Determine the set of points $z$ that satisfy the condition $|2z|>|1+z^2|$

I tried to redo this problem and got to this point

$|2z|>|1+z^2|$ $\Rightarrow$ $2|z|>1+|z^2|$ $\Rightarrow$ $2|z|>1+z\overline z$

Let $z=x+iy$ then





since $x,y$ are real number , so there is no solution for this inequality?

  • $\begingroup$ If you write $z = a + bi$, then you can note that your condition is equivalent to $(a^2 + (b-1)^2 - 2)(a^2 + (b+1)^2 - 2) < 0$, and is two circles with their intersection removed. I don't know of any way to see this easily, but this comment might help someone else. $\endgroup$ – George V. Williams Jan 30 '15 at 22:28
  • $\begingroup$ It should be $\cos 2\theta$ and not $\cos^2\theta$. Comparing the $2$ expressions you get $\frac{1}{r^2} + r^2 < 6-4\cos^2\theta$. How explicit do you need to be? I am not sure you can describe it in words... $\endgroup$ – benji Jan 31 '15 at 4:06
  • $\begingroup$ I always have problem with locus problem like this. I tried to explain and justify every step by algebra and calculating but keep getting into dead end. $\endgroup$ – Diane Vanderwaif Jan 31 '15 at 4:25
  • $\begingroup$ @DianeVanderwaif there must be solutions to the inequality, for example $z=i$ $\endgroup$ – benji Feb 4 '15 at 16:59
  • 1
    $\begingroup$ You can use $\Rightarrow$ $\Rightarrow$ or $\implies$ $\implies$. (Whichever you like better.) Of course, you can edit it back to =>, if you prefer it that way. $\endgroup$ – Martin Sleziak Feb 5 '15 at 19:19

Let $z=re^{i\theta}$. Then after squaring both sides we have $$ 4r^2 > r^4 + 2r^2\cos(2\theta) + 1\Rightarrow 0>(r^2-2r\sin(\theta)-1)(r^2+2r\sin(\theta)-1)\tag{1} $$ where I used the following identity $\cos(2\theta) = 1 -2\sin^2(\theta)$. Let's write $(1)$ in Cartesian coordinates so recall that $r^2 = x^2 + y^2$ and $r\sin(\theta) = y$. Then we have $$ 0>(x^2+y^2-2y-1)(x^2+y^2+2y-1)=[x^2+(y-1)^2-2][x^2+(y+1)^2-2] $$ Then \begin{align} x^2+(y-1)^2&< 2\\ x^2+(y+1)^2&< 2 \end{align} That is, we have two disc of radius $\sqrt{2}$ centered at $(0,\pm i)$. So the solution set is $$ S=\bigl\{z\in\mathbb{C}\colon \lvert z-i\rvert < \sqrt{2}\text{ or }\lvert z+i\rvert < \sqrt{2}\text{ but not in the intersection}\bigr\} $$

  • 1
    $\begingroup$ Nice answer. In the final equation, there should be $\sqrt{2}$ and not $2$. $\endgroup$ – mickep Feb 5 '15 at 7:37
  • $\begingroup$ @mickep yes you are correct. I will fix it. $\endgroup$ – dustin Feb 5 '15 at 7:37

let $D = \{z \colon 2|z| < |1 + z^2|\}.$ observe that the region $D$ is symmetric with respect to the transformations $z \to -z, z \to \bar z$ and $z \to \dfrac{1}{z}$ therefore it is enough to worry about the portion of $D$ in the first quadrant.

Let $o = 0, a = 1$ and $z = re^{it}, p = 1 + r^2e^{2it}$ so that $z$ is on the boundary of $D.$ applying rule of cosine to the triangle, we have $$\cos 2t = 2 - \frac{1}{2}\left( r^2 + \frac{1}{r^2} \right).$$

looking at the graph of $y = 2 - \frac{1}{2}(x + 1/x),$ you can verify that $-1 \le y \le 1$ for $3 - 2\sqrt 2 \le x \le 3 + 2\sqrt 2$ and $f(1) = 1, f(2 \pm \sqrt 3) = 0$

i plotted the region $D$ it contains the unit circle and has a hole(lens) containing the origin.

so part of $D$ in the first quadrant is given by $$\{re^{it} \colon t = \frac{1}{2}\cos^{-1}\left( 2 - \frac{1}{2}( r^2 + \frac{1}{r^2} ) \right), \sqrt{3 - 2\sqrt 2} \le r \le \sqrt{3 + 2 \sqrt 2} \}.$$

  • $\begingroup$ so my answer is wrong, I wonder where it's wrong? $\endgroup$ – Diane Vanderwaif Feb 2 '15 at 16:43
  • 1
    $\begingroup$ @DianeVanderwaif, you have $|1 + z^2 | = 1 +|z|^2$ which is not true except for $z = 1.$ $\endgroup$ – abel Feb 2 '15 at 17:03
  • 1
    $\begingroup$ I see, I applied the triangle inequality incorrectly. Thanks $\endgroup$ – Diane Vanderwaif Feb 2 '15 at 17:19
  • $\begingroup$ I'm a little confused . So you let $y=\cos (2t)$ then you use $f(1)$ with $1=r$? Then you let $r^2 \in [3-2\sqrt{2}, 3+2\sqrt{2}]$, when you take the square root,how do you know the sign won't change? because $r$ is always positive? $\endgroup$ – Diane Vanderwaif Feb 2 '15 at 23:31
  • $\begingroup$ and I think you got the inequality sign backward :D $\endgroup$ – Diane Vanderwaif Feb 2 '15 at 23:55

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.