0
$\begingroup$

I have an expression for the sum-of-divisors function defined as $$\sigma(n)=\sum_{d\mid n}d.$$ However I do not know how nontrivial or practical it actually is.

Let us define $$A=\frac{b}{2}-\frac{b}{\pi}\sum_{n=1}^\infty{\frac{\sin{\frac{2\pi nx}{b}}}{n}}=x\pmod b$$ where, when $\frac{x}{b}$ is a whole number, $A=\frac{b}{2}$

Now define $$B=\lim \limits_{n\to\infty}\frac{b}{2}-\frac{b}{\pi}\tan^{-1}\left(\cot\left(\frac{\pi x}{b}+\frac{1}{n}\right)\right)=x\pmod b$$ where, when $\frac{x}{b}$ is a whole number, $B=0$. The limit helps avoid division by zero in certain cases, but can be excluded in theory.

Now $$\sum_{b=1}^x A-\sum_{b=1}^x B=\frac{\sigma(x)}{2}.$$ So then $$\sigma(x)=2\left(\sum_{b=1}^x A-\sum_{b=1}^x B\right).$$

Can't this be considered a generating function for the $\sigma(x)$? If so, why hasn't it been seen before? Is this function somehow trivial?

$\endgroup$
  • $\begingroup$ I'm not sure that that's what "generating function" means. A(n ordinary) generating function of $\sigma(n)$ would be the function $\sum_{n=0}^\infty\sigma(n)x^n$. (There are other types of generating function, such as exponential or Dirichlet (di-ri-CLAY, sometimes di-ri-SHLAY).) $\endgroup$ – Akiva Weinberger Aug 31 '15 at 18:45
  • $\begingroup$ This classical Dirichlet series is a generating function for $\sigma_a$ (with $a=1$ in your case) :$$\zeta(s)\,\zeta(s-a)=\sum_{n=0}^\infty\frac {\sigma_a(n)}{n^s}$$ $\endgroup$ – Raymond Manzoni Aug 31 '15 at 19:26
1
$\begingroup$

A Generating function can be defined as a generating function is a formal power series in one indeterminate, whose coefficients contain information about a sequence of numbers $a_n$ which is indexed by the natural numbers. What you gave is not a formal power series and hence not a generating function. $\sigma(n)$ is an arithmetic function , its formal power series is the Dirichlet series which for an arithmetic function $f(n)$ is of the form $$ \ F(s)=\sum_{n=1}^{\infty}\frac{f(n)}{n^s}$$ And hence as Raymond Manzoni mentioned in his comment the generating function for $\sigma(n)$ is $$\zeta(s)\,\zeta(s-1)=\sum_{n=0}^\infty\frac {\sigma_1(n)}{n^s}$$

$\endgroup$

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.