Hey! I want to think about evaluating: \[\int_0^\infty \lfloor x\rfloor e^x dx \text{ where } \lfloor . \rfloor\text{ is greatest integer function }\]
trying to figure out the floor function thingy
got it that is a long command for the floor function
anyways... \[n \le x<n+1 \text{ then } \lfloor x \rfloor=n \text{ where } n \text{ is integer }\] now I think I want to do integration by parts...
trying to think how to integrate the greatest integer function
\[\int\limits_0^\infty \lfloor x \rfloor e^x dx \\ \int\limits_0^1 0 \cdot e^x dx+\int\limits_1^2 1 \cdot e^x + \int\limits_2^3 2 \cdot e^x + \cdots + \int\limits_n^{n+1} n \cdot e^{x} dx +\cdots \]
\[\sum_{n=1}^{\infty} \int\limits _n^{n+1} n \cdot e^x dx\]
\[\sum_{n =1}^\infty n \int\limits _n^{n+1} e^x dx \\ \sum_{n=1}^\infty n e^x|_n^{n+1} \\ \sum_{n=1}^\infty n[e^{n+1}-e^{n}]\]
actually I guess I don't need integration by parts
thats looks very neat, but it doesn't converge right
yeah wolfram says it doesn't diverge I thought this was suppose to converge those for some reason did i make a mistake above
I mean it does diverge sorry
you know what maybe if the exponent on the x was negative it would have converged
yeah exponential overtakes any polynomial, floor(x) is cheaper than a linear polynomial
did something wrong
oops didn't mean to delete that without copying it
\[\sum_{n=1}^{\infty}n(e^{-(n+1)}-e^{-n})\] whatever I know this converges I just need to find the sum now I was thinking this maybe a telescoping series
\[\int\limits_{0}^{\infty} \lfloor x \rfloor e^{-x} dx \\ =\sum_{n=1}^{\infty} \int\limits_{n}^{n+1} n e^{-x} dx \\ =\sum_{n=1}^{\infty} -n e ^{-x}|_n^{n+1} \\ \\ =\sum_{n=1}^{\infty} -n[e^{-(n+1)}-e^{-n}] \\ \text{ ratio test: } \\ \lim_{n \rightarrow \infty} \frac{-(n+1)[e^{-(n+2)}-e^{-(n+1)]}}{-n [e^{-(n+1)}-e^{-n}]}\] \[\lim_{n \rightarrow \infty}-(1+\frac{1}{n}) \frac{ e^{-(n+2)+(n+1)-e^{-(n+1)+(n+1)}}}{e^{-(n+1)+(n+1)}-e^{-n+(n+1)} } \\ -(1+0) \frac{e^{-1}-e^{0}}{e^{0}-e^{1}} =-1 \frac{e^{-1}-1}{1-e} =-1 \frac{1-e}{e-e^2} =\frac{e-1}{e-e^2} \\ =\frac{e-1}{e(1-e)}=\frac{-1}{e}\]
\[(e^{-2}-e^{-1})+2(e^{-3}-e^{-2})+3(e^{-4}-e^{-3})+4(e^{-5}-e^{-4}) +\cdots \] not exactly a telescoping series the way it is written at last
i pulle that from ur deleted replies
thanks @ganeshie8 now people can know what problem I'm talking about I'm trying to evaluate: \[\int\limits_{0}^{\infty} \lfloor x \rfloor e^{-x} dx\]
mimicing ur earlier work, \[\int\limits_{0}^{\infty} \lfloor x \rfloor e^{-x} dx = (e-1)\sum\limits_{n=1}^{\infty} \frac{n}{e^n} \]
so the first floor (x) e^x diverge or not?
i showed it diverge so I think it diverge (unless I made some mistake)
well wolfram showed my sum diverged :p
\(\color{blue}{\text{Originally Posted by}}\) @freckles \[\sum_{n =1}^\infty n \int\limits _n^{n+1} e^x dx \\ \sum_{n=1}^\infty n e^x|_n^{n+1} \\ \sum_{n=1}^\infty n[e^{n+1}-e^{n}]\] \(\color{blue}{\text{End of Quote}}\) from tis result ?
