3

So I have the two following series: $$\sum_{n=1}^\infty \frac{\prod_{k=1}^n(2k)^2}{(2n+2)!}$$ $$\sum_{n=0}^\infty \frac{\prod_{k=0}^n(2k+1)^2}{(2n+3)!}$$ I figured out the $n$th partial products: $$\prod_{k=1}^n(2k)^2=4^n(n!)^2$$ $$\prod_{k=0}^n (2k+1)^2=\frac{((2n+1)!)^2}{4^n(n!)^2}$$ So putting these back into my series they become the following: $$\sum_{n=1}^\infty \frac{\prod_{k=1}^n(2k)^2}{(2n+2)!}=\sum_{n=1}^\infty\frac{4^n(n!)^2}{(2n+2)!}$$ Now this diverges as expected by the limit test test. However when I look at my other series: $$\sum_{n=0}^\infty \frac{\prod_{k=0}^n(2k+1)^2}{(2n+3)!}=\sum_{n=0}^\infty\frac{( (2n+1)!)^2}{4^n(n!)^2(2n+3)!}$$ By the limit test maybe diverges or maybe doesn't, and the ratio test is inconclusive. Since I wasn't sure what to use for the a comparison test I threw this into wolfram alpha and it told me it converges which is baffling to me since both series are very similar if we write them out: $$\sum_{n=1}^\infty \frac{\prod_{k=1}^n(2k)^2}{(2n+2)!}=\frac{2^2}{4!}+\frac{2^24^2}{6!}+\frac{2^24^26^2}{8!}\cdot\cdot\cdot\cdot$$ $$\sum_{n=0}^\infty \frac{\prod_{k=0}^n(2k+1)^2}{(2n+3)!}=\frac{1^2}{3!}+\frac{1^23^2}{5!}+\frac{1^23^25^2}{7!}+\cdot\cdot\cdot$$ They both have the nth parial product of the even/odd integers squared in the numerator, and are over a factorial that is two greater than $n$, so I'm not sure why one is diverging and the other is converging. Is wolframalpha wrong, as it can be at times? Or is there someething here that I am missing?

Chris
  • 2,921
  • The first series also converges. How did you conclude it diverges? – Erick Wong Oct 21 '20 at 19:30
  • I agree with Erick, I'm getting an approximate numeric value of $1.225$. – K.defaoite Oct 21 '20 at 19:33
  • The terms shrink proportionately to $1/n^{3/2}$, so I suspect you’ll need something like Stirling’s approximation to determine convergence: limit or ratio/root test wouldn’t be sensitive enough. – Erick Wong Oct 21 '20 at 20:36
  • @ErickWong I mean I got that $\lim_{n\rightarrow\infty}\frac{4^n(n!)^2}{(2n+2)!}$ when to $\infty$ so thats how I got that the first series diverges. I was unaware that the limit test wouldn't be sensitive enough for that. Could you elaborate a little bit on that? – Chris Oct 21 '20 at 22:24
  • @ChristopherQuinnLaFondJr. I assume you mean “went to” instead of “when to”. But I think the limit is actually $0$ and you would probably agree the limit test is inconclusive if that is indeed true. – Erick Wong Oct 22 '20 at 03:00
  • @ErickWong I did mean went to. I may have messed up the limit but yes I would agree that it is inconclusive if that is true. – Chris Oct 22 '20 at 10:53

2 Answers2

4

Elaborating after @Erick Wong's comments.

You properly found that

$$a_n=\frac{4^n(n!)^2}{(2n+2)!}$$ Take logarithms $$\log(a_n)=n \log(4)+2\log(n!)-\log((2n+2)!)$$ Use Stirling approximation twice and continue with Taylor series to find $$\log(a_n)=\left(\frac{3}{2} \log \left(\frac{1}{n}\right)+\log \left(\frac{\sqrt{\pi }}{4}\right)\right)-\frac{11}{8 n}+O\left(\frac{1}{n^2}\right)$$ that is to say $$a_n \sim \frac{\sqrt \pi}{4 n^{\frac 32}}\exp\left(-\frac{11}{8 n}\right) <\frac{\sqrt \pi}{4 n^{\frac 32}}$$ $$\sum_{n=1}^\infty \frac{\sqrt \pi}{4 n^{\frac 32}}=\frac{\sqrt{\pi }}{4} \zeta \left(\frac{3}{2}\right)\approx 1.15758$$

Sooner or later, you will learn that $$\sum_{n=1}^\infty \frac{4^n(n!)^2}{(2n+2)!}=\frac{\pi ^2-4}{8}\approx 0.73370$$

Doing the same with $$b_n=\frac{(2n+1)!^2}{4^n(n!)^2(2n+3)!}$$ $$\log(b_n)=2\log((2n+1)!)-n \log(4)-2\log(n!)-\log((2n+3)!)$$ $$\log(b_n)=\left(\frac{3}{2} \log \left(\frac{1}{n}\right)+\log \left(\frac{1}{2 \sqrt{\pi }}\right)\right)-\frac{17}{8 n}+O\left(\frac{1}{n^2}\right)$$ that is to say $$b_n \sim \frac{1}{2 \sqrt \pi n^{\frac 32}}\exp\left(-\frac{17}{8 n}\right) < \frac{1}{2 \sqrt \pi n^{\frac 32}}$$ $$\sum_{n=1}^\infty \frac{1}{2 \sqrt \pi n^{\frac 32}}=\frac{1}{2 \sqrt \pi }\zeta \left(\frac{3}{2}\right)\approx 0.73694$$ Sooner or later, you will learn that $$\sum_{n=0}^\infty \frac{(2n+1)!^2}{4^n(n!)^2(2n+3)!}=\frac{ \pi -2}{2} \approx 0.57080$$

