8
$\begingroup$

For $n \in \mathbb{Z}_{\geq 0}$, let $[n]_q := (1-q^n)/(1-q) = (1+q+...+q^{n-1})$ as is customary, with $[0]_q=0$.

Let $R$ be the subring of $\mathbb{Q}(q)[x]$ consisting of all $f$ such that $f([n]_q) \in \mathbb{Z}[q]$ for all $n \in \mathbb{Z}_{\geq 0}$.

Define $f_0(x) = 1$ and $f_k(x) = f_{k-1}(x)\cdot \frac{x-[k-1]_q}{q^{k-1}[k]_q}$ for $k \geq 1$. Note $f_k([n]_q) = \frac{[n]_q!}{[n-k]_q! [k]_q!}=\binom{n}{k}_q \in \mathbb{Z}[q]$, so indeed $f_k \in R$ for all $k \geq 0$.

Is it the case that $R$ is spanned as a $\mathbb{Z}[q]$-module by the $f_k$? If so, what is known about the structure constants $c_{ij}^{k}(q)$?

EDIT: More information about this ring, such as the structure constants and a classification of all maps to fields, can be found in this preprint with Nate Harman: http://arxiv.org/abs/1601.06110v1.

$\endgroup$
8
  • $\begingroup$ This answer seems to say the classical structure constants $c_{ij}^{k}(1)$ have a simple form as a product of binomials: math.stackexchange.com/questions/1289526/…. If so, maybe the $c_{ij}^{k}(q)$ are just the obvious $q$-ifications? $\endgroup$ Commented Sep 19, 2015 at 16:10
  • $\begingroup$ Another implicit part of my question: have these $q$-deformations of numerical polynomials been studied anywhere? $\endgroup$ Commented Sep 19, 2015 at 18:21
  • $\begingroup$ Minor comment: We should probably take $\mathbb{Q}(q)$ instead of $\mathbb{Q}[q,q^{-1}]$ since the denominators aren't just powers of $q$. $\endgroup$
    – Nate
    Commented Sep 19, 2015 at 20:40
  • $\begingroup$ @Nate: Of course you are right. I always mix up this notation. $\endgroup$ Commented Sep 19, 2015 at 20:59
  • $\begingroup$ @SamHopkins does not the same trick with consecutive substitution of q-integers (as on the link) allow to calculate structure constants? $\endgroup$ Commented Sep 19, 2015 at 21:14

1 Answer 1

7
$\begingroup$

(Below is the proof that module $R$ is generated by $f_i(x)$, without calculation of structure constants.)

Polynomial $f(x)$ of degree $n$ may be interpolated in points $[0],[1],\dots,[n]$ (I omit index $q$): $f(x)=\sum_{i=0}^n f([i])\prod_{j\ne i} \frac{x-[j]}{[j]-[i]}$. For any summand leading term $f([i])\prod_{j\ne i} \frac1{[j]-[i]}$ is divisible in $\mathbb{Z}[q]$ by the leading term of $f_n(x)$. Hence $f(x)-Af_n(x)$ has degree less than $n$ for some $A\in \mathbb{Z}[q]$. This is induction step in the proof that $f$ is $\mathbb{Z}[q]$-linear combination of $f_0,f_1,\dots,f_n$.

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .