26
$\begingroup$

I am trying to calculate

$$ I=\frac{1}{\pi}\int_0^\pi \theta^2 \ln^2\big(2\cos\frac{\theta}{2}\big)d \theta=\frac{11\pi^4}{180}=\frac{11\zeta(4)}{2}. $$ Note, we can expand the log in the integral to obtain three interals, one trivial, the other 2 are not so easy, any ideas? We will use $$ \left( \ln 2 +\ln \cos \frac{\theta}{2} \right)^2=\ln^2(2)+\ln^2\cos\frac{\theta}{2}+2\ln (2)\ln \cos\big(\frac{\theta}{2}\big) $$ and re-write I as $$ \pi I=\ln^2(2)\int_0^\pi \theta^2d\theta +\int_0^\pi\theta^2 \ln^2 \cos \frac{\theta}{2}d\theta+2\ln 2 \int_0^\pi\theta^2 \ln \cos{\frac{\theta}{2}}d\theta. $$ Simplfying this further by using $y=\theta/2$ we obtain $$ \pi I=\frac{\pi^3\ln^2(2)}{3}+16\ln(2)\int_0^{\pi/2} y^2 \ln \cos (y) dy+8\int_0^{\pi/2} y^2 \ln^2 \cos (y) dy $$ Any Idea how to approach these two integrals? I know that $$ \int_0^{\pi/2} \ln \cos y dy= \frac{-\pi\ln(2)}{2}\approx -1.08879 $$ but I am unsure how to use that here. I do not think partial integration will work, Also the Riemann Zeta function is given by $$ \zeta(4)=\sum_{n=1}^\infty \frac{1}{n^4}=\frac{\pi^4}{90}. $$ Thanks!

$\endgroup$
13
$\begingroup$

$\newcommand{\+}{^{\dagger}} \newcommand{\angles}[1]{\left\langle #1 \right\rangle} \newcommand{\braces}[1]{\left\lbrace #1 \right\rbrace} \newcommand{\bracks}[1]{\left\lbrack #1 \right\rbrack} \newcommand{\ceil}[1]{\,\left\lceil #1 \right\rceil\,} \newcommand{\dd}{{\rm d}} \newcommand{\down}{\downarrow} \newcommand{\ds}[1]{\displaystyle{#1}} \newcommand{\equalby}[1]{{#1 \atop {= \atop \vphantom{\huge A}}}} \newcommand{\expo}[1]{\,{\rm e}^{#1}\,} \newcommand{\fermi}{\,{\rm f}} \newcommand{\floor}[1]{\,\left\lfloor #1 \right\rfloor\,} \newcommand{\half}{{1 \over 2}} \newcommand{\ic}{{\rm i}} \newcommand{\iff}{\Longleftrightarrow} \newcommand{\imp}{\Longrightarrow} \newcommand{\isdiv}{\,\left.\right\vert\,} \newcommand{\ket}[1]{\left\vert #1\right\rangle} \newcommand{\ol}[1]{\overline{#1}} \newcommand{\pars}[1]{\left( #1 \right)} \newcommand{\partiald}[3][]{\frac{\partial^{#1} #2}{\partial #3^{#1}}} \newcommand{\pp}{{\cal P}} \newcommand{\root}[2][]{\,\sqrt[#1]{\vphantom{\large A}\,#2\,}\,} \newcommand{\sech}{\,{\rm sech}} \newcommand{\sgn}{\,{\rm sgn}} \newcommand{\totald}[3][]{\frac{{\rm d}^{#1} #2}{{\rm d} #3^{#1}}} \newcommand{\ul}[1]{\underline{#1}} \newcommand{\verts}[1]{\left\vert\, #1 \,\right\vert}$ $\ds{I \equiv {1 \over \pi}\int_{0}^{\pi}\theta^{2}\ln^{2}\pars{2\cos\pars{\theta \over 2}} \,\dd\theta = {11\pi^{4} \over 180} = {11 \over 2}\,\zeta\pars{4}}$

