Ask your own question, for FREE!
Mathematics 13 Online
OpenStudy (legomyego180):

why

OpenStudy (legomyego180):

is \[\lim_{k \rightarrow \infty}\frac{ k^k }{ (k+1)^k} = \frac{ 1 }{ e } \]

OpenStudy (solomonzelman):

Yes

OpenStudy (solomonzelman):

You can prove this. Should I start you off?

OpenStudy (legomyego180):

please

OpenStudy (solomonzelman):

\(\displaystyle \lim_{k\to \infty}\frac{k^k}{(k+a)^k}=\lim_{k\to \infty}\left[\frac{k}{k+a}\right]^k=\lim_{t\to \infty}\left[\frac{t-a}{t}\right]^t\) You can achieve the third step in that line by setting k+1=t (that dictates k=t-1). (If you have a question about that, ask...) Then use \(\displaystyle y=e^{\ln y}\). \(\displaystyle \lim_{t\to \infty}\left[\frac{t-a}{t}\right]^t=\exp\left\{\ln~\lim_{t\to \infty}\left[\frac{t-a}{t}\right]^t\right\}\) \(\displaystyle =\lim_{t\to \infty}e^{\ln~\left[\frac{t-a}{t}\right]^t}=\lim_{t\to \infty}e^{t\ln~\left[\frac{t-a}{t}\right]}=\lim_{t\to \infty}e^{\ln~\left[\frac{t-a}{t}\right]~/~(1/t)}\) In the last step I re-wrote t as 1/t in the denominator. Now, take the L'Hospital's Rule.

OpenStudy (solomonzelman):

By the way, `exp{y}` is the same as `e^y`

OpenStudy (solomonzelman):

You can use 1, instead of "a", if abstractness troubles you. (B/c in your case you have 1, not a.) I was just going in general. \(\\[0.9em]\) [In my general proof startup, you should get e^(-a).]

OpenStudy (solomonzelman):

One more thing. It would reduce paint, if you re-write `(t-a)/t` as `1 - (a/t)`.

OpenStudy (legomyego180):

still here, working through it...

OpenStudy (solomonzelman):

Yes, just to show where we are, you are to take the L'H'S of \(\displaystyle \exp\left\{\lim_{t\to \infty}~\frac{\ln\left[1-\frac{a}{t}\right]}{1/t}\right\}\)

OpenStudy (solomonzelman):

you can ask me to repost it but without confusions. I will do so, if you need.

OpenStudy (legomyego180):

so, \(\color{blue}{\text{Originally Posted by}}\) @SolomonZelman \(\displaystyle \lim_{k\to \infty}\frac{k^k}{(k+a)^k}=\lim_{k\to \infty}\left[\frac{k}{k+a}\right]^k=\lim_{t\to \infty}\left[\frac{t-a}{t}\right]^t\) You can achieve the third step in that line by setting k+1=t (that dictates k=t-1). (If you have a question about that, ask...) Then use \(\displaystyle y=e^{\ln y}\). \(\displaystyle \lim_{t\to \infty}\left[\frac{t-a}{t}\right]^t=\exp\left\{\ln~\lim_{t\to \infty}\left[\frac{t-a}{t}\right]^t\right\}\) \(\displaystyle =\lim_{t\to \infty}e^{\ln~\left[\frac{t-a}{t}\right]^t}=\lim_{t\to \infty}e^{t\ln~\left[\frac{t-a}{t}\right]}=\color{red}\lim_{t\to \infty}e^{\ln~\left[\frac{t-a}{t}\right]~/~(1/t)}\) In the last step I re-wrote t as 1/t in the denominator. Now, take the L'Hospital's Rule. \(\color{blue}{\text{End of Quote}}\) So the limit in red is where Im having some trouble

OpenStudy (legomyego180):

Where did the 1/t come from?

OpenStudy (solomonzelman):

Oh. t = 1/ (1/t)

OpenStudy (solomonzelman):

Multiply by t, it same as dividing by 1/t, isn't it?

