Evaluating $\sum_{k\geq 1}\frac{\cos \left(\frac{\pi k}{2} \right)}{k^2}$

  • MHB
  • Thread starter alyafey22
  • Start date
In summary, the conversation discusses the evaluation of the series $\sum_{k\geq 1}\frac{\cos \left(\frac{\pi k}{2} \right)}{k^2}$. The solution involves using the second order Clausen functions and converting the integral into complex exponential form. This leads to a closed form expression for the series, which can be used to evaluate other similar series as well.
  • #1
alyafey22
Gold Member
MHB
1,561
1
Evaluate the following
\(\displaystyle \sum_{k\geq 1}\frac{\cos \left(\frac{\pi k}{2} \right)}{k^2}\)​
 
Mathematics news on Phys.org
  • #2
[sp]It looks like part of the Fourier series for \(\displaystyle x(\pi-x)\) at \(\displaystyle x=\pi/4\)

So I'm getting \(\displaystyle \pi/4(\pi-\pi/4)=\pi^2/6-S\).

Leading to S=\(\displaystyle -\pi^2/48\)...?

Bit of a cheat really. [/sp]
 
  • #3
ZaidAlyafey said:
Evaluate the following
\(\displaystyle \sum_{k\geq 1}\frac{\cos \left(\frac{\pi k}{2} \right)}{k^2}\)​

It is relatively easy...

$\displaystyle \sum_{k = 1}^{\infty} \frac{\cos (\frac{\pi k}{2})}{k^{2}} = - \frac {1}{4}\ (1 - \frac{1}{4} + \frac{1}{9} - \frac{1}{16} + ...) = - \frac{\pi^{2}}{48}$

Kind regards

$\chi$ $\sigma$
 
  • #4
ZaidAlyafey said:
Evaluate the following
\(\displaystyle \sum_{k\geq 1}\frac{\cos \left(\frac{\pi k}{2} \right)}{k^2}\)​
Hello Z! (Sun) Nice little problem there...

I hope no one minds the bump, but there's a general closed form for this series... Define the second order Clausen functions as follows:\(\displaystyle \text{Cl}_2(\theta)=-\int_0^{\theta}\log\Biggr|2\sin\frac{x}{2}\Biggr| \,dx=\sum_{k=1}^{\infty}\frac{\sin k\theta}{k^2}\)

\(\displaystyle \text{Sl}_2(\theta)=\sum_{k=1}^{\infty}\frac{\cos k\theta}{k^2}\)[N.B. that integral definition holds without use of the absolute value sign within the range \(\displaystyle 0 < \theta < 2\pi\,\)]

Now we convert the CL-integral into complex exponential form

\(\displaystyle \text{Cl}_2(\theta)=-\int_0^{\theta}\log\left(2\sin\frac{x}{2}\right)\,dx=-\int_0^{\theta}\log\left[2\left(\frac{e^{ix/2}-e^{-ix/2}}{2i}\right)\right]\,dx=\)

\(\displaystyle -\int_0^{\theta}\log\left(\frac{1-e^{-ix}}{ie^{-ix/2}}\right)\,dx=\)

\(\displaystyle \int_0^{\theta}\left(\log i-\frac{ix}{2}\right)\,dx-\int_0^{\theta}\log(1-e^{-x})\,dx=\)

\(\displaystyle \frac{\pi i\theta}{2}-\frac{i\theta^2}{4}-\int_0^{\theta}\log(1-e^{-ix})\,dx\)Now we expand the complex logarithm within the integrand as a power series:

\(\displaystyle \log(1-z)=-\sum_{k=0}^{\infty}\frac{z^k}{k}\)

\(\displaystyle \Rightarrow\)

\(\displaystyle \int_0^{\theta}\log(1-e^{-ix})\,dx=-\sum_{k=0}^{\infty}\frac{1}{k}\int_0^{\theta}e^{-ikx}\,dx=\)

\(\displaystyle -\sum_{k=0}^{\infty}\frac{1}{k}\int_0^{\theta}(\cos x-i \sin x)^{k}\,dx=-\sum_{k=0}^{\infty}\frac{1}{k}\int_0^{ \theta}(\cos kx-i \sin kx)\,dx=\)

