1
$\begingroup$

I tried to subtract two complex numbers in polar form without transforming them into the cartesian form. Therefore I used the approach made by Mark Viola in the following link.

Adding two polar vectors

I managed to get the following result.

$$e^{i(\phi-\phi_1)}=\frac{r_1-r_2e^{i(\phi_2-\phi_1)}}{\sqrt{r_1^2+r_2^2-2r_1r_2\cos (\phi_2-\phi_1)}} \tag 1$$

At this point I do not know how to achieve the final equation like mentioned in the link. I post the final equation for adding two complex numbers in polar form:

$${\phi=\phi_1+\operatorname{arctan2}\left(r_2\sin(\phi_2-\phi_1),r_1+r_2\cos(\phi_2-\phi_1)\right)} \tag 2$$

Any help or hints which leading to the final equation are appreciated.

$\endgroup$
  • 1
    $\begingroup$ Hint: to compute $z_1 - z_2$, you can compute $z_1 + (-z_2)$. So if you can negate $z_2$ (in polar form), then you can apply the addition formula from the other page to get the result you want. Hint: to negate $(r, \theta)$, try $(r, \theta + \pi)$. $\endgroup$ – John Hughes Jul 23 '18 at 16:26
  • $\begingroup$ Thanks for the hint and the quick answer, I will try your approach. $\endgroup$ – Pavel Jul 25 '18 at 19:12
1
$\begingroup$

In fact, you can't avoid the conversion from polar to Cartesian and back to polar, even if done in a single go (any expression you will find will be of a comparable complexity).

The combined equations are

$$r^2=(r_1\cos\phi_1-r_2\cos\phi_2)^2+(r_1\sin\phi_1-r_2\sin\phi_2)^2 \\=r_1^2-2r_1r_2\cos(\phi_1-\phi_2)+r_2^2$$

and

$$\tan\phi=\frac{r_1\sin\phi_1-r_2\sin\phi_2}{r_1\cos\phi_1-r_2\cos\phi_2}.$$

I don't think there is a nice simplification of the expression of the argument. And computationally, the simplification of the modulus is not so attractive as it involves an additional evaluation of a circular function.

$\endgroup$
  • $\begingroup$ Thank you for the quick answer. I get the point with the transformation. Maybe I can rearrange the expression when putting the $\phi_2$ and $\phi_1$ in relation. For example when both phase values are quite close to each other. Thanks again $\endgroup$ – Pavel Jul 25 '18 at 19:10

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.