0
$\begingroup$

Given a non-zero complex number $z$, write $\mathrm{arg}(z)$ for the prinicipal argument of $z$, defined as the unique $\theta \in (-\pi,\pi]$ for which there exists $r \in \mathbb{R}_{>0}$ such that $z = r e^{i\theta}.$

Then for all non-zero $z$, we have: $$\tan(\mathrm{arg}(z)) = \frac{\Im z}{\Re z}$$

Unfortunately, the "inverse" formula $$\mathrm{arg}(z)=\mathrm{tan}^{-1}\left(\frac{\Im z}{\Re z}\right)$$ only works for $\Re z \geq 0,$ due to injectivity issues.

We can fix this by introducing a correction factor. Define a function $$f : \mathbb{C}_{\neq 0} \rightarrow \mathbb{R}$$ by writing

$$f(z) = \begin{cases}\pi & \Re z < 0 \,\&\,\Im z \geq 0 \\ -\pi &\Re z < 0 \,\&\,\Im z < 0 \\ 0\end{cases}$$

Then if I'm not mistaken, we have: $$\forall(z \in \mathbb{C}_{\neq 0})\qquad \mathrm{arg}(z) = \mathrm{tan}^{-1}\left(\frac{\Im z}{\Re z}\right) + f(z).$$

More generally, for any half-open interval $I \subseteq \mathbb{R}$ of length $2\pi$, there should be a corresponding function $$f_I : \mathbb{C}_{\neq 0} \rightarrow \mathbb{R}$$ such that $$\forall(z \in \mathbb{C}_{\neq 0}) \qquad \mathrm{arg}_I(z) = \mathrm{tan}^{-1}\left(\frac{\Im z}{\Re z}\right)+f_I(z).$$

Question.

Q0. Is there a name for $f_I$, in either the special case given above corresponding to $I = (-\pi,\pi]$, or else in the general case?

Q1. What's the explicit formula for $f_I(z)$, for arbitrary $I = (a,b]$ where $b-a =2\pi$?

$\endgroup$
  • $\begingroup$ In computation related contexts atan2(y, x) is often available for that precise reason. $\endgroup$ – dxiv Aug 8 '17 at 3:42
  • $\begingroup$ @dxiv, that's a reasonable solution. I'd still like an answer to this question though, because I have a policy of doing things more-or-less as the book I'm reading is doing it, with small corrections like this to make things better wherever possible. That is, I don't bother completely rethinking everything until I'm pretty much done with whatever I'm reading. $\endgroup$ – goblin Aug 8 '17 at 3:44
1
$\begingroup$

I've always liked this one, because of its flexibility (overlapping half-planes):

$$ \operatorname{Arg}(x + iy) = \operatorname{atan2}(y,\, x) = \begin{cases} \arctan\left(\frac{y}{x}\right) &\text{if } x > 0, \\ \frac{\pi}{2} - \arctan\left(\frac{x}{y}\right) &\text{if } y > 0, \\ -\frac{\pi}{2} - \arctan\left(\frac{x}{y}\right) &\text{if } y < 0, \\ \arctan\left(\frac{y}{x}\right) \pm \pi &\text{if } x < 0, \\ \text{undefined} &\text{if } x = 0 \text{ and } y = 0. \end{cases} $$

another one that is short is

$$ \operatorname{Arg}(x + iy) = \operatorname{sgn}(y) \, \arccos\left( \frac{x}{\sqrt{x^2 + y^2}} \right) $$

$\endgroup$
0
$\begingroup$

The answer to $Q0$ is $\arg(z) = 2\tan^{-1}(y/(|z|+x))$ if $z=x+yi$ with no correction needed.

The advantage of this formula is that it only uses $\tan^{-1}$ and $|z|$ and works for all $z\neq 0$.

The answer to $Q1$ is $\:\arg(zc)/c\:$ for some $c\in C$ with $|c|=1$.

$\endgroup$
  • 1
    $\begingroup$ Shouldn’t this formula be homogeneous? $\endgroup$ – Lubin Aug 8 '17 at 3:19
  • $\begingroup$ @Lubin Thanks, of course, I forgot about that. $\endgroup$ – Somos Aug 8 '17 at 3:23

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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