\begin{align} I&={1 \over 2\pi}\int_{-\pi}^{\pi}\theta^{2} \ln^{2}\pars{2\root{1 + \cos\pars{\theta} \over 2}}\,\dd\theta \\[3mm]&={1 \over 2\pi} \int_{\verts{z} = 1 \atop {\vphantom{\Huge A}\verts{{\rm Arg}\pars{z}}\ <\ \pi}} \bracks{-\ln^{2}\pars{z}}\ln^{2}\pars{\root{2}\root{1 + {z^{2} + 1 \over 2z}}}\, {\dd z \over \ic z} \\[3mm]&={\ic \over 2\pi} \int_{\verts{z} = 1 \atop {\vphantom{\Huge A}\verts{{\rm Arg}\pars{z}}\ <\ \pi}} \ln^{2}\pars{z}\ln^{2}\pars{z + 1 \over z^{1/2}}\,{\dd z \over z} \\[3mm]&={\ic \over 2\pi}\lim_{\mu \to -1 \atop \nu \to 0} \partiald[2]{}{\mu}\partiald[2]{}{\nu} \int_{\verts{z} = 1 \atop {\vphantom{\Huge A}\verts{{\rm Arg}\pars{z}}\ <\ \pi}} z^{\mu}\pars{z + 1 \over z^{1/2}}^{\nu}\,\dd z \\[3mm]&={\ic \over 2\pi}\lim_{\mu \to -1 \atop \nu \to 0} \partiald[2]{}{\mu}\partiald[2]{}{\nu} \int_{\verts{z} = 1 \atop {\vphantom{\Huge A}\verts{{\rm Arg}\pars{z}}\ <\ \pi}}z^{\mu - \nu/2}\pars{z + 1}^{\nu}\,\dd z \qquad\qquad\qquad\qquad\qquad\qquad\qquad\quad\pars{1} \end{align}

The integration in $\pars{1}$ is given by: \begin{align} &\int_{\verts{z} = 1} z^{\mu - \nu/2}\pars{z + 1}^{\nu}\,\dd z \\[3mm]&=-\int_{-1}^{0}\pars{-x}^{\mu - \nu/2}\expo{\ic\pi\pars{\mu - \nu/2}} \pars{x + 1}^{\nu}\,\dd x -\int_{0}^{-1}\pars{-x}^{\mu - \nu/2}\expo{-\ic\pi\pars{\mu - \nu/2}} \pars{x + 1}^{\nu}\,\dd x \\[3mm]&=-\expo{\ic\pi\pars{\mu - \nu/2}}\int_{0}^{1}x^{\mu - \nu/2} \pars{-x + 1}^{\nu}\,\dd x +\expo{-\ic\pi\pars{\mu - \nu/2}}\int_{0}^{1}x^{\mu - \nu/2} \pars{-x + 1}^{\nu}\,\dd x \\[3mm]&=2\ic\sin\pars{\pi\bracks{{\nu \over 2} - \mu}} {\rm B}\pars{\mu - {\nu \over 2} + 1,\nu + 1}\tag{2} \end{align} where $\ds{{\rm B}\pars{x,y} = \int_{0}^{1}t^{x - 1}\pars{1 - t}^{y - 1}\,\dd t}$, $\ds{\pars{~\mbox{with}\ \Re\pars{x} > 0,\ \Re\pars{y} > 0~}}$ is the Beta Function.