OpenStudy (solomonzelman):

or, do you mean why I did that?

OpenStudy (legomyego180):

so you could apply L'H?

OpenStudy (solomonzelman):

Yes, because as t->∞, the following is true: 1) ln[ t-a / a ] = ln[1] = 0 2) 1/t = 1/∞ .... so to speak .... and that is =0. So we have a classic 0/0 case

OpenStudy (solomonzelman):

oh, (1) should be t-a / t

OpenStudy (solomonzelman):

sorry:)

OpenStudy (solomonzelman):

I am doing a complete copy of the proof in a different thread. If you want you can as well wait for that ...

OpenStudy (legomyego180):

Sure! Sorry, hopping between this and lecture right now. Final is next week so going a little crazy

OpenStudy (solomonzelman):

I got part of it down ... I will post I got so far. \(\color{blue}{\displaystyle \lim_{k\to \infty}\frac{k^k}{(k+\beta )^k}=\lim_{k\to \infty}\left(\frac{k}{k+\beta}\right)^k }\) \(\tiny \\[0.9em]\) I will make a substitution: \(\displaystyle k+\beta=t \) This condition dictates: \(\displaystyle k=t-\beta \) And when \(\displaystyle k\to \infty \), I know that \(\displaystyle t\to \infty \) as well. \(\tiny \\[0.7em]\) \(\color{blue}{\displaystyle \lim_{k\to \infty}\left(\frac{k}{k+\beta}\right)^k==\lim_{t\to \infty}\left(\frac{t-\beta}{t}\right)^{t-1} }\) \(\tiny \\[1.3em]\) Now, I will reduce this limit: \(\color{blue}{\displaystyle \lim_{t\to \infty}\left(\frac{t-\beta}{t}\right)^{t-1}=\lim_{t\to \infty}\left(\frac{t-\beta}{t}\right)^{t}\times \lim_{t\to \infty}\left(\frac{t-\beta}{t}\right)^{-1} }\) \(\tiny \\[1.7em]\) \(\color{blue}{\displaystyle =\lim_{t\to \infty}\left(\frac{t-\beta}{t}\right)^{t}\times \lim_{t\to \infty}\left(\frac{t}{t-\beta}\right)^{1} }\) \(\color{blue}{\displaystyle =\lim_{t\to \infty}\left(\frac{t-\beta}{t}\right)^{t}\times \color{red}{1} }\) So, so far I know the following equivalence: \(\color{black}{\displaystyle \lim_{k\to \infty}\frac{k^k}{(k+\beta )^k}:=\lim_{t\to \infty}\left(\frac{t-\beta}{t}\right)^t }\) I will proceed to evaluate the last limit (to prove the initial one). I will re-write the last limit as follows (using log. propreties): \(\color{black}{\large \displaystyle e^{\lim_{t\to \infty}~\ln \left[\left(\frac{t-\beta}{t}\right)^t\right] } }\) and to avoid writing exponent of e every time, will just examine \(\color{black}{ \displaystyle \lim_{t\to \infty}~\ln \left[\left(\frac{t-\beta}{t}\right)^t\right] }\). \(\color{blue}{ \displaystyle \lim_{t\to \infty}~\ln \left[\left(\frac{t-\beta}{t}\right)^t\right] =\lim_{t\to \infty}~t\cdot \ln \left[\frac{t-\beta}{t}\right]=\lim_{t\to \infty}~\frac{\ln \left[\frac{t-\beta}{t}\right] }{1/t} }\)

OpenStudy (solomonzelman):

this is where we got till, take your time to read if you need .... (sorry for making it long, tried to make it readable and simple)

OpenStudy (solomonzelman):

