5

Let $z_1,z_2,z_3$ be complex numbers such that $|z_1|=|z_2|=|z_3|=|z_1+z_2+z_3|=2$ and $|z_1–z_2| =|z_1–z_3|$,$(z_2 \ne z_3)$, then the value of $|z_1+z_2||z_1+z_3|$ is_______

My solution is as follow

${z_1} = 2{e^{i{\theta _1}}};{z_2} = 2{e^{i{\theta _2}}};{z_3} = 2{e^{i{\theta _3}}}$ & $Z = {z_1} + {z_2} + {z_3} = 2\left( {{e^{i{\theta _1}}} + {e^{i{\theta _2}}} + {e^{i{\theta _3}}}} \right)$

$\left| {{z_1} - {z_2}} \right| = \left| {{z_1} - {z_3}} \right| \Rightarrow \left| {{e^{i{\theta _1}}} - {e^{i{\theta _2}}}} \right| = \left| {{e^{i{\theta _1}}} - {e^{i{\theta _3}}}} \right|$

Let ${\theta _1} = 0$

$\left| {{z_1} - {z_2}} \right| = \left| {{z_1} - {z_3}} \right| \Rightarrow \left| {1 - \left( {\cos {\theta _2} + i\sin {\theta _2}} \right)} \right| = \left| {1 - \left( {\cos {\theta _3} + i\sin {\theta _3}} \right)} \right|$

$ \Rightarrow \left| {1 - \cos {\theta _2} - i\sin {\theta _2}} \right| = \left| {1 - \cos {\theta _3} - i\sin {\theta _3}} \right| \Rightarrow \left| {2{{\sin }^2}\frac{{{\theta _2}}}{2} - 2i\sin \frac{{{\theta _2}}}{2}\cos \frac{{{\theta _2}}}{2}} \right| = \left| {2{{\sin }^2}\frac{{{\theta _3}}}{2} - 2i\sin \frac{{{\theta _3}}}{2}\cos \frac{{{\theta _3}}}{2}} \right|$

$\Rightarrow \left| { - 2{i^2}{{\sin }^2}\frac{{{\theta _2}}}{2} - 2i\sin \frac{{{\theta _2}}}{2}\cos \frac{{{\theta _2}}}{2}} \right| = \left| { - 2{i^2}{{\sin }^2}\frac{{{\theta _3}}}{2} - 2i\sin \frac{{{\theta _3}}}{2}\cos \frac{{{\theta _3}}}{2}} \right| \Rightarrow \left| { - 2i\sin \frac{{{\theta _2}}}{2}\left( {\cos \frac{{{\theta _2}}}{2} + i\sin \frac{{{\theta _2}}}{2}} \right)} \right| = \left| { - 2i\sin \frac{{{\theta _3}}}{2}\left( {\cos \frac{{{\theta _3}}}{2} + i\sin \frac{{{\theta _3}}}{2}} \right)} \right|$

$ \Rightarrow \left| { - 2i\sin \frac{{{\theta _2}}}{2}\left( {{e^{i\frac{{{\theta _2}}}{2}}}} \right)} \right| = \left| { - 2i\sin \frac{{{\theta _3}}}{2}\left( {{e^{i\frac{{{\theta _3}}}{2}}}} \right)} \right|$

$ \Rightarrow \left| {2\sin \frac{{{\theta _2}}}{2}\left( {{e^{ - i\frac{\pi }{2}}}} \right)\left( {{e^{i\frac{{{\theta _2}}}{2}}}} \right)} \right| = \left| {2\sin \frac{{{\theta _3}}}{2}\left( {{e^{ - i\frac{\pi }{2}}}} \right)\left( {{e^{i\frac{{{\theta _3}}}{2}}}} \right)} \right| \Rightarrow \left| {2\sin \frac{{{\theta _2}}}{2}\left( {{e^{i\left( {\frac{{{\theta _2}}}{2} - \frac{\pi }{2}} \right)}}} \right)} \right| = \left| {2\sin \frac{{{\theta _3}}}{2}\left( {{e^{i\left( {\frac{{{\theta _3}}}{2} - \frac{\pi }{2}} \right)}}} \right)} \right|$

$\Rightarrow \left| {2\sin \frac{{{\theta _2}}}{2}} \right|\left| {\left( {{e^{i\left( {\frac{{{\theta _2}}}{2} - \frac{\pi }{2}} \right)}}} \right)} \right| = \left| {2\sin \frac{{{\theta _3}}}{2}} \right|\left| {\left( {{e^{i\left( {\frac{{{\theta _3}}}{2} - \frac{\pi }{2}} \right)}}} \right)} \right| \Rightarrow \left| {2\sin \frac{{{\theta _2}}}{2}} \right| = \left| {2\sin \frac{{{\theta _3}}}{2}} \right|$

${\theta _2} \ne {\theta _3}$

How do I proceed further?

