Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

While extending my calculation techniques, with aid of Mathematica, I found that

\begin{align*} \int_{0}^{1}\left( \frac{1}{\log x} + \frac{1}{1-x} \right)^{3} \, dx &= -6 \zeta '(-1) -\frac{19}{24}, \\ \int_{0}^{1}\left( \frac{1}{\log x} + \frac{1}{1-x} \right)^{4} \, dx &= -10 \zeta '(-2)-2 \zeta '(-1)-\frac{37}{72}, \\ \int_{0}^{1}\left( \frac{1}{\log x} + \frac{1}{1-x} \right)^{5} \, dx &= -\frac{35}{3} \zeta '(-3)-\frac{15}{2} \zeta '(-2)-\frac{5}{3} \zeta '(-1)-\frac{3167}{8640}, \\ \int_{0}^{1}\left( \frac{1}{\log x} + \frac{1}{1-x} \right)^{6} \, dx &=-\frac{21}{2} \zeta '(-4)-14 \zeta '(-3)-\frac{31}{4} \zeta '(-2)-\frac{3}{2} \zeta '(-1)-\frac{1001}{3600}. \end{align*}

I conjectures that these relations extends also to higher degrees:

My Guess. For $m \geq 3$, we can write $$ \int_{0}^{1}\left( \frac{1}{\log x} + \frac{1}{1-x} \right)^{m} \, dx = - \Bigg( q + \sum_{k=1}^{m-2} q_k \zeta'(-k) \Bigg) $$ for some positive rational numbers $q$ and $q_k$.

Is there any reference regarding this problem?


Addendum. Following i707107's advice, I obtained the following formula

\begin{align*} & \int_{0}^{1}\left( \frac{1}{\log x} + \frac{1}{1-x} \right)^{m} \, dx \\ &= -\frac{H_{m-1}}{(m-1)!} + \frac{1}{(m-1)!} \sum_{k=1}^{m-1} \left[{{m-1}\atop{k}}\right] \zeta(1-k) \\ &\quad - \frac{1}{(m-2)!}\sum_{j=1}^{m-1}\sum_{l=m-j}^{m-1} \binom{m}{j} \binom{m-2}{j-1} \left[{{j-1}\atop{l+j-m}}\right] \{ \zeta'(1-l) + H_{m-j-1} \zeta(1-l) \} \end{align*}

valid for $m \geq 2$, where $\left[{{n}\atop{k}}\right]$ denotes the unsigned Stirling's number of the first kind.

share|improve this question
1  
When I clicked the question I said: surely sos440 will answer this... –  Pedro Tamaroff Mar 25 '13 at 15:28
1  
@PeterTamaroff, I also wished I could have answered it. :) –  sos440 Mar 25 '13 at 15:37
    
I know for $m=1$ it's an integral representation of Euler's constant. But how do you evaluate the integral for $m=2$? –  Random Variable Mar 25 '13 at 22:10
    
@RandomVariable, You can refer to my blog posting for the case $m=2$. Indeed, the first method is what I began with for the case $m\geq 3$. –  sos440 Mar 25 '13 at 23:57
    
Could you prove positivity of $q$ as in your guess? It seems hard to prove it because $\zeta$ values alternates sign. –  i707107 Mar 28 '13 at 15:31

1 Answer 1

up vote 5 down vote accepted

You can try $$ \int_0^1\left(\frac{1}{\log x}+\frac{1}{1-x}\right)^m (-\log x)^{s-1}dt $$ for $s$ with sufficiently large real part.

This will give you an expression involving $\Gamma$ and $\zeta$ functions. Then use the analytic continuation to $\sigma>0$ and plug in $s=1$.

I checked this method for $m=2$ and got the right answer. I am sure that this will work in $m\geq 3$.

share|improve this answer
    
I just grasped what you are trying to say. It seems illuminating! Thank you. I will try it. –  sos440 Mar 27 '13 at 1:48
    
Thanks! I obtained an analytic formula for this integral, which coincided with my previous calculation for other values of $m$ as well. Now it remains to identify the limit of the integral as $s \to 1$. –  sos440 Mar 27 '13 at 2:52

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.