Try This Sum

Calculus Level 5

n = 2 e 2 n n 2 1 \large\sum_{n=2}^{\infty}\frac{{e}^{-2n}}{{n}^{2}-1}

Consider the infinite sum above. If the solution can be expressed in the form 1 A ( 1 + e B C ) + sinh ( D ) ln ( 1 e E ) \frac{1}{A}\left(1+\frac{{e}^{-B}}{C}\right)+\sinh{(D)}\ln{(1-{e}^{-E})}

find A × B × C × D × E . A\times B\times C\times D\times E.


The answer is 32.

This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try refreshing the page, (b) enabling javascript if it is disabled on your browser and, finally, (c) loading the non-javascript version of this page . We're sorry about the hassle.

3 solutions

Kartik Sharma
Jun 17, 2015

The adjacent solution is great and perfect for the problem so great like this.

Anyways here is how I did it

S = n = 2 e 2 n n 2 1 \displaystyle S = \sum_{n=2}^{\infty}{\frac{{e}^{-2n}}{{n}^{2}-1}}

What the function under the sigma operation can remind? First of all, it can create a lot of tension for sure.

Can it be written in the form of a Laurent series and then supposedly Residue Theorem? Well, I don't think so. Then, some other integration under the sum sign?

Yeah. Laplace transform!

L 1 ( e 2 n n 2 1 ) = u 2 ( x ) s i n h ( x 2 ) \displaystyle {L}^{-1}\left(\frac{{e}^{-2n}}{{n}^{2}-1}\right) = {u}_{2}(x)sinh(x-2)

as it is of the form e c s F ( s ) \displaystyle {e}^{-cs} F(s) , F ( s ) = m s 2 m 2 \displaystyle F(s) = \frac{m}{{s}^{2}-{m}^{2}} and s = n , m = 1 \displaystyle s = n, m = 1

where L 1 \displaystyle {L}^{-1} means inverse Laplace transform, u c ( x ) \displaystyle {u}_{c}(x) means heaviside step function. For finding the inverse Laplace transform, one can either peek it from the table or the Bromwich integral(although latter is quite hard to do).

Hence,

L 1 ( S ) = n = 2 u 2 ( x ) s i n h ( x 2 ) \displaystyle {L}^{-1}(S) = \sum_{n=2}^{\infty}{{u}_{2}(x)sinh(x-2)}

S = n = 2 0 u 2 ( x ) s i n h ( x 2 ) e n x d x \displaystyle S = \sum_{n=2}^{\infty}{\int_{0}^{\infty}{{u}_{2}(x)sinh(x-2){e}^{-nx}} dx}

Using GP formula,

S = 0 u 2 ( x ) e x e x 1 s i n h ( x 2 ) d x \displaystyle S = \int_{0}^{\infty}{{u}_{2}(x) \frac{{e}^{-x}}{{e}^{x}-1} sinh(x-2) dx}

which is just

S = L ( u 2 ( x ) s i n h ( x 2 ) e x 1 ) \displaystyle S = L\left(\frac{{u}_{2}(x)sinh(x-2)}{{e}^{x}-1}\right)

equals

S = 1 2 ( 1 + e 2 2 ) + s i n h ( 2 ) l n ( 1 e 2 ) \displaystyle S = \frac{1}{2}\left(1 + \frac{{e}^{-2}}{2}\right) + sinh(2)ln(1-{e}^{-2})

I have not provided details of how to find Laplace (inverse) transform as it will just make the solution longer, you have to at last peep from the table only :P

Kartik Sharma - 5 years, 12 months ago

You missed the d x \mathrm dx in all the integrals.

Prasun Biswas - 5 years, 11 months ago

Log in to reply

Thanks a lot! That is a very common mistake of mine and I need to improve there. BTW, how did you do it?

Kartik Sharma - 5 years, 11 months ago

Log in to reply

Haha, I did it in a way similar to Oussama below. The only difference is that I used the Maclaurin series of ln ( 1 + x ) \ln(1+x) (with a bit of modification) instead of evaluating I I like he did.

To be honest, I don't understand squat about how you evaluated it. I need to learn all these integral transforms soon. :)

Prasun Biswas - 5 years, 11 months ago
Oussama Boussif
Jun 17, 2015

Before we start evaluating this sum let's simplify it into an easier one: S = n = 2 e 2 n n 2 1 = 1 2 n = 2 ( e 2 n n 1 e 2 n n + 1 ) 2 S = S S S\quad =\quad \sum _{ n=2 }^{ \infty }{ \frac { { e }^{ -2n } }{ { n }^{ 2 }-1 } } \\ \quad \quad =\quad \frac { 1 }{ 2 } \sum _{ n=2 }^{ \infty }{ (\frac { { e }^{ -2n } }{ n-1 } } \quad -\quad \frac { { e }^{ -2n } }{ n+1 } )\\ \quad \Rightarrow 2S\quad =\quad S'-S''

Let's simplify S S' and S S'' :

