Given the function $Z(s,N)= \sum \limits_{n=1}^{N}n^{-s}$.

In the limit $N \to \infty$ the function $Z(s,N) \to \zeta (s)$ Riemann Zeta function.

My question is: Is there a Functional equation for this function? I mean a relationship of the form $ Z(s,N)=G(s) Z(1-s,N)$.

  • 7
    $\begingroup$ $\text{}{}n=0$? $\endgroup$ – anon Dec 2 '11 at 23:54
  • 2
    $\begingroup$ "... in the limit $N \to \infty$ ...": this claim is only true if $\Re(s) > 1$. $\endgroup$ – Matt E Dec 3 '11 at 2:10
  • $\begingroup$ OH. sorry :( i meant n=1 :) otherwise the series would be divergent ... $\endgroup$ – Jose Garcia Dec 3 '11 at 11:28
  • $\begingroup$ Edited in accord with comments above. $\endgroup$ – Gerry Myerson Feb 27 '12 at 22:27

Well, here's an answer - no, because then $G(s)=Z(s,N)Z(1-s,N)^{-1}$ would be invariant with respect to $N$, but it clearly isn't (check the graphs for various $N$ on W|A if you want to, or just plug in the trivial $N=1$ case). If you let $G$ vary with $N$ then of course there are such functions, but I doubt there is a nice formula that covers all the cases. The fact that a "$p=\infty$" factor can be added to $\zeta(s)$ to "complete" it so it obtains symmetry about $\operatorname{Re}(s)=1/2$ derives from the functional equation of the theta function (see my previous answer), and I don't think this can even have a finite analogue here because the Poisson summation formula requires a lattice to work with (of the discrete subgroup, not poset, variety), which will necessarily be infinite in the Euclidean setting.

There is one other thing I want to point out. The functional equations for $L$ functions are more suggestive in forms like $\chi(1-s)L(1-s)=\chi(s)L(s)$ because this showcases the "completed" zeta function, which may be a more "natural" object than $L$ itself, despite first appearances. As I hinted at earlier on the "$p=\infty$" factor $\pi^{-s/2}\Gamma(s/2)$, there is an underlying adelic framework that explains this sort of completion for Riemann's zeta (see this or other discussions of Tate's thesis).

Also, $Z(s,N)$ is typically denoted $H_{N,s}$ as a generalization of the harmonic numbers.

As Gerry notes, there is an approximate reflection equation given by the Riemann-Siegel formula:

$$\zeta(s)=\sum_{n=1}^N \frac{1}{n^s}+\pi^{1/2-s}\frac{\Gamma\left(\frac{s}{2}\right)}{\Gamma\left(\frac{1-s}{2}\right)}\sum_{k=1}^M\frac{1}{k^{1-s}}-\frac{\Gamma(1-s)}{2\pi i}\oint_{\gamma(2\pi M)}\frac{(-z)^{s-1}e^{-Nz}}{e^z-1}dz $$

See Wikipedia or MathWorld.


There is something like a functional equation in Section 3 of http://www.stanford.edu/group/journal/2005/pdfs/Carl.pdf, although on my screen a lot of characters are missing - maybe they print correctly but just don't show up on my screen.


there is the "approximate functional equation" https://en.wikipedia.org/wiki/Riemann%E2%80%93Siegel_formula

which is described in the Tichmarsh and on the french wikipedia of zeta-Riemann function but not on the english one. someone should add a link to that Riemann Siegel formula to the wiki.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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