\(\displaystyle -\sum_{k=0}^{\infty}\frac{1}{k}\Biggr[\frac{1}{k}\left(\sin kx+i \cos kx\right)\Biggr]_0^{\theta}=\)

\(\displaystyle -\sum_{k=0}^{\infty}\frac{1}{k}\Biggr[\frac{1}{k}\left(\sin k\theta+i \cos k\theta-i\right)\Biggr]=\)

\(\displaystyle -\sum_{k=0}^{\infty}\frac{\sin k\theta}{k^2}-i\sum_{k=0}^{\infty}\frac{\cos k\theta}{k^2}+i\sum_{k=0}^{\infty}\frac{1}{k^2}=\)

\(\displaystyle -\text{Cl}_2(\theta)-i\text{Sl}_2(\theta)+i\zeta(2)=-\text{Cl}_2(\theta)-i\text{Sl}_2(\theta)+\frac{i\pi^2}{6}\)Combining this with our earlier results, and equating the imaginary parts to zero, we deduce that

\(\displaystyle \text{Sl}_2(\theta)=\frac{\pi^2}{6}-\frac{\pi\theta}{2}+\frac{\theta^2}{4}\)
Or, equivalently...\(\displaystyle \sum_{k\ge 1}\frac{\cos k\theta}{k^2}= \frac{\pi^2}{6}-\frac{\pi\theta}{2}+\frac{\theta^2}{4}\, .\, \Box\)-------------------------------------------

For \(\displaystyle m\in \mathbb{N}^+\,\), all series of the form\(\displaystyle \text{Sl}_{2m}(\theta)=\sum_{k=0}^{\infty}\frac{ \cos k\theta}{k^{2m}}\)

\(\displaystyle \text{Sl}_{2m+1}(\theta)=\sum_{k=0}^{\infty}\frac{\sin k\theta}{k^{2m}}\)Are expressible in closed form, as a polynomial in \(\displaystyle \theta\,\) and \(\displaystyle \pi\,\)
 
  • #5
Nice , generalization !
 
  • #6
Thanks Z! :DIt can make tricky-looking results quite easy. Such as\(\displaystyle \sum_{k=1}^{\infty}\frac{\cos (\pi k/10)}{k^2}=\frac{\pi^2}{6}-\frac{\pi}{4}\sqrt{\frac{5+\sqrt{5}}{2}}+\frac{5+ \sqrt{5}}{16}\)
 
  • #7
DreamWeaver said:
Thanks Z! :DIt can make tricky-looking results quite easy. Such as\(\displaystyle \sum_{k=1}^{\infty}\frac{\cos (\pi k/10)}{k^2}=\frac{\pi^2}{6}-\frac{\pi}{4}\sqrt{\frac{5+\sqrt{5}}{2}}+\frac{5+ \sqrt{5}}{16}\)
Oops! :eek::eek::eek:

Sorry folks... Wasn't really concentrating yesterday... I put in value of \(\displaystyle \cos \theta\) rather than the argument, \(\displaystyle \theta\)...[hangs head in shame]
 

FAQ: Evaluating $\sum_{k\geq 1}\frac{\cos \left(\frac{\pi k}{2} \right)}{k^2}$

What is the value of the given series?

The value of the series is approximately 0.176. This can be calculated by plugging in the values of k from 1 to infinity and adding them up.

How is the value of the series calculated?

The value of the series is calculated using a mathematical technique called the Euler-Maclaurin summation formula. This formula allows us to approximate the value of a series by using a combination of numerical integration and summation.

What is the significance of the cosine function in the series?

The cosine function in the series is what makes the series converge. Without the cosine function, the series would diverge and not have a finite value. The cosine function helps to balance out the terms in the series and ensure convergence.

How does the value of the series change as k increases?

As k increases, the value of the series becomes closer to its true value. However, since the series is an infinite sum, it will never reach its exact value. As k gets larger, the value of the series will approach 0.176 but will never equal it.

Can the value of the series be approximated using other methods?

Yes, the value of the series can also be approximated using numerical integration techniques such as Simpson's rule or the trapezoidal rule. These methods involve dividing the area under the curve into smaller sections and summing them up to approximate the value of the series.

Similar threads

Back
Top