Find the function for this Taylor series

Join the discussion
Ask a follow-up here, or get your own question answered by working scientists, mathematicians and engineers — people, not an autocomplete.
Real named experts · corrections over time · the nuance an AI answer skips
14 replies · 3K views
arithmetic
Messages
32
Reaction score
0
[itex]\sum_{m=0}^\infty \frac{(m-1)^{m-1}x^{m}}{m!}[/itex]


Interesting result...
 
Last edited:
Physics news on Phys.org
I doubt there is a closed form for it. Plus, it only converges if |x|<1/e. After looking at the graph, it's similar to the Lambert W function when |x|<1/e: [tex]-W(-x)=\sum_{m=1}^{\infty}{\frac{m^{m-1}x^{m}}{m!}}[/tex] Subtracting 1 from both sides will make it approximately equal to your sum. However, there is no way to manipulate my series to put yours in terms of the W function.
 
Last edited:
Another interesting thing of note, if [tex]f(x)=\sum_{m=0}^\infty \frac{(m-1)^{m-1}x^{m}}{m!}[/tex] then [tex]f'(x)=\sum_{n=0}^\infty \frac{n^{n}x^{n}}{n!}[/tex] as long as |x|<1/e.
 
Last edited:
Lagrange inversion...

1/(1- ...) = ...

voilá
 
The Lagrange Inversion Theorem is used to find the inverse of a function. You didn't ask for the inverse in your original post :P
 
TheFool said:
After looking at the graph, it's similar to the Lambert W function when |x|<1/e: [tex]-W(-x)=\sum_{m=1}^{\infty}{\frac{m^{m-1}x^{m}}{m!}}[/tex] Subtracting 1 from both sides will make it approximately equal to your sum. However, there is no way to manipulate my series to put yours in terms of the W function.
How about this? Integrate W wrt x:
[tex]\sum_{m=1}^{\infty}{\frac{m^{m-1}x^{m+1}}{(m+1)!}} = \sum_{m=2}^{\infty}{\frac{(m-1)^{m-2}x^{m}}{m!}} = x\sum_{m=2}^{\infty}{\frac{(m-1)^{m-2}x^{m-1}}{m!}}[/tex]
Now divide by x and differentiate:
[tex]\sum_{m=2}^{\infty}{\frac{(m-1)^{m-1}x^{m-2}}{m!}}[/tex]
.. and finally multiply by x2

[edit] ... finally finally, add in the m = 1 term
 
Last edited:
Well, it would seem I don't belong posting in this forum. I shouldn't have missed that.
 
TheFool said:
Well, it would seem I don't belong posting in this forum. I shouldn't have missed that.

No, that`s wrong.

Yours is shorter and better. From what you stated, you just need one step further ...
and voilá, 1/ 1-...
 
arithmetic said:
[itex]\sum_{m=0}^\infty \frac{(m-1)^{m-1}x^{m}}{m!}[/itex]
Is this right?
[itex]f(x) = \sum_{m=0}^\infty \frac{(m-1)^{m-1}x^{m}}{m!} = xG(x) - ∫G(x)[/itex]
where G(x) = -W(-x).
Hence [itex]f(x) = ∫x.dG(x)[/itex]
But I don't know where we're trying to get to.
 
micromass said:
Did you actually know the answer to this problem??

yes, I know













p.3, 14
 
Last edited:
Even though I showed I really am a fool for missing something so obvious, I decided to finish this.

[tex]y=\sum_{m=0}^{\infty}\frac{(m-1)^{m-1}x^{m}}{m!}[/tex]
[tex]y'=\sum_{m=0}^{\infty}\frac{m^{m}x^{m}}{m!}[/tex]
Move on to the W function.
[tex]-W(-x)=\sum_{m=1}^{\infty}\frac{m^{m-1}x^{m}}{m!}[/tex]
Differentiate.
[tex]\frac{-W(-x)}{x(W(-x)+1)}=\sum_{m=1}^{\infty}\frac{m^{m}x^{m-1}}{m!}[/tex]
[tex]\frac{-W(-x)}{W(-x)+1}=\sum_{m=1}^{\infty}\frac{m^{m}x^{m}}{m!}[/tex]
Take note the subscript in that sum is m=1 instead of m=0.
[tex]y'=1-\frac{W(-x)}{W(-x)+1}[/tex]
[tex]y=\frac{x}{W(-x)}+C[/tex]
In another form.
[tex]y=-e^{W(-x)}+C[/tex]
Both are true when |x|<1/e.
 
Last edited:
Congratulations FOOL.
There are other ways to get the function, but yours is great and different to others I have seen in the literature.

Also congratulations to haruspex who did a great job, indeed, I wonder how you could posibly have found the function from your first post where you handled the sums, their indexes and the x's and x^2's...Just another note:
FOOL, you say |x|<1/e

however W(-1/e)=-1 and
y={1/e}/{W(-1/e)}+C= -1/e +C

What do you say about that?
 
Last edited:
I say that |x|<1/e since that is the radius of convergence of [tex]y=\sum_{m=0}^{\infty}\frac{(m-1)^{m-1}x^{m}}{m!}[/tex]