29
$\begingroup$

Compute $$\int_0^ex^{1/x}\;\mathrm dx.$$

There is an analytical anti-derivative found in this answer. How does one compute this?

Using the anti-derivative approach we have

$$\int x^{1/x}\;\mathrm d x=x + \frac{\log^2x}{2}-\sum^\infty_{n=2}\sum^n_{k=0}\frac{\log^{n-k}x\;}{x^{n-1}(n-k)!(n-1)^{k+1}}+C$$

Now there is a problem for this anti-derivative as it gets near $0$ (apparent from the $\log$ s). I do not know how to prove this but if $f(x)=x^{1/x}$ and $\lim_{x\to0}f(x)=0$ can be defined then I want to say that $\lim_{x\to0}F(x)=0$. Assuming my non-rigorous logic and some logic from this question we have $$\int_0^ex^{1/x}\;\mathrm dx=\sum _{n=2}^{\infty } \frac{ \Gamma (n+1,\;n-1)}{(n-1)^{n+1}\;n!}+\frac{1}{2}+e$$


Update: Using Mathematica I computed some numerical integrals to find that $F(0)\approx1.53328$. Therefore we have $$\int_0^ex^{1/x}\;\mathrm dx=\sum _{n=2}^{\infty } \frac{ \Gamma (n+1,\;n-1)}{(n-1)^{n+1}\;n!}+\frac{1}{2}+e-\lim_{x\to0}\int x^{1/x}\;\mathrm dx$$

$\endgroup$
2
  • 1
    $\begingroup$ The number is tabulated in oeis.org/A175994 which cites a paper with a graph of the integrad. $\endgroup$
    – user113880
    Commented Dec 6, 2013 at 14:36
  • $\begingroup$ Mathematica 9 gives $2.661825705380417828497039337651395830214970820983303548214678485091470210657175166246828293562435140 \cdots$ $\endgroup$ Commented Jun 17, 2014 at 6:54

3 Answers 3

2
$\begingroup$

Here is a way a physicist is approaching the problem.

First, the variable change $x=\frac{1}{y}$ gives:

$$\int_0^ex^{1/x}\;\mathrm dx=\int_{\frac{1}{e}}^{\infty}\frac{y^{-y}}{y^2}dy$$

Now, as a first approximation, we use the following inequality:

$$\int_{\frac{1}{e}}^{\infty}\frac{y^{-y}}{y^2}dy<\int_{\frac{1}{e}}^{\infty}\frac{y^{-\frac{1}{e}}}{y^2}dy=\int_{\frac{1}{e}}^{\infty}y^{-\frac{1}{e}-2}dy=\frac{e^{2+\frac{1}{e}}}{e+1}$$

Finally:

$$0<\int_0^ex^{1/x}\;\mathrm dx<\frac{e^{2+\frac{1}{e}}}{e+1};\;e>0$$

This result can be probably improved.

$\endgroup$
0
0
$\begingroup$

ISC for 2.66182570538 yields only $$ {\frac {7}{8}}+\sum _{n=1}^{\infty }\frac{60}{ 6\,{5}^{n}+8\,{n}^{3}- 3\,{n}^{2}+7\,n } $$ which is perhaps simpler than your double sum.

$\endgroup$
2
  • $\begingroup$ Thank you for the answer but I am trying to compute this without numerical approximation. Do you reckon that the answer can be found analytically? $\endgroup$ Commented Nov 10, 2013 at 18:05
  • 6
    $\begingroup$ This unfortunately does not converge to the OP's integral. The OP's converges to $$2.6618257053804178284970...$$ while this converges to $$2.661825705395275140900...$$ They disagree on the last digit of your ISC search (and it's odd and unfortunate that your summation, in fact, disagrees with ISC's...) $\endgroup$
    – user98602
    Commented Jan 6, 2014 at 1:09
-2
$\begingroup$

The answer is approximately 2.66183. I hope this helps. I evaluated it, but I am new to the website and wasn't sure how to write out the calculations.

$\endgroup$
1
  • $\begingroup$ This answer was flagged for not being an answer, but retracted in case it would have been declined. It can be a comment instead $\endgroup$ Commented Feb 20 at 20:01

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .