MHB Can You Solve These Two Difficult Integrals?

  • Thread starter Thread starter sbhatnagar
  • Start date Start date
  • Tags Tags
    Integrals
sbhatnagar
Messages
87
Reaction score
0
Prove that

\[\int_0^1 \frac{\log(1+x)\log(x)}{1-x}dx=\zeta(3)-\frac{\pi^2}{4}\log(2)\]

\[\int_0^1 \frac{\log(1+x^2)}{1+x}dx=\frac{3}{4}\log^2(2) -\frac{\pi^2}{48}\]
 
Mathematics news on Phys.org
I think these kind of problems have gone out of fashion these days.

Anyway, here is my solution to (1).

Problem 1

Step 1 - Reduction to Euler Sum

\[
\begin{aligned}
\int_0^1 \frac{\log(1+x)\log(x)}{1-x}dx &= \sum_{n=1}^\infty \frac{(-1)^{n+1}}{n}\int_0^1 \frac{x^n \log(x)}{1-x}dx \\
&= \sum_{n=1}^\infty \frac{(-1)^{n+1}}{n} \left( H_n^{(2)}-\frac{\pi^2}{6}\right)
\\ &= \sum_{n=1}^\infty \frac{(-1)^{n+1}}{n}H_n^{(2)}-\frac{\pi^2}{6}\log(2)
\end{aligned}
\]

where \(\displaystyle H_n^{(2)} = \sum_{k=1}^n \frac{1}{n^2}\)

Step 2 - Evaluation of Euler Sum

The evaluation of the Euler Sum is tricky.

Note that

\(\displaystyle \int_0^1 (-r)^{n-1} \ dr = \frac{(-1)^{n+1}}{n}\) and
\(\displaystyle \int_0^1 \int_0^1 (s t)^{k-1} \ ds \ dt = \frac{1}{k^2}\)

Plugging these into the sum, we obtain

\[
\begin{aligned}
\sum_{n=1}^\infty \frac{(-1)^{n+1}}{n}H_n^{(2)} &= \sum_{n=1}^\infty \frac{(-1)^{n+1}}{n}\sum_{k=1}^n \frac{1}{k^2} \\
&= \sum_{n=1}^\infty \sum_{k=1}^n \int_0^1 (-r)^{n-1} \ dr \int_0^1 \int_0^1 (s t)^{k-1} \ ds \ dt \\
&= \sum_{k=1}^\infty \int_0^1 \left( \sum_{n=k}^\infty (-r)^{n-1} \right)dr \int_0^1 \int_0^1 (s t)^{k-1} \ ds \ dt \\
&= \sum_{k=1}^\infty \int_0^1 \frac{(-r)^{k-1}}{1+r}dr\int_0^1 \int_0^1 (s t)^{k-1} \ ds \ dt \\
&= \int_0^1 \int_0^1 \int_0^1 \frac{1}{(1+r)(1+rst)}dr \ ds \ dt \\
&= \int_0^1 \int_0^1 \frac{\log(1+rs)}{(1+r)(rs)}ds \ dr \\
&=-\int_0^1 \frac{\text{Li}_2(-r)}{(1+r)r}dr \\
&=\int_0^1 \text{Li}_2(-r) \left(\frac{1}{1+r}-\frac{1}{r} \right)dr \\
&= -\log(2)\frac{\zeta(2)}{2}+\int_0^1 \frac{\log^2(1+r)}{r}dr +\frac{\zeta(3)}{4} \\ &= \zeta(3)-\frac{\zeta(2) \log(2)}{2}
\end{aligned}
\]

Step 3 - Now, the combination of these efforts results in

\[\begin{aligned}\int_0^1 \frac{\log(x) \log(1+x)}{1-x}dx &= \left( \zeta(3)-\frac{\zeta(2) \log(2)}{2}\right) - \frac{\pi^2}{6}\log(2) \\ &= \zeta(3)-\frac{\pi^2}{12}\log(2)-\frac{\pi^2}{6}\log(2) \\ &= \zeta(3)-\frac{\pi^2}{4}\log(2)\end{aligned}\]
 
According to our suggested guidelines for posting challenge problems posted here:

http://www.mathhelpboards.com/f28/guidelines-posting-answering-challenging-problem-puzzle-3875/

we ask that our members be given at least a week to respond before posting the solution(s). This gives people a good chance to respond if they have a solution. (Sun)
 
MarkFL said:
According to our suggested guidelines for posting challenge problems posted here:

http://www.mathhelpboards.com/f28/guidelines-posting-answering-challenging-problem-puzzle-3875/

we ask that our members be given at least a week to respond before posting the solution(s). This gives people a good chance to respond if they have a solution. (Sun)

I am sorry.:D I thought it would be nice to give some ideas on Euler Sums first.

Anyway, if it is against the rules I shall not post a solution so early.
 
Well, I wouldn't say the guidelines are as strict as rules, but when you post a challenging problem, it is best to give people a reasonable amount of time in which to respond. :D
 
Insights auto threads is broken atm, so I'm manually creating these for new Insight articles. In Dirac’s Principles of Quantum Mechanics published in 1930 he introduced a “convenient notation” he referred to as a “delta function” which he treated as a continuum analog to the discrete Kronecker delta. The Kronecker delta is simply the indexed components of the identity operator in matrix algebra Source: https://www.physicsforums.com/insights/what-exactly-is-diracs-delta-function/ by...
Fermat's Last Theorem has long been one of the most famous mathematical problems, and is now one of the most famous theorems. It simply states that the equation $$ a^n+b^n=c^n $$ has no solutions with positive integers if ##n>2.## It was named after Pierre de Fermat (1607-1665). The problem itself stems from the book Arithmetica by Diophantus of Alexandria. It gained popularity because Fermat noted in his copy "Cubum autem in duos cubos, aut quadratoquadratum in duos quadratoquadratos, et...
I'm interested to know whether the equation $$1 = 2 - \frac{1}{2 - \frac{1}{2 - \cdots}}$$ is true or not. It can be shown easily that if the continued fraction converges, it cannot converge to anything else than 1. It seems that if the continued fraction converges, the convergence is very slow. The apparent slowness of the convergence makes it difficult to estimate the presence of true convergence numerically. At the moment I don't know whether this converges or not.
Back
Top