Questions concerning analytic and singularities of power series

102 Views Asked by At

In the book 'The Concrete Tetrahedron' by Manuel Kauers and Peter Paule on p30 reads or implies the function $$f: C\backslash \{ ... ,-2\pi,-\pi,0,\pi,2\pi, ...\}\rightarrow C,\, f(z)=\frac z{\sin(z)}$$(where I have used $C$ to mean complex numbers) has a removable singularity at $z\!=\!0$, and the extension $f(0):=1$ turns it into a function that is analytic in a neighborhood of the origin. What exactly is meant by the extension $f(0):=1$ ? Could anyone explain it in different words ? If left off the $f: C\backslash \{ ... ,-2\pi,-\pi,0,\pi,2\pi, ...\}\rightarrow C$ part , that is just plain $\frac z{\sin(z)}$, does not have a singularity at $z\!=\!0$ and has radius of convergence $r\!=\!\pi$ ? Is adding the extension $f(0):=1$ to $$f: C\backslash \{ ... ,-2\pi,-\pi,0,\pi,2\pi, ...\}\rightarrow C,\, f(z)=\frac z{\sin(z)}$$ essentially the same as $\frac z{\sin(z)}$ within $r\!=\!\pi$ ?

On p31 it reads or implies coefficients of the formal power series $$\frac x {\sin(x)}-(\frac \pi {\pi-x}-\frac\pi{-\pi-x} ) $$ are bounded in absolute value by $(\frac1{2\pi}+\eta)^n$ when n is large. In other words $$|a_n-\pi^{-n}+({-\pi})^{-n}|<(\frac1{2\pi}+\eta)^n,\,n\!\rightarrow\!\infty$$ for every $\eta>0$, it follows that $a_{2n}\sim 2\pi^{-2n}$ as $n\!\rightarrow\!\infty.$ Is it a text error and it should have been written $a_{2n}\sim (2\pi)^{-2n}$ instead of $a_{2n}\sim 2\pi^{-2n}$ ? Is that correct ? Next suppose instead it were coefficients of the formal power series $$\frac x {\sin(x)}-(\frac \pi {\pi-x}-\frac\pi{-\pi-x} )-(\frac{2\pi} {2\pi-x}-\frac{2\pi}{-2\pi-x}) $$ then would $a_{2n}\sim (3\pi)^{-2n}$ ?

1

There are 1 best solutions below

3
On

For the extension problem, consider the following simpler case.

Let $g:\mathbb{C}\setminus\{0\}$ be $g(z)=\frac{\sin z}{z}$. Define a function $h:\mathbb{C}\to\mathbb{C}$ with $h(z)=g(z)$ for $z\ne 0$ and $h(0)=1$. Then the function $h$ is called an extension of $g$. Moreoever, $h$ is analytic on $\mathbb{C}$, particularly around $z=0$.

The asymptomatic expression should indeed be $a_{2n}\sim(2\pi)^{-2n}$ since the constant matters by the definition on page 165 of the book.

For your last question, the answer is yes. Because your new function, denoted it as $g$, has removable singularities at $0$, $\pm \pi$, $\pm 2\pi$; $g$ can thus be extended to an analytic function in a disk of radius $3\pi$ around the origin. Now you can follow the same logic for the function on page 31.