With $\pars{1}$ and $\pars{2}$, $\ds{I}$ is reduced to: $$ I=-\,{1 \over \pi}\lim_{\mu \to -1 \atop \nu \to 0} \partiald[2]{}{\mu}\partiald[2]{}{\nu}\bracks{% \sin\pars{\pi\bracks{{\nu \over 2} - \mu}} {\rm B}\pars{\mu - {\nu \over 2} + 1,\nu + 1}} $$

Since $\ds{{\rm B}\pars{x,y}= {\Gamma\pars{x}\Gamma\pars{y} \over \Gamma\pars{x + y}}}$ ( $\ds{\Gamma\pars{z}}$ is the GammaFunction ): \begin{align} I&=-\,{1 \over \pi}\ \lim_{\mu \to -1 \atop \nu \to 0} \partiald[2]{}{\mu}\partiald[2]{}{\nu}\bracks{% \sin\pars{\pi\bracks{{\nu \over 2} - \mu}} {\Gamma\pars{\mu - \nu/2 + 1}\Gamma\pars{\nu + 1} \over \Gamma\pars{\mu + \nu/2 + 2}}} \\[3mm]&=-\ \overbrace{\lim_{\mu \to -1 \atop \nu \to 0} \partiald[2]{}{\mu}\partiald[2]{}{\nu}\bracks{% {\Gamma\pars{\nu + 1} \over \Gamma\pars{\nu/2 - \mu}\Gamma\pars{\mu + \nu/2 + 2}}}} ^{\ds{=\ -\,{11\pi^{4} \over 180}}} \end{align} where we used the identity $\ds{\Gamma\pars{z}\Gamma\pars{1 - z} = {\pi \over \sin\pars{\pi z}}}$

Then, $$ I \equiv \color{#00f}{\large% {1 \over \pi}\int_{0}^{\pi}\theta^{2}\ln^{2}\pars{2\cos\pars{\theta \over 2}} \,\dd\theta} = \color{#00f}{\large{11\pi^{4} \over 180} = {11 \over 2}\,\zeta\pars{4}} $$

$\endgroup$
  • $\begingroup$ This is an excellent solution. Thanks a lot Felix Marin. I had thought I had commented earlier on this. Sorry for the delay, this answer is GREAT. Nice Integral $\endgroup$ – Jeff Faraci Mar 20 '14 at 0:00
  • 1
    $\begingroup$ @Jeff Thanks a lot. You're welcome. $\endgroup$ – Felix Marin Mar 20 '14 at 0:33
  • $\begingroup$ you may appreciate this: math.uwo.ca/~dborwein/cv/zeta4.pdf $\endgroup$ – Jeff Faraci Apr 27 '14 at 19:05
  • 1
    $\begingroup$ @Integrals That link is very helpful. Manipulation of Harmonic Numbers are always somehow tricky and that paper has a lot about them. Thanks. $\endgroup$ – Felix Marin Apr 27 '14 at 21:49
3
$\begingroup$

One way to go about this is to use the identity:

$$\int_{0}^{\frac{\pi}{2}}\cos^{p-1}(x)\cos(ax)dx=\frac{\pi}{2^{p}}\cdot \frac{\Gamma(p)}{\Gamma\left(\frac{a+p+1}{2}\right)\Gamma\left(\frac{p-a+1}{2}\right)}....(1)$$

Then, diff this twice w.r.t 'a', and let a=0.

Then, diff twice w.r.t p and let p=1.

The diffing on the right side may be a little tedious, but get tech to do it.

diffing once w.r.t p will give you $$\int_{0}^{\frac{\pi}{2}}x^{2}\ln(\cos(x))dx=\frac{-\pi^{3}}{24}\ln(2)-\frac{\pi}{3}\zeta(3)$$.

Then, diff again to get the integral in question.

a fun way to go about evaluating this integral is to use contours.

consider $$f(z)=zlog^{3}(1+e^{2iz})$$ over a rectangular contour with vertices $$-\frac{\pi}{2}, \;\ \frac{\pi}{2}, \;\ \frac{\pi}{2}+Ri, \;\ \frac{-\pi}{2}+Ri$$, with quarter-circle indents around $\pm \frac{\pi}{2}$.

A good while back, Nick Strehle wrote up a nice post on this method of evaluating log-trig integrals via residues. It is on the site somewhere if you nose around.

$\endgroup$

Your Answer

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

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