Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Question: What is the correct notion of a product of integral (or rational) polytopes which induces a factorization of its Ehrhart (quasi-)polynomial into two primitive Ehrhart (quasi-)polynomials corresponding to its constituent polytopes, viz., $L_{P \times Q}(t) = L_{P}(t) L_{Q}(t)$?

(Motivation) Given two closed integral polytopes $P$ and $Q$ each with vertices at $\{ \mathbf{0} , b_{1} \mathbf{e}_{1}, \dots, b_{n} \mathbf{e}_{n} \}$ and $\{ \mathbf{0} , d_{1} \mathbf{e}_{1}, \dots, d_{m} \mathbf{e}_{m} \}$, respectively, where $n, b_{i}, m, d_{j} \in \mathbb{N}$, define the integral polytope $R$ with vertices at $\{ \mathbf{0}, b_{1} \mathbf{e}_{1}, \dots, b_{n} \mathbf{e}_{n}, d_{1} \mathbf{e}_{n+1}, \dots, d_{m} \mathbf{e}_{n+m} \}$.

The above construction cannot be the sought after product $P \times Q$. Suppose $P$ and $Q$ are defined by $b_{1} = b_{2} = d_{1} = d_{2} = 2$. It is easy to show that $L_{P}(1) = L_{Q}(1) = 6$. Define $R$ as above with vertices of $P$ and $Q$. It is true that $L_{R}(1) = 15 \neq 6^{2}$.

Question: What is $R$ in terms of $P$ and $Q$? Is it special in some way?

Thanks!

share|improve this question

2 Answers 2

up vote 4 down vote accepted

Isn't this obvious (or have I missed something?)

The appropriate product of rational polytopes $P\subseteq\mathbb R^m$ and $Q\subseteq\mathbb R^n$ is surely their Cartesian product $P\times Q\subseteq\mathbb R^{m+n}$. For a positive integer $t$, the integer points in $t(P\times Q)$ are those of the form $(a,b)$ where $a\in tP$ and $b\in tQ$, so there are $L_P(t)L_Q(t)$ of them.

share|improve this answer
1  
Thanks for the post! Two questions: What are the $n m$ vertices of the cartesian product $P \times Q$? Does this factorization continue to hold for rational polytopes and their quasi-polynomials (whose coefficient functions may have distinct periods)? –  user02138 Nov 23 '10 at 16:14
    
The vertices of $P\times Q$ are $(a,b)$ where $a$ is a vertex of $P$ and $b$ is a vertex of $Q$. In this answer I didn't assume that $P$ and $Q$ had integer point vertices. –  Robin Chapman Nov 23 '10 at 16:57

The (Motivation) description appears to be similar to a direct sum (a.k.a. tegum product) as defined in Section 1.2 of this paper:

http://people.reed.edu/~davidp/homepage/seniors/mcmahan.pdf

The difference is that for a direct sum, 0 is in the relative interior of both of the factor polytopes, rather than a common vertex.

The same article answers your question about what the nm vertices of a Cartesian product are.

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.