  • 4
    +1 to your answer, changing its score from $-3$ to $-2$. I am astonished that anyone would not upvote your answer, let alone downvote it. You showed a very good effort in your work. – user2661923 Aug 23 '21 at 06:20
  • @user2661923 I agree, OP constantly asks questions showing high effort – DatBoi Aug 23 '21 at 06:22

6 Answers6

5

Consider $$u=z_2/z_1,v=z_3/z_1$$ and then we have $$|u-1|=|v-1|\tag{1}$$ and $$u\neq v, |u|=|v|=1\tag{2}$$ and $$|1+u+v|=1\tag{3}$$ Then from first equation we get $$(u-1)(\bar{u} - 1)=(v-1)(\bar{v}-1)$$ or $$u+\bar{u}=v+\bar{v}\tag{4}$$ This means that $u, v$ have same real part and using $(2)$ we can see that they are conjugates. If $u=x+iy, v=x-iy$ then from $(3)$ we get $$|1+2x|=1$$ so that either $x=0$ or $x=-1$. But $x=-1$ gives $y=0$ so that $u=v$ which is not allowed.

Thus $u, v=\pm i$ and the expression whose value is to be found is $$|z_1|^2|1+u||1+v|=4|1+i||1-i|=8$$


The key to this problem is to understand that the individual values of $z_1,z_2,z_3$ don't matter, but what matters is their ratios. Also the choice of dividing by $z_1$ to get $u, v$ is because all the constraints in question are symmetric in $z_2,z_3$ and $z_1$ plays a special role as being equidistant from both $z_2,z_3$.

4

We have three points on the circle of radius $2$ centered at the origin. We know that the triangle formed by these points is isosceles. We can rotate the triangle by some angle so that point $z_1$ is on either the real axis or the imaginary axis. Note that the modulus will not change because rotation = multiplying by $e^{iθ}$ for some $θ$. Now the point $z_2,z_3$ are symmetric about the axis you chose; what can you deduce about their sum?

Spoilers

After the rotation $z_1$ is either $\pm 2$ or $\pm 2i$. WLOG assume it's $2$. Then the rotated $z_2$, $z_3$ are $x+iy$ and $x-iy$. We have $|2+2x|=2$ meaning $x=0$. Thus, $z_2=\pm 2i$ and $z_3=\mp 2i$. Thus, $|z_1+z_2||z_1+z_3|=8$.

justadzr
  • 1,478
  • I guess one of us is wrong. It seems to me that $(z_1, z_2, z_3) = (-2, 2i, 2)$ satisfies the constraints, in which case $|z_1 + z_3| = 0.$ – user2661923 Aug 23 '21 at 06:26
  • 1
    @user2661923 $|z_1-z_2|=|z_1-z_3|$, hence you would have $|-2-2i|=|-4|$, but this is wrong. – Jean-Claude Arbaut Aug 23 '21 at 06:28
  • 1
    There is however a mistake: $|2-2i|\cdot|2+2i|=8$. – Jean-Claude Arbaut Aug 23 '21 at 06:31
  • @Jean-ClaudeArbaut Edited. Thanks for the review. – justadzr Aug 23 '21 at 06:33
  • @Jean-ClaudeArbaut You are right, I am going to have to re-consider. I read the question too quickly. I will delete/edit/undelete my answer. – user2661923 Aug 23 '21 at 06:33
  • @user2661923 I have never assumed that $z_1=2$... I deduced that $z_1$, which original has the form $z_1=2e^{i\theta}$, can be rotated to $z_1=2$ by multiplying $e^{-i\theta}$. There are no "uniqueness" concerns because we don't care about the exact values of the points; we only care about their modulus. My answer is a valid proof yet yours contains guessing the values of $z_1, z_2, z_3$... – justadzr Aug 23 '21 at 07:40
  • Yes, you are right. I left you a comment, and then deleted it. There is a uniqueness concern. However, you addressed this concern by explicitly stating that when $z_1$ is rotated to $2i$, that $z_2, z_3$ must have form $x + iy, x - iy$ which does lead to the conclusion that when $z_1 = 2$ that $x$ must equal $0$. I simply didn't read your answer closely enough. In fact, I started to Addendum my answer to complete the analysis, and then realized that I am merely repeating your $x+iy, x-iy$ analysis. – user2661923 Aug 23 '21 at 07:44
3

This is a kludgy metacheating simplification of the problem. From the constraints, all 3 points are on the circle of radius 2, centered at the origin.

