Find a formula for the partial sum
\[\large 1+4r+9r^2+\cdots+n^2r^{n-1}\]

- ganeshie8

Find a formula for the partial sum
\[\large 1+4r+9r^2+\cdots+n^2r^{n-1}\]

- Stacey Warren - Expert brainly.com

Hey! We 've verified this expert answer for you, click below to unlock the details :)

- katieb

I got my questions answered at brainly.com in under 10 minutes. Go to brainly.com now for free help!

- UnkleRhaukus

\[\sum_{i=1}^{n}i^2r^{i-1}\]

- anonymous

What have you tried so far?

- ganeshie8

I actually found two different solutions, just trying to understand this problem better by looking at how others approach this

Looking for something else?

Not the answer you are looking for? Search for more explanations.

## More answers

- anonymous

\[
f(r) = \sum_{k=0}^{n}r^k=\frac{1-r^{n+1}}{1-r}\\
f'(r) = \sum_{k=1}^{n}kr^{k-1}\\
rf'(r) = \sum_{k=1}^{n}kr^{k}\\
(rf'(r))' = \sum_{k=1}^{n}k^2r^{k-1} = rf''(r) + f'(r)
\]

- ganeshie8

Wow! that's almost same as the first solution I have :
\[\sum_{i=1}^{n}i^2r^{i-1} =\dfrac{d}{dr} r\sum_{i=1}^{n} \left(r^i\right)' =\dfrac{d}{dr} r\left(\sum_{i=1}^{n} r^i\right)' = \dfrac{d}{dr} r \left(r\dfrac{r^n-1}{r-1}\right)' \]

- ganeshie8

actually it is identical xD

- ganeshie8

The second method uses some kind of convolution of two interleaving sequences

- anonymous

\[
f(r) = \sum_{k=0}^{n}r^k
\]And \[
S_n = \sum_{k=1}^nkr^{k-1}
\]Then\[
rS_n = \sum_{k=1}^nkr^{k}
\]Let \(k=m-1\):\[
rS_n = \sum_{m=2}^{n+1}(m-1)r^{m-1} = \sum_{m=2}^{n+1}mr^{m-1} -\sum_{m=2}^{n+1}r^{m-1}
\]Undo the sub for the second term:\[
rS_n= \sum_{m=2}^{n+1}mr^{m-1} -\sum_{k=1}^{n}r^{k}
\]Add missing terms and remove extraneous terms:\[
rS_n+1 - (n+1)r^{n+1} = S_n - f(r)
\]So we get \[
S_n = \frac{1-(n+1)r^{n+1}+f(r)}{1-r}
\]When we do:\[
\Psi _n = \sum_{k=1}^nk^2r^{k-1}
\]We get \[
r\Psi_n = \sum_{m=2}^n(m-1)^2r^{m-1}
\]And we can use previous methods to solve for it.

- ganeshie8

Looks pretty neat, so if i understand correctly we need to find \(\sum kr^{k-1}\) before finding \(\sum k^2r^{k-1}\)

- ganeshie8

\[r\Psi_n = \sum_{m=\color{red}{1}}^n(m-1)^2r^{m-1} = \Psi_n -2\sum_{m=\color{red}{1}}^nmr^{m-1} + \sum_{m=\color{red}{1}}^n r^{m-1}\]

- ganeshie8

we can simply plugin the previous result for the sum in middle, awesome!

- anonymous

\[
f(m,r) = \sum_{k=0}^nk^mr^{k}
\]And \[
rf(m,r) = \sum_{k=0}^nk^mr^{k+1} =\sum_{k'=1}^{n+1}(k'-1)^mr^{k'}
\]\[
rf(m,r)+(-1)^m-n^mr^{n+1}= \sum_{p=0}^{q}{q\choose p}(-1)^pf(q-p,r)
\]This is sort of our recursive solution

- anonymous

Hmmm, I guess to get it explicit:\[
f(m,r) = \frac{(-1)^m-n^mr^{n+1}+\sum_{k=0}^{n-1}{n\choose k}(-1)^pf(q-p,r)}{1-r}
\]

- anonymous

I know I made a small error in it

- ganeshie8

that was a typo in the exponent, its okay it is a beautiful solution !

- anonymous

\[
f(m,r) = \frac{(-1)^m-n^mr^{n+1}-\sum_{k=0}^{m-1}{m\choose k}(-1)^pf(m-k,r)}{1-r}
\]This one seems to work for \(m=0\).

- anonymous

Hmm, I still shouldn't have that n term, they should be m most likely

- sparrow2

what is r here?

- ganeshie8

I think r could be any real number

- ganeshie8

\[\large f(m,r) = \frac{1}{1-r}\left[n^m(r^n-1)+\sum\limits_{k=2}^n \sum\limits_{i=0}^{m-1} \binom{m}{i}k^i(r^k-1)\right]\]

- anonymous

The whole point though is to get rid of that outer sum. And you shouldn't have any \(n\) terms.