I could have use the ratio test to come up with that know
ues from that last line
ah ok i guess ration test can testify divergence
we can simply use the limit test
oh yes! that limit does not go to 0
\[\sum\limits_{n=1}^{\infty} \frac{n}{e^{nx}}=\sum\limits_{n=1}^{\infty} \frac{d}{dx}e^{-nx}=\frac{d}{dx}\sum\limits_{n=1}^{\infty} e^{-nx}= \frac{d}{dx}\frac{e^{-x}}{1-e^{-x}}=\cdots\]
and we know \[\frac{d}{dx} \frac{e^{-x}}{1-e^{-x}} \\ \frac{d}{dx} \frac{1}{e^x-1} \\ \frac{d}{dx}(e^x-1)^{-1} \\ -1(e^x-1)^{-2}(e^x) \\ - \frac{e^x}{(e^x-1)^2}\] but we should evaluate this at x=1 since we had that we really wanted to evaluate \[\int\limits\limits_{0}^{\infty} \lfloor x \rfloor e^{-x} dx = (e-1)\sum\limits_{n=1}^{\infty} \frac{n}{e^n} \] \[(e-1) \frac{-e}{(e-1)^2}=\frac{-e}{e-1}\] but I think we are to get 1/(e-1)
Oh right \[\int\limits\limits_{0}^{\infty} \lfloor x \rfloor e^{-x} dx =-\sum_{n=1}^\infty n[e^{-(n+1)}-e^{-n}] = \frac{e-1}{e}\sum\limits_{n=1}^{\infty} \frac{n}{e^n}\]
algebra mistakes
thanks @ganeshie8 ok I was sorta failing at the algebra a little too but I see it now
I can't remember who gave me this problem but @Loser66 this is also on your gre practice exam if you want to look at this.
i remember working this problem differently sometime back here in openstudy but google isn't helping at the moment..
And I bet you have answered so many questions since this that going back through your old questions is horrifying and it also doesn't work to well sometimes. Sometimes it automatically throws me out of the looking through my old question thing. Not sure if you know what I mean.
Thanks so mmmmmmmmuch
truly i could not see how you threw differentiation in there @gabylovesu how is that sum equal to that differentiation ?
@ganeshie8
wrong tagging lol
Have you tried parameterizing and applying the Laplace transform? I'm thinking something along the lines of \[I(s)=\int_0^\infty \lfloor x\rfloor e^{-sx}\,dx=\mathcal{L}\{\lfloor x\rfloor\}\] The transform definitely exists because \(\lfloor x\rfloor\) is piecewise continuous and of exponential order.
\[\color{red}{\sum\limits_{n=1}^{\infty} \frac{n}{e^{nx}}=-\sum\limits_{n=1}^{\infty} \frac{d}{dx}e^{-nx}}=-\frac{d}{dx}\sum\limits_{n=1}^{\infty} e^{-nx}= -\frac{d}{dx}\frac{e^{-x}}{1-e^{-x}}=\cdots\] I think it would be easier to make sense of it by differentiating \(\large \color{red}{e^{-nx}}\) and seeing that you get back the starting expression...
hey @SithsAndGiggles I haven't thought of that but I would be interested in seeing that way if you wanted to show that way. I honestly I haven't seen any laplace transform action in like 10 years so I kind of forgot all of that. :p
The idea would be to express \(\lfloor x\rfloor e^{-x}\) in terms of the step function \[\theta(x-c)=\begin{cases}1&\text{for }x\ge c\\0&\text{for }x<c\end{cases}\]
ooh i see it now
\[\lfloor x\rfloor e^{-x}=\sum_{c=1}^\infty ce^{-x}\Bigg(\theta(x-c)-\theta(x-c-1)\Bigg)\] So \[\begin{align*}\mathcal{L}\left\{\lfloor x\rfloor\right\}&=\int_0^\infty \lfloor x\rfloor e^{-sx}\,dx\\ &=\int_0^\infty \sum_{c=1}^\infty ce^{-sx}\Bigg(\theta(x-c)-\theta(x-c-1)\Bigg)\,dx\\ &=\sum_{c=1}^\infty c\int_0^\infty e^{-sx}\Bigg(\theta(x-c)-\theta(x-c-1)\Bigg)\,dx\\ &=\sum_{c=1}^\infty c\mathcal{L}\{\theta(x-c)-\theta(x-c-1)\}\\ &=\sum_{c=1}^\infty c\frac{e^{-cs}-e^{-(c+1)s}}{s} \end{align*}\] Setting \(s=1\) gives the series @ganeshie8 was working with. Nothing new I suppose :P
does seem similar but still pretty @SithsAndGiggles anyways this is making me hungry so peace and thanks for the fun
I will come back later though to analyze more deeply what you said
As for the actual result: \[\sum_{c=1}^\infty c(e^{-c}-e^{-(c+1)})=\sum_{c=1}^\infty c\left(\frac{1}{e^c}-\frac{1}{e^{c+1}}\right)=\frac{1}{e}+\frac{1}{e^2}+\frac{1}{e^3}+\cdots=\frac{1}{1-\frac{1}{e}}-1\]
Join our real-time social learning platform and learn together with your friends!