4
$\begingroup$

Are there any studies about this function? $$f(x,n)=\sum_{k=1}^{n} x^{1/k}=x+x^{1/2}+x^{1/3}+x^{1/4}+\cdots +x^{1/n}$$

EDIT:

My first notes about it.

$f(1,n)=n$

$f'(1,n)=H_n$

$\int_0^1 \frac{f(x,n)}x\cdot dx=\frac{n^2+n}2$

where $f'(x,n)=\frac{d}{dx}f(x,n)$

To avoid fractional powers we can let $x=y^{n!}$ to change it into integer-powers, but the obtained polynomial does not have uniform paterm of powers.

$$f(y,n)=\sum_{k=1}^{n}y^{n!/k}=y^{n!}+y^{n!/2}+y^{n!/3}+y^{n!/4}+\cdots+y^{n!/n}$$ $$=y^{n!/n}\left(y^{n!-(n!/n)}+y^{(n!/2)-(n!/n)}+y^{(n!/3)-(n!/n)}+y^{(n!/4)-(n!/n)}+\cdots+y^{(n!/(n-1))-(n!/n)}+1\right)$$

As example let $n=4$ $$f(y,4)=y^6(y^{18}+y^6+y^2+y+1)$$

$\endgroup$
5
  • $\begingroup$ I don't know about you, but using $a/b$ in exponents is more readable, since the exponents aren't very complicated here. $\endgroup$ Oct 18 '16 at 0:08
  • $\begingroup$ You did the right edit my friend @SimpleArt $\endgroup$
    – Pentapolis
    Oct 18 '16 at 0:09
  • $\begingroup$ I recommend you make it clear that $f'(1,n)$ refers to $\frac{d}{dx}f(x,n)$. It's not impossible to differentiate with respect to $n$ if one allows such extension. $\endgroup$ Oct 18 '16 at 0:09
  • $\begingroup$ $f(x,n)\sim n$ as $n\to\infty$ $\endgroup$ Oct 18 '16 at 0:16
  • $\begingroup$ $$\int_1^{n+1}x^{1/z}dz\le f(x,n)\le n+\int_1^nx^{1/z}dz$$for $x>1$ by the integral test. $\endgroup$ Oct 18 '16 at 0:27

Your Answer

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

Browse other questions tagged or ask your own question.