- ganeshie8

"n" is the index, nth partial sum right

- anonymous

oh, right, I forgot

- ganeshie8

I was just messing with your work and ended up wid above formula
it is kind of recursive too as the right side part has one exponent less..

- anonymous

Everything goes to hell when you let \(r=1\) though.

- anonymous

\[
f(m, 1) = \sum_{k=0}^nk^m
\]

- ganeshie8

Haha lets just say \(r\ne 1\) then

- anonymous

But \(r=1\) is an interesting series.

- ganeshie8

Indeed, it can be shown that that series asymptotically approaches \(\dfrac{n^{m+1}}{m+1}\)
\[f(m,1) \sim \dfrac{n^{m+1}}{m+1}\]

- anonymous

Is there a formula for that series?

- ganeshie8

Yes, was working on some definite integral last night and ended up wid above result...
let me pull up..

- anonymous

I think \(f(-1,1)\) is known not to have an elementary function.

- ganeshie8

Sorry it is not a formula, just a limit which gives the asymptotic sum

- ganeshie8

\[\lim\limits_{n\to\infty}\dfrac{f(m,1)}{n^{m+1}} = \frac{1}{m+1}\]

- sparrow2

for r=1 sum if n(n+1)(2n+1)/6 you can prove using induction

- ganeshie8

http://math.stackexchange.com/questions/936924/limit-of-the-sequence-frac1k2k-nknk1

- ganeshie8

right @sparrow2 thats sum of squares of first n natural numbers, fun one!

- sparrow2

i think the point is that we just should write the sum in the form so when you give me a natural number for ex n=39 i will give you the answer not the sum written in other form

- sparrow2

like in sum of squares of first n

- ganeshie8

you mean explicit formula

- sparrow2

i don't know what it is called

- sparrow2

i don't know math's english terminology very well :)

- ganeshie8

got you :) basically we're trying to find a recursive relation for
\[\large 1^m + 2^mr + 3^mr^2+\cdots+n^mr^{n-1}\]
when \(r=1\), we get the series
\[1^m+2^m+3^m+\cdots+n^m\]

- ganeshie8

yeah the original question is a special case, \(m=2\)

- Empty

I am a little late to the party, but I thought I'd try to challenge myself to come up with a different approach. I think it doesn't help much, but is there anything to be gained by writing the squares as:
\[k^2 = \sum_{n=1}^k (2n-1) \]
My thought was to switch the order of summation signs, however it won't work here since the upper limit k is summed over in the other summation, oh well.

- ganeshie8

Thats similar to what wio did in his recurrence relation,
that works pretty nicely as it reduces the exponent.. .

- Empty

Oh I thought he just did the derivatives of the geometric series approach.

- ganeshie8

He did it in two ways..

- sparrow2

and do we know that there is a formula for this?

- ganeshie8

check this @sparrow2
https://en.wikipedia.org/wiki/Faulhaber%27s_formula

- sparrow2

ok thanks

- ganeshie8

generating functions might also work i think
post the work @ikram002p

- ganeshie8

Let \[f(n,m,r)=\sum\limits_{k=1}^n k^mr^{k-1} \]
Let \(a_k = k^m\) and \(b_k = r^{k-1}\),
then \(B_n=\sum\limits_{k=1}^n b_k = \dfrac{r^n-1}{r-1}\)
It is easy to see that \(b_n = B_{n+1}-B_n\)
\[\begin{align}\color{blue}{f(n,m,r)}&=\sum\limits_{k=1}^n a_k b_k \\~\\
&=\sum\limits_{k=1}^n a_k (B_{n+1}-B_n) \\~\\
&= a_nB_n+\sum\limits_{k=2}^n (a_{k-1}-a_{k}) B_k \\~\\
&= n^m\dfrac{r^n-1}{r-1}+\frac{1}{1-r}\sum\limits_{k=1}^n [(k-1)^m-k^m] (r^k-1) \\~\\
\end{align}\]

- anonymous

Liam did an investigation to see how water temperature affects the amount of salt that will dissolve in the water. He filled 4 beakers with exactly 100 milliliters of water each. He then heated each beaker to a different temperature and tested salt solubility at each different temperature. What was Liam's independent variable?
amount of water in each beaker
the number of beakers
the amount of salt that dissolved in each beaker
the temperature of the water in each beaker

- anonymous

A student conducts an experiment to determine how the additional of salt to water affects the density of the water. The student fills 3 beakers with equal amounts of water. He then adds 1 cup of salt to the first beaker, 2 cups of salt to the second beaker and no salt to the third beaker. What is the dependent variable in the student's investigation?
the amount of salt in the water
the temperature of the water
the density of the water
the ph of the water

Looking for something else?

Not the answer you are looking for? Search for more explanations.