Group Abstract Group Abstract

Message Boards Message Boards

0
|
69 Views
|
3 Replies
|
4 Total Likes
View groups...
Share
Share this post:

Finding general terms of a series

Posted 1 day ago

I want to find general terms of this series I already used command ''FindSequenceFunction''. Help me

FindSequenceFunction[{E^-m, E^-m (2 - m), 
  FullSimplify[ (1 - 2 m + m^2/2)], FullSimplify[(-m + m^2 - m^3/6)], 
  FullSimplify[ (m^2/2 - m^3/6 + m^4/24)], 
  FullSimplify[ (-(m^3/6) + m^4/24 - m^5/120)]}, m]
POSTED BY: Somveer Keshav
3 Replies

Just to be clear, what is sought is to find the general term $a_n$ of the sequence $$ \eqalign{ a_0&=e^{-m}, \cr a_1&=e^{-m} (2-m), \cr a_2&=\frac{1}{2} (m-4) m+1, \cr a_3&=-\frac{1}{6} m ((m-6) m+6), \cr a_4&=\frac{1}{24} m^2 ((m-4) m+12), \cr a_5&=-\frac{1}{120} m^3 ((m-5) m+20), \cr } $$ Right?

There's not much of a pattern there. A factor of $(-1)^n/n!$ Another factor of $m^{n-2}$ for $n\ge2$, which mysteriously becomes $e^{-m}$ for $n=0,1$, perhaps. Finally, again for $n\ge2$, a factor of the form $(m-b_n)m+n(n-1)$, where $b_n$ is a seemingly random sequence, $b_2=4$, $b_3=6$, $b_4=4$, $b_5=5$. Except for the factor $e^{-m}$, the term $a_1$ could be seen to fit the pattern of $a_n$ for $n\ge2$. Then $b_n=2n$ for $n=1,2,3$ and $b_n=n$ for $n=4,5$ — who knows what happens for $n>5$?

Here's the sequence based on the pattern for $a_2$, $a_3$, if two terms constitute a pattern:

Table[
   Subscript[a, n] == (-1)^n/n! m^(n - 2) ((m - 2 n) m + n (n - 1)),
   {n, 0, 5}] // Column // TeXForm

$$ \eqalign{ a_0&=1, \cr a_1&=2-m, \cr a_2&=\frac{1}{2} ((m-4) m+2), \cr a_3&=-\frac{1}{6} m ((m-6) m+6), \cr a_4&=\frac{1}{24} m^2 ((m-8) m+12), \cr a_5&=-\frac{1}{120} m^3 ((m-10) m+20), \cr } $$

In fact, we see that both $a_0$ and $a_1$ in the OP's sequence would fit this formula if the mysterious factor $e^{-m}$ were a mistake and could be dropped from each term. Where this sequence differs from the OP's suggests that the OP's sequence is much more complicated, that is, assuming it contains no errors. In my experience, I have to give FindSequenceFunction[] a couple dozen or more terms for it to discover a formula for a complicated sequence.

POSTED BY: Michael Rogers

Ignoring the exponential terms, the closest I arrived is this:

Table[(-m)^(n - 1)/(n + 1)!
  (n + n^2 + (m - a[n]) m),
 {n, -1, 4}]

where a[n] should come from

FindSequenceFunction[{0, 2, 4, 6, 4, 5}, n]

if only it gave an answer. I was disappointed that this too gave no answer:

FindSequenceFunction[{2, 6, 24, 120}, n]
POSTED BY: Gianluca Gorni

I will venture to guess that the middle terms of the first and second elements are not quite what was intended. This would make the task more difficult.

POSTED BY: Daniel Lichtblau
Reply to this discussion
Community posts can be styled and formatted using the Markdown syntax.
Reply Preview
Attachments
Remove
or Discard