A Why Question:Taylor Polynomial of e^x over x?

  • Thread starter Thread starter RadiantL
  • Start date Start date
  • Tags Tags
    E^x Polynomial
AI Thread Summary
The discussion centers on the use of Taylor and Laurent series for approximating functions like e^x/x at x=0. It clarifies that a Taylor polynomial cannot be used for e^x/x due to the function having a pole at that point, necessitating a Laurent series instead. The conversation explains that while the Taylor series for e^x can be divided by x, it does not represent the Taylor series of e^x/x. The participants also explore whether similar methods can be applied to other functions, concluding that Taylor polynomials can be used for functions without poles. Overall, the distinction between Taylor and Laurent series is emphasized in the context of function approximation.
RadiantL
Messages
31
Reaction score
0
So I was just wondering why when you approximate using the Taylor Polynomials for something like e^x/x at x = 0 you can just find the approximation for e^x and make it all over x, could you do the same for like e^x/x^2 or e^x/x^3?

I hope my question makes sense... thanks
 
Mathematics news on Phys.org
You can't find a Taylor polynomial approximation of \frac{e^x}{x} at x=0. The function has a pole at 0. Finding a Taylor approximation at 0, requires the function to exist at that point (and be continuous, differentiable, etc. there).

However, you can make a Laurent series approximation. A Laurent series allows terms like \frac{1}{x},\frac{1}{x^2},... in its expansion.

The Laurent series is unique, so if you found one expression of your function in a Laurent series, then you found it. In our case, we can indeed do

\frac{e^x}{x}=\frac{1}{x}+1+\frac{x}{2}+...+\frac{x^n}{(n+1)!}+...

The same thing will work for \frac{e^x}{x^2} or \frac{e^x}{x^3}. But don't call this a Taylor approximation!
 
Oh man, that's funky... my book under integration using the Taylor Polynomials it gives an example

∫e^x/x dx ≈∫ T5(x)/x dx

is that the same thing as the Laurent series?
 
Your book is correct. The T_5(x) is indeed the Taylor series of e^x (because e^x exists at x=0 and is smooth there).

However, it would be incorrect to say that \frac{T_5(x)}{x} is the Taylor series of \frac{e^x}{x}. Here you have to use Laurent series.
 
Ah, so I'm wrong um so what my book did was just take an approximation for a part of that function? Specifically e^x
 
RadiantL said:
Ah, so I'm wrong um so what my book did was just take an approximation for a part of that function? Specifically e^x

Yes, that's exactly what your book did!
 
Intresting, so can I do that for everything else? for example...

x^2+x/x could I just take the Taylor polynomial of x^2+x and then make it over x?
 
RadiantL said:
Intresting, so can I do that for everything else? for example...

x^2+x/x could I just take the Taylor polynomial of x^2+x and then make it over x?

Yes. But the Taylor polynomial of x^2+x will just be x^2+x... (if you take the degree of the polynomial >1).
 
Haha of course, was trying to think up a random example :P Anyway, wow thanks so much your help was very much appreciated. I feel like you should be getting paid for this
 
Back
Top