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.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

How to solve the q-shift difference equation $W(x)=(x-a)W(xq^{-2})$? Here $W(x)$ is the unknown function. Thank you very much.

share|cite|improve this question
Have you tried using that equation like a fixed-point equation, iterating and trying to see if the process converges? – Mariano Suárez-Alvarez Jan 17 '12 at 19:21
@Mariano, the process will run forever. – LJR Jan 17 '12 at 19:27
That is exactly why I suggested to see of the process converges. I very much doubt there is a solution to the equation in finite terms! – Mariano Suárez-Alvarez Jan 17 '12 at 19:30
Do ypu understand what 'see if the process converges' means? – Mariano Suárez-Alvarez Jan 17 '12 at 19:35
No, that's not what I mean. You should review —in a calculus book— the properties of infinite products (it usually comes right after those of infinite series). The little book by Knopp titled Theory and application of infinite series is a great source. – Mariano Suárez-Alvarez Jan 17 '12 at 19:48
up vote 1 down vote accepted

Assume $a\neq 0$ and $q>1$. We can define $G(u) = a^{-u} f(u) W(q^{2u})$, where $f(u)$ is a function satisfying $f(u+1)=-f(u)$. Then $$ \begin{eqnarray} G(u) &=& a^{-u}f(u)W(q^{2u}) \\ &=& a^{-u}f(u)(q^{2u} - a)W(q^{2u}q^{-2}) \\ &=& a^{-(u-1)}f(u-1)(1-a^{-1}q^{2u}) W(q^{2(u-1)}) \\ &=& (1-a^{-1}q^{2u})G(u-1). \end{eqnarray} $$ This gives $G(u)$ as a convergent infinite product times a scale factor (equal to $\lim_{u\rightarrow-\infty}G(u)$), which we will take to be $1$: $$ G(u) = (1-a^{-1}q^{2u})(1-a^{-1}q^{2(u-1)})\cdots = (a^{-1}q^{2u};q^{-2})_{\infty}, $$ where $(\alpha;\beta)_\infty$ is the q-Pochhammer symbol. Now, taking $x=q^{2u}$, we can invert this to obtain $$ \begin{eqnarray} W(x) = \frac{a^{u}}{f(u)}\left(\frac{x}{a};q^{-2}\right)_{\infty}=e^{(i\pi+\ln a)u}\left(\frac{x}{a};q^{-2}\right)_{\infty}, \end{eqnarray} $$ choosing $f(u)=e^{-i\pi u}$, or $$ W(x) = \exp\left({\frac{i\pi + \ln a}{2 \ln q}\cdot\ln x}\right)\left(\frac{x}{a};q^{-2}\right)_{\infty}. $$ This is a particular solution, given by a particular choice of $f$. To obtain the general solution, multiply $W(x)$ by any function of $u=\ln x / (2 \ln q)$ that is periodic with period $1$.

share|cite|improve this answer
thank you very much. Why do you know how to define $G(u)$ to solve this problem? – LJR Jan 18 '12 at 16:37
The change of variable $x=q^{2u}$ turns the multiplicative shift into an additive shift. You then have an infinite-looking product, but its terms approach $(-a)$ instead of $1$ (which they need to approach for the product to converge). Factoring out the $(-a)$, in the form of $a^{-u}f(u)$, leaves the remaining product convergent. – mjqxxxx Jan 18 '12 at 16:54

Your Answer


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.