4
$\begingroup$

In a textbook I'm reading, the author states without proof that $$ \zeta(s,mz) = \frac{1}{m^{s}} \sum_{k=0}^{m-1} \zeta \left(s,z+\frac{k}{m} \right), \tag{1}$$ where $\zeta(s,z) $ is the Hurwitz zeta function

Supposedly, this isn't hard to prove. But is it possible to prove $(1)$ using simply the series definition of the Hurwitz zeta function, that is, $ \displaystyle\zeta(s,z) = \sum_{n=0}^{\infty} \frac{1}{(z+n)^{s}}$?

It might be interesting to note that the polygamma functions (excluding the digamma function) can be expressed in terms of the Hurwitz zeta function.

So from $(1)$ we can derive the multiplication formula $$\psi_{n}(mz) = \frac{1}{m^{n+1}} \sum_{k=0}^{m-1}\psi_{n} \left(z+ \frac{k}{m} \right) , \quad n \in \mathbb{Z}_{>0}.$$

$\endgroup$
2
  • $\begingroup$ it is obvious ${}{}$ $\endgroup$
    – reuns
    Dec 11, 2016 at 3:20
  • $\begingroup$ Given any totally multiplicative series, whose Fourier transform exists, one can construct a function obeying the multiplication theorem.The Wikipedia article multiplication theorem explains why and how. $\endgroup$
    – Linas
    Jul 7, 2020 at 19:51

1 Answer 1

7
$\begingroup$

We start with $\displaystyle \zeta(s,z) = \sum_{n=0}^{\infty} \frac{1}{(z+n)^{s}}$.

First, substitute in $mz$ for $z$:

$$\begin{align} \zeta(s,mz) &= \sum_{n=0}^{\infty} \frac{1}{(mz+n)^{s}} \\ &= \frac{1}{m^s}\sum_{n=0}^{\infty} \frac{1}{(z+\frac{n}{m})^{s}}. \end{align}$$

Now write $n = n'm + n''$, where now $n'$ will range from $0$ to $\infty $ and $n''$ will range from $0$ to $m-1$. Putting this above gives $$\begin{align} &= \frac{1}{m^s}\sum_{\substack{n'\geq 0 \\ n'' \bmod m}} \frac{1}{(z+\frac{n''}{m} + n')^{s}} \\ &= \frac{1}{m^s} \sum_{k = 0}^{m-1} \sum_{n' \geq 0} \frac{1}{(z + \frac km + n')^s} \\ &= \frac{1}{m^s}\sum_{k =0}^{m-1}\zeta\left(s,z+\frac{k}{m}\right), \end{align}$$

as we wanted to show.

$\endgroup$
2
  • $\begingroup$ So you just used the division algorithm? $\endgroup$ Dec 8, 2013 at 11:20
  • $\begingroup$ @RandomVariable: Yes, that's how I split $n$ into $n'$ and $n''$. $\endgroup$
    – davidlowryduda
    Dec 8, 2013 at 12:12

You must log in to answer this question.

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