Further, directly from the constraints, $z_1$ is on the perpendicular bisector of the line segment connecting $z_2$ and $z_3$. Finally, $z_2, z_3$ must be chosen so that $|z_1 + z_2 + z_3|$ is also on the same circle. This suggests that the vectors $z_2, z_3$ should cancel each other out, which suggests that $z_2 + z_3$ must equal $0$.

Initially, I considered $(z_2, z_1, z_3) = (-2, 2i, 2)$ as one obvious way of satisfying these constraints. Then, I realized, based on someone else's (now deleted) comment, that this answer could be harmlessly rotated by any angle $\theta$.

Although I can't prove it, it seems probable to me that the only way to satisfy these constraints is by some rotation of either $(z_2, z_1, z_3) = (-2, 2i, 2)$ or $(z_2, z_1, z_3) = (-2, -2i, 2).$

This means that (for example) the triangle formed by the vertices $(0,0), z_1, z_2$ must be a 45-45-90 right triangle, whose hypotenuse is $|z_1 + z_2| = 2\sqrt{2}$.

Consideration between $z_1$ and $z_3$ is identical.

Therefore, the answer is $\left[2\sqrt{2}\right]^2 = 8.$


Note
A case may be made that this answer is incomplete, since I did not actually prove that there was no other way to satisfy the constraints. I am therefore relying heavily on the fact that the tone of the question indicates that the answer must be unique.


Addendum
Completing the problem, by proving (for example) that when $z_1 = 2i$ that $z_2, z_3$ must be on the real axis.

I started to analyze this, and then realized that I am merely repeating the analysis already supplied in the answer given by justadzr.

That is, any given $(z_2, z_1, z_3)$ that satisfies the constraints continues to satisfy the constraints when all three points are rotated by $\theta$. When such a rotation takes $z_1$ to $2i$, then, very similar to what justadzr's answer indicates, $z_2, z_3$ must have form $x + iy, -x +iy$, which subsequently implies that $y = 0$.

This places $z_2, z_3$ on the real axis, which places them at $-2, +2$ (in some order), since these are the only two places where the circle intersects the real axis.

user2661923
  • 35,619
  • 3
  • 17
  • 39
  • In this special case $|z_1-z_{2/3}|$ is indeed equal to $|z_1+z_{2/3}|$. But the hypotenuse you are referring to is not $|z_1+z_{2/3}|$ but the former modulus. – justadzr Aug 23 '21 at 06:55
  • @justadzr Simplifying to the case of $(z_2, z_1) = (-2, 2i)$, then the hypotenuse is represented by the line segment from $(0,0)$ to $(z_1 + z_2) = (-2, 2i)$ which does form a 45-45-90 triangle. Further, rotating all 3 of the vertices by $\theta$ does not alter this. Therefore, I am confused by your comment. – user2661923 Aug 23 '21 at 07:00
  • The hypotenuse of the triangle with vertices $0, z_1, z_2$ is the segment from $z_1$ to $z_2$. Yet this value can be computed by $|z_1+z_2|$, but this is only true if you know the values of $z_1, z_2, z_3$, in which case your arguments on the triangle is redundant. – justadzr Aug 23 '21 at 07:07
  • @justadzr Here, I disagree. You don't need to know the explicit values of $z_2, z_1, z_3.$ Instead, all you need to know is that $(0,0), z_1, z_2$ form a 45-45-90 right triangle. As I said, this is true in the basic $(z_2, z_1) = (-2, 2i)$ configuration, and must remain unchanged by any rotation. – user2661923 Aug 23 '21 at 07:12
  • Knowing the value of $z_1, z_2, z_3$ after the rotation gives you that $0, z_1, z_2$ forms such a triangle. But you don't need this triangle to compute the result - the fact that the modulus does not change under rotations is simply about the points $z_1, z_2, z_3$, not the triangle. One could immediately get $|z_1+z_3|$ if you assume they are $-2, 2i, 2$ after a rotation. – justadzr Aug 23 '21 at 07:21
  • @justadzr Very good point, which I totally overlooked. I was focusing on rotation not changing the angles, while you were focusing on rotation not changing the length of the diagonal to the parallelogram (i.e. the diagonal whose length is represented by $|z_1 + z_3|$. – user2661923 Aug 23 '21 at 07:31
3

Algebra + Geometry approach

Let $z_1,z_2,z_3$ be complex numbers such that $|z_1|=|z_2|=|z_3|=|z_1+z_2+z_3|=2$ and $|z_1–z_2| =|z_1–z_3|$,$(z_2 \ne z_3)$, then the value of $|z_1+z_2||z_1+z_3|$ is_______

Let's try interpret the problem geometrically.

