3

$$y=\lim_{n\to \infty}\left(\sin π/2n \times \sin 2π/2n \times... \sin(n-1)π/2n\right)$$ Find y.  I took log on both the sides and I got  $$\ln y= \int_0^1 {\ln \sin πx/2\, dx}$$ but I'm stuck here.

2 Answers2

1

If we wish to evaluate the limit of interest by use of Riemann sums, then we can proceed as follows. Let $y_n$ be the product defined by

$$y_n=\prod_{k=1}^{n-1}\sin\left(\frac{k\pi}{2n}\right) \tag 1$$

Taking the logarithm of both sides of $(1)$ reveals

$$\log y_n=\sum_{k=1}^{n}\log\left(\sin\left(\frac{k\pi}{2n}\right)\right) \tag 2$$

where we used the fact that $\sin \pi/2=1$, thereby permitting our extending the limit on the sum to $n$. Note that we can write the Riemann sum of the integral

$$\begin{align} \int_0^{\pi/2}\log\left(\sin x\right)\,dx&=\lim_{n\to \infty}\,\frac{\pi/2}{n}\,\sum_{k=1}^{n}\log\left(\sin\left(\frac{k\pi}{2n}\right)\right)\\\\ &=-\frac12 \pi \log 2 \tag 3 \end{align}$$

Comparing the $(2)$ and $(3)$ we see that

$$\lim_{n\to \infty}\log y_n =-\infty$$

which implies therefore that

$$\bbox[5px,border:2px solid #C0A000]{\lim_{n\to \infty}y_n=0}$$

Mark Viola
  • 179,405
0

Obviously to compute the closed form of the product is a chance, but there are at least two good reasons for which that limit is zero that require very few computations:

  1. $f(\theta)=\log\sin\theta$ is an improperly Riemann-integrable function over $\left(0,\frac{\pi}{2}\right)$;
  2. for any $\theta\in\left(0,\frac{\pi}{2}\right)$, we have $0<\sin\theta<\theta.$
Jack D'Aurizio
  • 353,855