# Closed form for (in)finite sums

1. Apr 17, 2013

### henpen

I have a set of questions concerning the perennial sum
$$\large \sum_{k=1}^{n}k^p$$
and its properties.

1. For $p \ge 0$, the closed form of this is known (via Faulhaber's formula).
I know little about divergent series, but I've read that in some sense there exists a value associated with these sums as $n \rightarrow \infty$: for instance $\lim_{n \rightarrow \infty}\sum_{k=1}^{n}k = \frac{-1}{12}$ (here)? I think this is called zeta function regularisation.

a. What is the significance of such sums? Are they useful realistic compatable-with-rest-of-mathematics, or wholly artificial? What evidence is there of these pseudosums' existence?

b. What are they for $p \ge -1$ (hoping that the Harmonic series, $p = -1$, has such a 'pseudo-sum')?

2. Moving on to $p < 0$. For even $p$ the limit as $n \rightarrow \infty$ infinity is known (Euler).

a. Is there a Faulhaber equivalent in this case (a closed form of the sum up to $n$ for either even or integer $n$)?

b. Are the odd $p$'s infinite series likely to be found?

For comparison, with constant $k$ and changing $p$:
$$\large \sum_{p=1}^{n}k^p= \frac{1-k^{p+1}}{1-k}$$
I would love if for constant $p$ and changing $k$ was a formula as complete as this.

Thank you for volunteering to suffer my question.

Last edited: Apr 17, 2013