Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

(Note: I chose a general title, because I believe this discussion will be applicable to all other hyperbolic functions having an asymptote at infinity, but I will specifically be focusing on $\coth(x)$ in the following)

I encounter problems when trying to perform a Taylor expansion of $f(x) = \coth(x)$ at infinity (the term "asymptotic expansion" might be more accurate here).

I'm using the method described in Taylor expansion around infinity of a fraction. I.e. I define a function $$g(x) = f(1/x) = \coth(1/x),$$ Taylor expand $g(x)$ "at" $x = 0$, i.e. $$g(x) \approx T(x)$$ for $x << 1$, where $T(x)$ is (a truncated version of) the Taylor series of $g(x)$. I then introduce the variable substitution $y = 1/x$ and obtain $$g(1/y) = f(y) = \coth(y) \approx T(1/y)$$ for $y >> 1$.

The problem is obtaining $T(x)$. I've tried using Mathematica's "Series"-function, but it won't give me an answer. I then tried to calculate it as $$T(x) = \lim_{x\to 0}g(x) + \frac{\lim_{x\to 0}g'(x)}{1!}(x-0) + \frac{\lim_{x\to 0}g''(x)}{2!}(x-0)^2 + \ldots$$ but while, obviously, $\lim_{x\to 0}g(x) = 1$, it seems that $\lim_{x\to 0}g^{(n)}(x) = 0$ for $n>0$, no matter how high $n$ is (I've tried up to and including $n = 20$ in Mathematica).

So my question is: is the Taylor (asymptotic) expansion of $\coth(x)$ at infinity simply 1 (which would surprise me, since g(x) != 1, and since I've always thought that Taylor expansions of non-polynomial functions contain an infinite number of terms), or have I made a mistake somewhere, in my calculations or my general understanding of the problem?

share|improve this question
1  
Yeah, I don't think this is going to work because the exponential function has an essential singularity at infinity. –  Jason Zimba Mar 20 at 16:52
1  
Does it help for your purposes to write $\coth x = (1+e^{-2x})/(1-e^{-2x}) = (1 + e^{-2x})(1 + e^{-2x} + e^{-4x} + e^{-6x} + \dots)$? –  Jason Zimba Mar 20 at 17:25
    
That helps a lot, thanks! I've expanded the idea behind your comment in an answer below. –  andreasdr Mar 21 at 13:36

1 Answer 1

up vote 0 down vote accepted

I'll expand on Jason Zimbas comment: $$f(x) = \coth(x) = 1 + 2e^{-2x} + 2e^{-4x} + \ldots$$ $$g(x) = 1 + 2e^{-2/x} + 2e^{-4/x} + \ldots$$ $$g'(x) = \frac{4e^{-4/x}(2+e^{2/x})}{x^2} + \ldots$$ $$g''(x) = -\frac{8e^{-4/x}(2(x-2) + e^{2/x} (x-1))}{x^4} + \ldots$$ Both derivatives above go to $0$ when $x \to 0$ (also if higher-order terms are included). It seems that this is true for all higher order derivatives as well (although I don't have a stringent proof). I.e. the Taylor expansion of $f(x)$ at infinity (in the sense described in the question above) is simply 1.

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.