Show a set of maps of the complex plane $x\alpha = ax + b$ is of finite order for given variables $a, b$ where $a \ne 1$ and $a$ is root of unity

72 Views Asked by At

I'm trying some exercises from the Introduction to Group Theory book by Ledermann and Weir and I'm already stuck on an exercise from Chapter 1:

Let $\alpha$ be the map of the complex plane defined by $x\alpha = ax + b$, where $a$ and $b$ are given complex numbers and $a \ne 1$. Obtain a formula for $\alpha^n$ where $n$ is a positive integer, and show that $\alpha$ is of finite order if and only if $a$ is a root of unity.

My complex numbers knowledge is a bit basic so I'm unsure how to start here. I guess if a complex number is not a root of unity, repeatedly applying this $\alpha$ to $x$ would carry on yielding more and more different numbers? But how could I put that into a formula that demonstrates it algebraically?

3

There are 3 best solutions below

3
On BEST ANSWER

The group here is going to be a set of maps acting on the complex plane. The maps in question are of the form $x\mapsto ax+b$. To form a group, the maps all have to have inverses, so we can't allow maps with $a=0$, but otherwise $a$ and $b$ can be any complex numbers.

Important things to check are that the identity map is $x\mapsto x=1x+0$ and that two maps, $x\mapsto a_1x+b_1$, $x\mapsto a_2x+b_2$ are the same map if and only if $a_1=a_2$ and $b_1=b_2$.

To answer your question, you need to determine under what circumstances a group element $\alpha$ given by $x\alpha=ax+b$ has the property that $\alpha^n$ is the identity map, $x\mapsto x=1x+0$. If $a=1$ and $b\ne0$, you should be able to see that this will never happen. So we are left with the case $a\notin\{0,1\}$.

Try iterating the map $\alpha$ a few times, $$ \begin{aligned} x\alpha&=ax+b\\ x\alpha^2&=a^2x+ab+b\\ x\alpha^3&=a^3+a^2b+ab+b, \end{aligned} $$ and you should be able to find the general formula for $x\alpha^n$. Since we are assuming $a\ne1$, you can simplify this using the formula for geometric sums, $1+a+x^2+\ldots+a^{n-1}=\frac{a^n-1}{a-1}$. You should then be able to determine the circumstances under which $x\mapsto x\alpha^n$ is the same map as $x\mapsto x$.

4
On

It is not hard to prove by induction that$$\alpha^n(x)=a^nx+a^{n-1}b+a^{n-2}b+\cdots+b.$$If $a^n=1$ and $a\ne1$, then, since$$0=a^n-1=\overbrace{(a-1)}^{\phantom0\ne0}\left(a^{n-1}+a^{n-2}+\cdots+1\right),$$we have\begin{align}a^nx+a^{n-1}b+a^{n-2}b+\cdots+b&=x+\left(a^{n-1}+a^{n-2}+\cdots+1\right)b\\&=x.\end{align}

0
On

As Jose Carlos Santos mentioned $\alpha^n(x) $ will be of that form replacing $b^i$'s by $b$ .

For the converse part consider the equality, $\alpha ^n(x) =x\ \forall x \in \mathbb C$ so, it's an identity hence $a^n=1$.