$|z_1| = |z_2 | = |z_3|$ means that the three points lie on a circle.

$|z_1 - z_2 | = |z_1 - z_3|$ means that the chord substended by $z_1$ and $z_2$ is equal to the chord subtended by $z_1$ and $z_3$. This means that $z_1$ is the mid point of the arc on the circle starting at $z_3 $ and ending at $z_2$.

It is also given that $|z_1 + z_2 + z_3 |= 2$, we can rearrange this to give: $|1 + \frac{z_2}{z_1} +\frac{z_3}{z_1} | = 1$ but we know that $\frac{z_2}{z_1} = e^{i \theta}$ and $\frac{z_3}{z_1} = e^{- i \theta}$, this leads to:

$$| 1+ 2 \cos \theta | = 1$$

It is clear the the only two principle solutions for $\theta$ are $\theta = \{\pi, \frac{\pi}{2} \}$... but if $\theta=\pi$ then $z_2 = z_3$ violating the condition in question (draw a picture!)

We go back to the question : $|z_1 + z_2 | |z_1 + z_3| = \frac{|z_1^2 - z_2^2| | z_1^2 - z_3^2|}{|z_1 - z_2| |z_1 - z_3|}$

We know when we square a complex number, we square the magnitude and double the angle. Hence, $z_1^2 - z_2^2$ represents a chord having twice the angle as $z_1 - z_2$ and on a circle with squared the radius and similarly does $|z_1 - z_3|^2$.

Ok, but what is the expression for the length of chord which subtends an angle $\theta$ in a circle of radius $r$? By the law of cosines, I find it as $ 2r \sin \frac{\theta}{2}$... now we begin the bash:

$$|z_1 + z_2 | |z_1 + z_3| = \frac{|z_1^2 - z_2^2| | z_1^2 - z_3^2|}{|z_1 - z_2| |z_1 - z_3|} = \frac{ (2r^2 \sin \theta)^2}{(2r \sin \frac{\theta}{2})^2}=4r^2 \cos^2 \frac{\theta}{2}$$

We know $r=2$ and $ \theta= \frac{\pi}{2}$, which gives the final answer as $8$

Kevin.S
  • 3,706
0

WLOG, it can be assumed that $z_1$ is on Imaginary axis(or y-axis).

Since $\left|\frac{z_1+z_2+z_3}{3}\right|=\frac{2}{3}$

It means that distance of the Centroid of triangle formed by joining $z_1,z_2,z_3$ from the centre of circle is $\frac{2}{3}$ and consequently distance of centroid from $z_1$ is $\frac{4}{3}$.

So perpendicular distance of $z_1$ from line segment joining $z_2$ and $z_3$ is

$\frac{1}{2}(\frac{4}{3})+\frac{4}{3}=2$

(since centroid divides median in ratio 2:1)

which is radius of the circle on which $z_1,z_2,z_3$ lie.

Thus $z_2$ and $z_3$ lie on the x-axis and are thus end-points of the diameter of the circle.

So clearly $z_2=-z_3$ and $z_3=-z_2$.

So, $|z_1+z_2||z_1+z_3|=|z_1-z_3||z_1-z_2|=(2\sqrt{2})(2\sqrt{2})=8$

Maverick
  • 9,172
  • 2
  • 30
  • 61
0

Let $z_1=2\cdot e^{i \alpha}, z_2=2\cdot e^{i \beta}, z_3=2\cdot e^{i \gamma}$

and It is given that $z_2\neq z_3$ so we must not get $\beta=\gamma+2k\pi$

$z_1+z_2+z_3=2(e^{i\alpha}+e^{i\beta}+e^{i\gamma})\implies |z_1+z_2+z_3|=2|e^{i\alpha}+e^{i\beta}+e^{i\gamma}|$

