Let $$z^3+bz^2+cz+d=0$$ be a cubic equation with complex coefficients. Suppose $z_1, z_2$ and $z_3$ are its roots.
I need to find a condition on $b,c,d$ so that $$|z_1|=|z_2|.$$ How can I find such a relationship? Any help would be appreciate. Thank you.

  • 2
    $\begingroup$ You could try reverse engineering: start with a factored cubic, and assuming two roots have the same modulus, see what happens when you multiply (note that complex conjugates have the same modulus, which certainly gives a large class of the cubics in question). $\endgroup$ – pjs36 Jan 19 '16 at 6:49
  • $\begingroup$ Any thoughts on the answers that have been posted? $\endgroup$ – Gerry Myerson Jan 21 '16 at 11:41
  • $\begingroup$ Earth to Nilan, come in please. $\endgroup$ – Gerry Myerson Jan 23 '16 at 4:25
  • $\begingroup$ @GerryMyerson: :) $\endgroup$ – Bumblebee Jan 28 '16 at 3:41

If $|z_1| = |z_2|$, you can write $z_1 = u e^{i\theta}$ and $z_2 = u e^{-i\theta}$ for some real $\theta$ and complex $u$. The cubic is then $$(z - u e^{i\theta})(z - u e^{-i\theta})(z - z_3)= (z^2 - 2 u \cos(\theta) + u^2)(z - z_3)$$ We then get $$ \eqalign{b &= -2 u \cos(\theta) - z_3\cr c &= 2 u z_3 \cos(\theta) +u^2\cr d &= -u^2 z_3\cr} $$

EDIT: We can eliminate the complex parameters $u$ and $z_3$ to obtain an equation involving $b,c,d$ and $\cos(\theta)$: $$ 64\,{d}^{2} \cos^6 \left( \theta \right) -16\,d \left( bc+3\,d \right) \cos^4 \left( \theta \right) + 4\, \left( {b}^{3}d-bcd+{c}^{3}+3\,{d}^{2} \right) \cos^2 \left( \theta \right) -(bc - d)^2 = 0 $$ If we write $\cos^2(\theta) = v$, that is a cubic in $v$, and a condition (maybe the only one needed) on $b,c,d$ is that this cubic has a solution in the real interval $[0,1]$.

  • 1
    $\begingroup$ Shouldn't it be $2u\cos(\theta)z$? $\endgroup$ – amcerbu Jan 19 '16 at 7:31
  • 1
    $\begingroup$ You have only considered the case when $z_1$ is the conjugate of $z_2$. What about the case of repeated roots, i.e., $z_1 = z_2$ $\endgroup$ – user297008 Jan 19 '16 at 7:50
  • $\begingroup$ @Tanuj, that's $\theta=0$. $\endgroup$ – Gerry Myerson Jan 19 '16 at 8:37
  • 1
    $\begingroup$ @GerryMyerson The problem is that you have defined it such that a) $b,c,d\in\mathbb{R}$, and b) $z_1=-z_2$, which is a direct result of your assumption that $z_1$ and $z_2$ have opposite $\theta$'s. $\endgroup$ – Simply Beautiful Art Jan 19 '16 at 13:16
  • 2
    $\begingroup$ @Simple, I can't speak for OP, but we can take $u$ to be either one of the solutions to $u^2=z_1z_2$, and then get $\theta$ from $e^{i\theta}=z_1/u$. Or, we can first get $\theta$ from $z_1/z_2=e^{2i\theta}$, and then get $u$ from $u=z_2e^{i\theta}$. $\endgroup$ – Gerry Myerson Jan 19 '16 at 22:05

In comparison to Robert Israel's answer, I note the following:

$$|ue^{\alpha\cdot i}|=u$$

for $\alpha\in\mathbb{R}$ and $u>0$.

Generally, Euler form of a complex number has the magnitude of the complex number as a coefficient and the real angle multiplied by $i$ with base $e$, which allows us to encompass all possible complex numbers with a slight exception of maybe $0+0i$.

Now, we can define our problem such that we have:

$$z_1=ue^{xi}, z_2=ue^{yi}$$

So now we have:


And you could multiply them all out if you want.

Another answer goes along the line of finding the root of a cubic polynomial, which may be found on Wolfram|Alpha.

Take the absolute value of the thing and try to solve using definitions of the absolute value operation:



$$|z|=z, z\ge0$$



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.