2

I'm having trouble with this: As w moves between $-i$ and $i$ on $w = \frac {1+zi}{z+i}$, prove z moves across positive imaginary axis.

I've got the two cartesian equations:

$u = \frac {2x}{x^2 + (y+1)^2}$ and $v = \frac {x^2 + y^2 -1}{x^2 + (y+1)^2}$

and I cannot seem to get to a y>0 if I work from w to z, nor can I seem to get $-1<v<1$ if I move from z towards w. I'm not after the solution but any tips on how to proceed would be greatly appreciated.

Closest I've come is by starting with $v = \frac {y^2 -1}{(y+1)^2}$ (as $x = 0$). Rearranging and solving for y, I get $v < \pm 1$, but I can't seem to get $-1<v<+1$

Thanks.

Noobcoder
  • 227

2 Answers2

1

I believe the method I'm using below is easier than what you're trying to do. First, you can determine $z$ in terms of $w$ as follows:

$$\begin{equation}\begin{aligned} w & = \frac{1 + zi}{z + i} \\ w(z + i) & = 1 + zi \\ wz + wi & = 1 + zi \\ wz - zi & = 1 - wi \\ z(w - i) & = 1 - wi \\ z & = \frac{1 - wi}{w - i} \end{aligned}\end{equation}\tag{1}\label{eq1A}$$

Since you asked for hints, I trust this will do. Nonetheless, just in case you would like to see the rest of the solution, I've placed it below in spoilers.

It asks about $w$ moving between $-i$ and $i$. Thus, you have $w = ki$ for real $k$ where $-1 \lt k \lt 1$.

Substituting this into \eqref{eq1A} gives

$$z = \frac{1 - (ki)i}{ki - i} = \frac{1 + k}{(k - 1)i} = \left(\frac{1 + k}{1 - k}\right)i \tag{2}\label{eq2A}$$

The remainder of the solution is:

Let the coefficient of $i$ be $j = \frac{1 + k}{1 - k}$. As $k$ goes from $-1$ to $1$, note $j$ is always positive, but with it going from just above $0$ when $k$ is close to $-1$ to approach positive infinity as $k$ gets close to $1$. Thus, as asked, this shows $z$ is starting from the origin and moving across the positive imaginary axis.

Update: Using Cartesian coordinates instead, your correctly get to

$$v = \frac{y^2 - 1}{(y+1)^2} \tag{3}\label{eq3A}$$

As you stated, you next need to solve for $y$ in terms of $v$. Doing this gives

$$\begin{equation}\begin{aligned} & v = \frac{y^2 - 1}{y^2 + 2y + 1} \\ & v(y^2 + 2y + 1) = y^2 - 1 \\ & vy^2 + 2vy + v = y^2 - 1 \\ & (v - 1)y^2 + 2vy + (v + 1) = 0 \end{aligned}\end{equation}\tag{4}\label{eq4A}$$

This is a quadratic equation, so using the quadratic formula gives

$$\begin{equation}\begin{aligned} y & = \frac{-2v \pm \sqrt{(2v)^2 - 4(v - 1)(v + 1)}}{2(v - 1)} \\ & = \frac{-2v \pm \sqrt{4v^2 - (4v^2 - 4)}}{2(v - 1)} \\ & = \frac{-v \pm 1}{v - 1} \end{aligned}\end{equation}\tag{5}\label{eq5A}$$

One solution of $y = \frac{-v + 1}{v - 1} = -1$ gives in \eqref{eq3A} that $v = \frac{0}{0}$, so it's not valid. Instead, you have

$$y = \frac{-v - 1}{v - 1} = \frac{1 + v}{1 - v} \tag{6}\label{eq6A}$$

This is the same as \eqref{eq2A}, in the spoiler, with $k$ replaced by $v$.

John Omielan
  • 47,976
  • I came to this too. I got that y tends to +infinity when v tends to 1, but I became too fixated with the denominator and totally missed how the numerator is affected by v close to -1. Thanks for this. – Noobcoder Apr 21 '20 at 11:32
  • is it possible to do this using the cartesian equations? I had to derive it for the first part of the question, so i'm assuming they wanted us to use it. – Noobcoder Apr 21 '20 at 11:36
  • @Noobcoder You're welcome for the help. As for your query, yes, it's possible to also do this using the Cartesian coordinate equations. I've updated my answer to show how to finish what you had in your question text. – John Omielan Apr 21 '20 at 20:14
0

The locus of $z$ is an upward hyperbola existing for $y\ge 1$

We can write $$|w+i|+|w-i|=2~~~(1)$$ Introducing $$w=\frac{1+iz}{z+i}$$ in (1), we get $$\left |\frac{2+2iz}{z+i}\right|+ \left| \frac{2}{z+i}\right|=2 \implies |z+i|-|z-i|=1$$ Which is the upward parabola existing for $y \ge 1$.

Z Ahmed
  • 43,235