Hence $|e^{i\alpha}+e^{i\beta}+e^{i\gamma}|=1\implies |(\cos\alpha+\cos\beta+\cos\gamma)+i(\sin \alpha+\sin\beta+\sin\gamma)|=1\implies 3+2 \cos(\alpha-\beta)+2\cos(\beta-\gamma)+2\cos(\gamma-\alpha)=1$

Hence $$\cos(\alpha-\beta)+\cos(\beta-\gamma)+\cos(\gamma-\alpha)=-1.........(1)$$

From $|z_1-z_3|=|z_1-z_2|\implies |2e^{i\alpha}-2e^{i\gamma}|=|2e^{i\alpha}-2e^{i\beta}|\implies |(\cos\alpha-\cos\gamma)+i(\sin\alpha-\sin\gamma)|=|(\cos\alpha-\cos\beta)+i(\sin\alpha-\sin\beta)|$

$\implies 2-2\cos(\alpha-\gamma)=2-2\cos(\alpha-\beta)$

$\implies \cos(\alpha-\gamma)=\cos(\alpha-\beta).......(2)$

$\implies \alpha-\gamma=\beta-\alpha \implies2\alpha=\beta+\gamma..........(3)$

Now using $(2), (3)$ Equation $(1)$ turns to

$\cos(\alpha-\beta)+\cos(2\alpha-\gamma-\gamma)+\cos(\alpha-\beta)=-1.......(4)$

And using the $\alpha-\gamma=\beta-\alpha$ again in Equation $(4)$ turns to

$2\cos(\alpha-\beta)+\ \cos(2(\beta-\alpha))=-1\implies \cos(\alpha-\beta)=0$ or $-1$

If $\cos(\alpha-\beta)=-1$ then $\alpha-\beta=(2k+1)\pi\;$ and from $(3)$ we get $\beta=\gamma+2k\pi\;$

So $\cos(\alpha-\beta)$ must be $=0$

Now According to Question

$|z_1+z_2||z_1+z_3|\implies |z_1^2+z_1z_2+z_1 z_3+z_2z_3|$

$\implies \;4|e^{i2\alpha}+e^{i(\alpha+\beta)}+e^{i(\alpha+\gamma)}+e^{i(\beta+\gamma)}|$

$\implies$ $4|e^{2i\alpha}+\cos(\alpha+\beta)+i\sin(\alpha+\beta)+\cos(\alpha+\gamma)+i\sin(\alpha+\gamma)+e^{2i\alpha}|$

$\implies 4\bigg|2\cdot e^{2i\alpha}+2\cos\bigg(\dfrac{2\alpha+\beta+\gamma}{2}\bigg)\cos\bigg(\dfrac{\beta-\gamma}{2}\bigg)+i\cdot 2\sin\bigg(\dfrac{2\alpha+\beta+\gamma}{2}\bigg)\cos\bigg(\dfrac{\beta-\gamma}{2}\bigg)\bigg|$

$\implies$ $4\bigg|2\cdot (\cos2\alpha+i\sin2\beta)+2\cos\bigg(\dfrac{\beta-\gamma}{2}\bigg)(\cos2\alpha+i\sin2\beta)\bigg|$

$\implies 8\bigg|\bigg(\cos2\alpha+i\sin2\beta\bigg)\bigg(1+\cos\bigg(\dfrac{\beta-\gamma}{2}\bigg)\bigg|$

$\implies 8\bigg| \bigg(1+\cos\bigg(\dfrac{\beta-\gamma}{2}\bigg) \bigg |$

$\implies8\bigg| \bigg(1+\cos(\alpha-\beta)\bigg) \bigg |=8$

Note:

$e^{i\theta}=\cos(\theta)+i\sin(\theta)$

$z_{1}^2=4\cdot e^{2i\alpha},\;z_{1}z_{3}=4\cdot e^{i(\alpha+\gamma)},\;z_{1}z_{2}=4\cdot e^{i(\alpha+\beta)},\;z_{3}z_{2}=4\cdot e^{i(\gamma+\beta)}$

$\cos \alpha +\cos \beta=2\cos\bigg(\dfrac{\alpha+\beta}{2}\bigg)\cos\bigg(\dfrac{\alpha-\beta}{2}\bigg)$

$\sin \alpha +\sin \beta=2\sin\bigg(\dfrac{\alpha+\beta}{2}\bigg)\cos\bigg(\dfrac{\alpha-\beta}{2}\bigg)$

$\beta-\gamma=2(\alpha-\beta)$ using Equation $(3)$

mathophile
  • 3,225