\(\color{blue}{ \displaystyle \lim_{t\to \infty}~\frac{\ln \left[\frac{t-\beta}{t}\right] }{1/t}=\lim_{t\to \infty}~\frac{\ln \left[1-\frac{\beta}{t}\right] }{1/t}=\lim_{t\to \infty}~\frac{\frac{d}{dt}\left\{\ln \left[1-\frac{\beta}{t}\right]\right\} }{\frac{d}{dt}\left\{1/t\right\} } }\) Applying the rule for the derivative of ln(x), and the chain rule for inner argument, which is a power rule ... for the bottom, just the power rule \(\color{blue}{ \displaystyle=\lim_{t\to \infty}~\frac{\frac{1}{1-\frac{\beta}{t} }\times \frac{\beta }{t^2} }{-1/t^2} }\)

OpenStudy (solomonzelman):

now, what is needed is just to simply this limit. You can cancel 1/t^2 on top and bottom. \(\color{blue}{ \displaystyle=\lim_{t\to \infty}~\frac{\frac{\beta}{1-\frac{\beta}{t} } }{-1} =-\beta\cdot \lim_{t\to \infty}~\frac{1}{1-\frac{\beta}{t} } }\) \(\color{blue}{ \displaystyle=-\beta\cdot \lim_{t\to \infty}~\frac{1}{\frac{t-\beta}{t} }=-\beta\cdot \lim_{t\to \infty}~\frac{t}{t-\beta } =-\beta\times \color{red}{1} }\)

OpenStudy (solomonzelman):

Came to an end, just need to put the pieces to get: That limit that I examined (which was supposed to be in the exponent of e) is equal to \(-\beta\), therefore the whole limit is \(e^{-\beta}\). And thus you know that \(\color{blue}{\displaystyle \lim_{k\to \infty}\frac{k^k}{(k+\beta )^k}=e^{-\beta} }\)

OpenStudy (solomonzelman):

It is actually not that hard, I just wrote it in a very lengthy way ... if you have questions, ask.

OpenStudy (solomonzelman):

Well, if you really need something very short, then you can remember that \(\displaystyle \lim_{x \rightarrow \infty}\left[\frac{ x+a }{ x}\right]^x = e^a\) (your initial limit is also a version of this, just that a=-1 in your particular case, and to now this, you can just set a sub x+a=t, like I did in my proof.)

satellite73 (satellite73):

think this depends largely on your definition of \(e\)

OpenStudy (agent0smith):

I feel like there's a far easier/shorter method here. First remember the definition of e http://output.to/sideway/images/knowledge/mathematics/calculus/001/limit_a_01j.png Do some simplifying \[\large \lim_{k \rightarrow \infty}\frac{ k^k }{ (k+1)^k} = \] \[\large \lim_{k \rightarrow \infty}\left( \frac{ k }{ k+1} \right)^k=\] \[\large \lim_{k \rightarrow \infty}\left( \frac{ k+1 }{ k} \right)^{-k}=\] \[\large \lim_{k \rightarrow \infty}\left( 1+\frac{ 1 }{ k} \right)^{-k}=\]apply the power limit law seen here http://image.slidesharecdn.com/lecture5-limitlaws-140916111716-phpapp01/95/lecture-5-limit-laws-7-638.jpg?cb=1410866284 \[\large \left( \lim_{k \rightarrow \infty}\left( 1+\frac{ 1 }{ k} \right)^{k} \right)^{-1}=\]Now look back at the definition of e.

OpenStudy (agent0smith):

\[\large \left( \lim_{k \rightarrow \infty}\left( 1+\frac{ 1 }{ k} \right)^{k} \right)^{-1}= e^{-1}\]

satellite73 (satellite73):

guess @agent0smith beat me to it. if that is your definition of \(e\) then you are done in basically two steps

OpenStudy (legomyego180):

Ah, I see. Very cool. I like both ways, Soloman's way really showed me how it works and Agent show'd me the rule you can use to also verify it. Thanks for the detailed and well thought out responses guys, it's really helpful.

Can't find your answer? Make a FREE account and ask your own questions, OR help others and earn volunteer hours!

Join our real-time social learning platform and learn together with your friends!
Can't find your answer? Make a FREE account and ask your own questions, OR help others and earn volunteer hours!

Join our real-time social learning platform and learn together with your friends!