3
$\begingroup$

If we know the series sum given below converges to a value $C$(constant) $$\sum_{n=0}^{\infty}a_n =C \tag 2$$ Can we generate following in terms of C. values of $a_n$ will tend to zero as n goes to infinity and sum too converges to C..

  1. $\sum_{n=0}^{\infty}na_n $
  2. $\sum_{n=0}^{\infty} \frac{a_n}{n!}$
$\endgroup$
3
  • 3
    $\begingroup$ Note for 1, that if you take $a_n=\frac{1}{n^2}$ the sum $\sum_{n=0}^{\infty}{na_n}$ doesnt converge at all.. $\endgroup$
    – Snufsan
    Commented Aug 8, 2014 at 10:41
  • $\begingroup$ I can only say equation 2 is a convergent sum .. Means for large values each $a_n$ will tend to zero.. For 1 and 2 no such guarentee $\endgroup$
    – Nirvana
    Commented Aug 8, 2014 at 10:49
  • $\begingroup$ The fact that $a_n$ tends to zero doesent mean that the sum convergent (for instance the harmonic sum). For 2 that fact that you "reduced" the size of each $a_n$ is what makes the sum coverge (formally, you can use the limit comparison test with $a_n$). $\endgroup$
    – Snufsan
    Commented Aug 8, 2014 at 11:03

2 Answers 2

1
$\begingroup$

Counterexample for question #$1$:

  • $ a_n= \begin{cases} 8 & \text{$n=0$}\\ 0 & \text{$n>0$}\\ \end{cases} $

  • $\sum\limits_{n=0}^{\infty}a_n=8$

  • $\sum\limits_{n=0}^{\infty}na_n=0$

So you cannot express $\sum\limits_{n=0}^{\infty}na_n$ in terms $C$ for every given $a_n$.

$\endgroup$
1
$\begingroup$

no; let $a_n=2^{-(n+1)}$, $b_0=1, b_n=0$ otherwise $(n>0)$. $C=1$ in both cases, yet for $a_n$:
1. has value 1
2. has value $exp(1/2)$

while for $b_n$
1. has value 0
2. has value 1

EDIT: It seemed to me as if the only given was $C$, and not the general term of the series. There is no direct link between the value of C and the values of 1., 2. though. Simply permuting the elements of $a_n$ changes the values of both 1. and 2., so we see there are infinitely many values they can take for a given C. In fact, by constructing the series carefully, we can make 1. and 2. be anything given C (in infinitely many ways for each number we decide to make 1. and 2. equal to). Thus, in any case you wouldn't be relating the answer to C, but simply summing the resulting series and finding values 1. and 2. (unless you consider $(\sum_{n=0}^{\infty}na_n/C)*C$ a valid, meaningful relationship :D)
TL;DR C is irrelevant, all that matters is the general term.

However, Here is a trick for series 1.
Assuming everything is nice, write $f(x) = \sum_{n=0}^{\infty}na_nx^n$. We seek $f(1)$. Note that $f$ is the derivative of $F(x) =\sum_{n=0}^{\infty}a_nx^{(n+1)}$. If we can find $F$, then we can find $f$, and hence $f(1)$. This works with $a_n=2^{-(n+1)}$.

$\endgroup$
1
  • 2
    $\begingroup$ In the same vein, you can have many series, which converge to $\pi$ but have different terms (& thus, different values of $1.$ and $2.$). However your answer doesn't really show that you can't express $1.$ and $2.$ in terms of $C$ if given the general term. It seems like it would be difficult to do so though. $\endgroup$
    – Jam
    Commented Aug 8, 2014 at 10:33

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .