Limit of sum of sequences at infinity
Limit of sum of sequences at infinity
Given two sequences $$a_n=int_0^1 (1-x^2)^n dx$$ and $$b_n=int_0^1 (1-x^3)^n dx$$ ,($nin N$) then find the value of $$L=lim_nto infty (10 sqrt [n]a_n +5sqrt [n]b_n)$$
My try:
$$a_n=int_0^1 (1-x^2)^n dx= frac 12 Bleft( n+1, 1/2right)=frac 12cdot frac Gamma(n+1)Gamma(1/2)Gammaleft(n +frac 32right)$$
Similarly $$b_n=frac 13cdot frac Gamma(n+1)Gamma(1/3)Gammaleft(n +frac 43right)$$
Now let's consider $$J=lim_nto infty (b_n)^1/n =lim_nto infty left(frac 13right)^1/ncdot left(frac Gamma(n+1)Gamma(1/3)Gammaleft(n +frac 43right)right) ^1/n$$ $$=lim_ntoinfty left(frac 13right)^1/ncdot left(lim_nto infty left(frac Gamma(n+1)Gamma(1/3)Gammaleft(n +frac 43right)right) ^1/nright)$$ $$=1cdot lim_nto infty left(frac Gamma(n+1)Gamma(1/3)Gammaleft(n +frac 43right)right) ^1/n$$ $$=exp left( lim_ntoinfty frac ln(Gamma(n+1)) +ln(Gamma(1/3))-ln(Gamma(n +4/3))nright)$$
Now by L'Hospital we have $$exp left( lim_nto infty frac ln(Gamma(n+1)) +ln(Gamma(1/3))-ln(Gamma(n +4/3))nright) =expleft( lim_nto infty left[psi(n+1) -psi(n+ 4/3)right] right)$$
But for large $x$, $psi(x)sim ln (x)$ Hence $$exp left( lim_nto infty left[psi(n+1) -psi(n+ 4/3)right]right)=exp left( lim_nto infty left[ln(n+1) -ln(n+ 4/3)right]right)=1$$
Similarly $$lim_nto infty (a_n)^1/n= 1$$
Hence giving me the answer as $15$.
I want to know whether I am correct or not. And if I am correct I want to know if there is any other simple and elementary method to solve the question.
Edit:
By some other method I dont mean that just converting the Gamma function to factorial representation and then using Stirling's approximation
2 Answers
2
A much simpler approach. Let $k$ be a positive integer. Then for $xin [0,1]$,
$$1-xleq 1-x^kleq 1$$
and therefore
$$frac1n+1=int_0^1 (1-x)^n dxleq a_n(k):=int_0^1 (1-x^k)^n dxleq 1.$$
Hence, as $ntoinfty$,
$$1leftarrowfrac1sqrt[n]n+1lesqrt[n]a_n(k)leq sqrt[n]1=1$$
and by the Squeeze Theorem it follows that $sqrt[n]a_n(k)rightarrow 1$. In your case, we need the limit for $k=2$ and $k=3$.
There is a very simple method of evaluating $L$. Fort any bounded measurable function $f$ on $[0,1]$ we have $lim (int_0^1 |f|^p)^1/p to |f|_infty$ as $ p to infty$. Hence $a_n^1/n to max (1-x^2): 0leq x leq 1=1$. Similarly $b_n^1/n to 1$. So $L=10+5=15$. Actually, th result I have quoted is easy to prove in this case. Note that $int_0^1 (1-x^2)^n leq 1$. Also $int_0^1 (1-x^2)^n geq int_0^t (1-x^2)^n geq t(1-t^2)^n$ for each $t$. Take n-th root and conclude that $lim inf a_n^1/n geq 1-t^2$ for anty $t>0$. Let $t to 0$.
$begingroup$
This is a 'popular' exercise from Rudin's book and you can get it easily from MSE.
$endgroup$
– Kavi Rama Murthy
Sep 7 '18 at 6:28
$begingroup$
Some MSE links please?????
$endgroup$
– Digamma
Sep 7 '18 at 6:29
$begingroup$
@Manthanein I have included a proof now.
$endgroup$
– Kavi Rama Murthy
Sep 7 '18 at 6:36
$begingroup$
Does this result have it's origin in the space theory or measure theory or something related. If yes then I am really sorry sir because I don't have quite an idea about those chapters. I am just in the 12th standard as of now. But I will try to understand the proofs so that I could use this result in near future.
$endgroup$
– Digamma
Sep 7 '18 at 6:45
Thanks for contributing an answer to Mathematics Stack Exchange!
But avoid …
Use MathJax to format equations. MathJax reference.
To learn more, see our tips on writing great answers.
Required, but never shown
Required, but never shown
By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.
$begingroup$
Where does this result come from? Any references might help me a bit
$endgroup$
– Digamma
Sep 7 '18 at 6:27