S = n = 2 e 2 n n 1 = e 2 n = 2 e 2 ( n 1 ) n 1 = e 2 k = 1 e 2 k k = e 2 I S'\quad =\quad \sum _{ n=2 }^{ \infty }{ \frac { { e }^{ -2n } }{ n-1 } } \quad =\quad { e }^{ -2 }\sum _{ n=2 }^{ \infty }{ \frac { { e }^{ -2(n-1) } }{ n-1 } } \\ \qquad \qquad \qquad \qquad \quad =\quad { e }^{ -2 }\sum _{ k=1 }^{ \infty }{ \frac { { e }^{ -2k } }{ k } } \quad =\quad { e }^{ -2 }I And : S = n = 2 e 2 n n + 1 = e 2 n = 2 e 2 ( n + 1 ) n + 1 = e 2 k = 3 e 2 k k = e 2 ( k = 1 e 2 k k ( e 2 + e 4 2 ) ) = e 2 ( I ( e 2 + e 4 2 ) ) S''\quad =\quad \sum _{ n=2 }^{ \infty }{ \frac { { e }^{ -2n } }{ n+1 } } \quad =\quad { e }^{ 2 }\sum _{ n=2 }^{ \infty }{ \frac { { e }^{ -2(n+1) } }{ n+1 } } \\ \qquad \qquad \qquad \qquad \quad =\quad { e }^{ 2 }\sum _{ k=3 }^{ \infty }{ \frac { { e }^{ -2k } }{ k } } \\ \quad \quad \quad \quad \quad \quad \quad \quad \quad \quad \quad \quad \quad =\quad { e }^{ 2 }(\sum _{ k=1 }^{ \infty }{ \frac { { e }^{ -2k } }{ k } } -\quad ({ e }^{ -2 }+\frac { { e }^{ -4 } }{ 2 } ))\\ \quad \quad \quad \quad \quad \quad \quad \quad \quad \quad \quad \quad \quad =\quad { e }^{ 2 }(I-({ e }^{ -2 }+\frac { { e }^{ -4 } }{ 2 } ))

Now let's evaluate I I : We consider the following generating function y = f ( x ) = k = 1 e 2 k k x n = x e 2 + k = 1 e 2 ( k + 1 ) k + 1 x k + 1 y = e 2 + k = 1 e 2 ( k + 1 ) x k y = e 2 + e 2 k = 1 e 2 k x k y\quad =\quad f(x)\\ \quad \quad =\quad \sum _{ k=1 }^{ \infty }{ \frac { { e }^{ -2k } }{ k } } { x }^{ n }\\ \quad \quad =\quad x{ e }^{ -2 }+\sum _{ k=1 }^{ \infty }{ \frac { { e }^{ -2(k+1) } }{ k+1 } } { x }^{ k+1 }\\ \Rightarrow y'\quad =\quad { e }^{ -2 }+\sum _{ k=1 }^{ \infty }{ { e }^{ -2(k+1) } } { x }^{ k }\\ \quad \quad y'\quad =\quad { e }^{ -2 }+{ e }^{ -2 }\sum _{ k=1 }^{ \infty }{ { e }^{ -2k } } { x }^{ k } And : y = f ( x ) = k = 1 e 2 k k x n y = k = 1 e 2 k x k 1 x y = k = 1 e 2 k x k y\quad =\quad f(x)\\ \quad \quad =\quad \sum _{ k=1 }^{ \infty }{ \frac { { e }^{ -2k } }{ k } } { x }^{ n }\\ \Rightarrow y'\quad =\quad \sum _{ k=1 }^{ \infty }{ { e }^{ -2k } } { x }^{ k-1 }\\ \quad xy'\quad =\quad \sum _{ k=1 }^{ \infty }{ { e }^{ -2k } } { x }^{ k }

So, we get the following differential equation and it's solution :

y = e 2 + e 2 x y y ( 1 x e 2 ) = e 2 y = e 2 1 x e 2 y = l n ( 1 x e 2 ) + C y'={ e }^{ -2 }+{ e }^{ -2 }xy'\\ y'(1-x{ e }^{ -2 })={ e }^{ -2 }\\ y'=\frac { { e }^{ -2 } }{ 1-x{ e }^{ -2 } } \\ \Rightarrow y\quad =\quad -ln(1-x{ e }^{ -2 })+C

We can verifiy that for x = 0 x=0 , y = 0 y=0 , So C = 0 C=0 :

I = f ( 1 ) = l n ( 1 e 2 ) I=f(1)=-ln(1-{ e }^{ -2 })

Now we can find the first Sum:

2 S = I ( e 2 e 2 ) + ( 1 + e 2 2 ) 2S={ I(e }^{ -2 }-{ e }^{ 2 })+(1+\frac { { e }^{ -2 } }{ 2 } ) And by substituting I I , we get:

S = l n ( 1 e 2 ) e 2 e 2 2 + 1 2 ( 1 + e 2 2 ) S = l n ( 1 e 2 ) s i n h ( 2 ) + 1 2 ( 1 + e 2 2 ) S=ln(1-{ e }^{ -2 })\frac { { e }^{ 2 }-{ e }^{ -2 } }{ 2 } +\frac { 1 }{ 2 } (1+\frac { { e }^{ -2 } }{ 2 } )\\ S=ln(1-{ e }^{ -2 })sinh(2)+\frac { 1 }{ 2 } (1+\frac { { e }^{ -2 } }{ 2 } )

You have a typo in the last three lines. It should be S S , not S S' .


Just for the record, there's a simpler way if you recall the Maclaurin series for ln ( 1 + x ) \ln(1+x) which is k = 1 ( 1 ) k 1 x k k \sum\limits_{k=1}^\infty(-1)^{k-1}\dfrac{x^k}{k} which converges iff x 1 x ( 1 ) |x|\leq 1~\land~x\neq (-1) , you can easily manipulate it a bit (change x x to ( x ) (-x) and then multiply the sum by ( 1 ) (-1) keeping note of the convergence conditions) and then substitute x = e 2 x=e^{-2} to get the value of I I .

Prasun Biswas - 5 years, 11 months ago

Log in to reply

Thanks for pointing that typo out. And I feel very stupid now because I didn't think of that Maclaurin series.

Oussama Boussif - 5 years, 11 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...