Edit

Notice that $$\sum_{n=1}^\infty a_n\,x^n=\frac{\sin ^{-1}(x)^2-x^2}{2 x^2}$$ $$\sum_{n=0}^\infty b_n\,x^n=\frac{\sin ^{-1}(x)-x}{x^3}$$

  • What do you mean that sooner or later I will learn that those two sums are equivalent to those exact values? – Chris Oct 22 '20 at 10:56
  • 1
    I will give them the benefit of the doubt and say an analysis of your series may not fit into the answer(?) The first series is related to the Taylor expansion of $\sin^{-1}(x^2)$. The second is related to the series $$\pi = 2 \sum_{k=0}^\infty \frac {(2 k - 1)!!}{(2 k + 1) (2 k)!!}$$ listed in Wolfram Research without context. I'm not sure which expansion this is from. – player3236 Oct 22 '20 at 11:37
  • @player3236. Have a look at my edit for the infinite series. Cheers :-) – Claude Leibovici Oct 23 '20 at 03:24
  • (+1) However, the second sum starts at $n=0$, so the answer should be $\frac16$ greater. – robjohn Oct 23 '20 at 12:39
3

Convergence

Using the asymptotic approximation given in inequality $(10)$ of this answer, we get $$ \binom{2n}{n}\sim\frac{4^n}{\sqrt{\pi n}}\tag1 $$ Therefore, $$ \begin{align} \frac{\prod\limits_{k=1}^n(2k)^2}{(2n+2)!} &=\frac{4^nn!^2}{(2n)!(2n+1)(2n+2)}\\ &=\frac{\color{#090}{4^n}}{\color{#090}{\binom{2n}{n}}\color{#C00}{(2n+1)(2n+2)}}\\ &\sim\frac{\color{#090}{\sqrt{\pi n}}}{\color{#C00}{4n^2}}\\ &=\frac{\sqrt\pi}{4}\frac1{n^{3/2}}\tag2 \end{align} $$ and $$ \begin{align} \frac{\prod\limits_{k=0}^n(2k+1)^2}{(2n+3)!} &=\frac{(2n+1)!^2}{4^nn!^2(2n+3)!}\\ &=\frac{\color{#090}{\binom{2n}{n}}\color{#C00}{(2n+1)}}{\color{#090}{4^n}\color{#C00}{(2n+2)(2n+3)}}\\ &\sim\frac1{\color{#090}{\sqrt{\pi n}}\,\color{#C00}{2n}}\\ &=\frac1{2\sqrt\pi}\frac1{n^{3/2}}\tag3 \end{align} $$ The sums of both $(2)$ and $(3)$ converge by comparison to a $p$-series with $p=3/2$.


Evaluation

In this answer, it is shown that $$ \begin{align} \arcsin^2(x) &=\sum_{k=1}^\infty\frac{4^kx^{2k}}{2k^2\binom{2k}{k}}\\ &=\sum_{k=1}^\infty\frac{4^k}{\binom{2k}{k}}\frac{x^{2k}}{2k^2}\\ &=\sum_{k=0}^\infty\frac{4^k}{\binom{2k}{k}}\frac{2x^{2k+2}}{(2k+1)(2k+2)}\tag4\\ \end{align} $$ and in this answer, it is shown that $$ \begin{align} \arcsin(x) &=\sum_{k=0}^\infty\frac2{2k+1}\binom{2k}{k}\left(\frac{x}{2}\right)^{2k+1}\\ &=\sum_{k=0}^\infty\frac{\binom{2k}{k}}{4^k}\frac{x^{2k+1}}{2k+1}\\ &=x+\sum_{k=0}^\infty\frac{\binom{2k}{k}}{4^k}\frac{(2k+1)x^{2k+3}}{(2k+2)(2k+3)}\tag5 \end{align} $$ Applying $(4)$, we get $$ \begin{align} \sum_{n=1}^\infty\frac{\prod\limits_{k=1}^n(2k)^2}{(2n+2)!} &=\sum_{n=1}^\infty\frac{4^n}{\binom{2n}{n}(2n+1)(2n+2)}\\ &=\frac12\arcsin(1)^2-\frac12\\ &=\frac{\pi^2}8-\frac12\tag6 \end{align} $$ Applying $(5)$, we get $$ \begin{align} \sum_{n=0}^\infty\frac{\prod\limits_{k=0}^n(2k+1)^2}{(2n+3)!} &=\sum_{n=0}^\infty\frac{\binom{2n}{n}(2n+1)}{4^n(2n+2)(2n+3)}\\ &=\arcsin(1)-1\\[6pt] &=\frac\pi2-1\tag7 \end{align} $$

robjohn
  • 345,667