS OF RATIONAL CONVEX POLYTOPES. Richard P. STANLEY*. Department of Mathematics, Massachusetts Institute of Technology, Cambridge, MA 02139 ...
  20. [20]
    [PDF] Computing the Continuous Discretely - matthias beck
    The mathematical interplay between polytopes and lattices comes to life when we study the relationships between the continuous volume of a polytope and.
  21. [21]
    [PDF] Rational Ehrhart Theory
    Ehrhart. “Sur les polyèdres rationnels homothétiques à n dimensions”. C. R. Acad. Sci. Paris 254 (1962), pp. 616–618. [11] M. H. J. Fiset and A. M. Kasprzyk ...
  22. [22]
    [PDF] Periods of Ehrhart coefficients of rational polytopes
    Apr 7, 2016 · We also establish Theorems 1.3 and 1.4 in the dimension n = 2 case. We are interested in the period sequences of quasi-polynomials. This period ...<|control11|><|separator|>
  23. [23]
    [PDF] maximal periods of (ehrhart) quasi-polynomials - UC Davis Math
    Prominent examples of quasi- polynomials appear in Ehrhart's theory as integer-point counting functions for rational polytopes, and McMullen gives upper bounds ...
  24. [24]
    [PDF] Rational Ehrhart Theory - matthias beck
    There is an Ehrhart theory for the quasipolynomial ehrQ(P;λ) in the rational (equivalently, real) variable λ. Examples ehrQ([1,2];λ) = b2λc−dλe ...
  25. [25]
    [PDF] Classification of Ehrhart quasi-polynomials of half-integral polygons
    In the 1960s, Eugene Ehrhart developed Ehrhart theory to enumerate lattice points in convex polytopes. An important tool in Ehrhart theory is the Ehrhart ...
  26. [26]
    [PDF] Toric Varieties David Cox John Little Hal Schenck - mimuw
    Jan 1, 2017 · Page 1. Toric Varieties. David Cox. John Little. Hal Schenck ... These are the chapters included in the version you downloaded. The ...
  27. [27]
    THE GEOMETRY OF TORIC VARIETIES - IOPscience
    THE GEOMETRY OF TORIC VARIETIES. V I Danilov. © 1978 The British Library and The London Mathematical Society Russian Mathematical Surveys, Volume 33, ...
  28. [28]
    [PDF] Toric varieties, characteristic classes and lattice point counting
    Counting lattice points. Pick's formula. Danilov: #(P ∩ M) = χ(XP,O(DP)).
  29. [29]
    [PDF] arXiv:math/0111331v2 [math.CO] 3 Jun 2003
    an Ehrhart-type reciprocity law. This generalizes the classical ... [3] I. G. Macdonald, Polynomials associated with finite cell complexes,. J ...
  30. [30]
    [PDF] A Brief Introduction to Valuations on Lattice Polytopes
    Oct 1, 2018 · Theorem 1.4 (Betke-Kneser Theorem [6]). The family of real-valued, lattice- invariant valuations on lattice polytopes is a (d+1)-dimensional ...
  31. [31]
    Tensor valuations on lattice polytopes - ScienceDirect
    Oct 15, 2017 · The Ehrhart polynomial and the reciprocity theorems by Ehrhart & Macdonald are extended to tensor valuations on lattice polytopes.<|control11|><|separator|>
  32. [32]
    None
    Summary of each segment:
  33. [33]
    [PDF] Ehrhart Series and Lattice Triangulations - UT Math
    May 11, 2007 · Here we use the notation h∗, following Stanley [20] and Athanasiadis [1], to emphasize the analogy with h-polynomials of simplicial polytopes ...
  34. [34]
    Weighted Ehrhart theory and orbifold cohomology - ResearchGate
    We introduce the notion of a weighted delta-vector of a lattice polytope. Although the definition is motivated by motivic integration, we study weighted ...
  35. [35]
    None
    ### Summary of Ehrhart Polynomials of Order Polytopes