In a recent MO question the author asks: Could the Riemann zeta function be a solution for a known differential equation?

I don't understand why the answer isn't a trivial yes? For example, $y' = -\sum log(n)/n^s $ is an equation with y being the zeta function. It's easy to come up with many trivial examples of this. What am I misunderstanding?

  • 2
    $\begingroup$ The differential equation should not contain any nontrivial functions of $s$ except via $y$ and its derivatives. $\endgroup$ – Count Iblis Jan 22 '17 at 22:16
  • $\begingroup$ Here is a link to the MO post the OP is (probably) talking about, for reference: mathoverflow.net/questions/260196/… $\endgroup$ – A.Sh Jan 22 '17 at 22:19
  • $\begingroup$ Yes, it's really easy to make more. $\endgroup$ – mtheorylord Jan 22 '17 at 22:44
  • $\begingroup$ For example the $\Gamma$ function is a solution for a very basic functional equation $\Gamma(s+1) = s \Gamma(s)$. Quite similarly, the Riemann zeta function is a solution for $\zeta(s) = \chi(s) \zeta(1-s)$ where $\chi(s) = \pi^{s-1} 2^s \sin(\pi s/2) \Gamma(1-s)$. So those functions obeys to some functional equations, but we don't know any similar differential equations for them. $\endgroup$ – reuns Jan 22 '17 at 22:44

The zeta function is trivially a solution to many differential equations

You are correct that we can construct differential equations for which $\zeta(s)$ is a solution in this manner. But what you have. . .

$y'(s) = \displaystyle \sum_{n=1} ^\infty \frac{\log(n)}{n^s}$

. . . is quite a complicated equation. Much more complicated that, for example, a finite linear equation like. . .

$A_n(s)y^{(n)} + A_{n-1}(s)y^{(n-1)}+ \ldots + A_{1}(s)y' + A_{0}(s)y(s) = 0$

. . . where $A_i(s)$ are some functions. The paper points out that $\zeta(s)$ does not solve any finite linear equation. Then it asks how complicated we need to make an equation in order for $\zeta(s)$ to solve it. Does it need to be as complicated as yours or can we do better?

The answer given in the paper is that $\zeta(s)$ solves an infinite linear equation of the form

$ \ldots + A_n(s)y^{(n)} + A_{n-1}(s)y^{(n-1)}+ \ldots + A_{1}(s)y' + A_{0}(s)y(s) = 0$.

This equation is simpler than yours in that there are no functions of $s$ other than $y(s)$ and its derivatives. However it is more complicated in the sense that it includes the higher derivatives of $y(s)$.

  • $\begingroup$ $(s-1) \zeta(s)$ solves such an infinite differential equation, as any entire function ? $\endgroup$ – reuns Jan 22 '17 at 22:48

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.