Must They All Be Equal To Each Other?

Algebra Level 2

x = 2 z 2 1 + z 2 , y = 2 x 2 1 + x 2 , z = 2 y 2 1 + y 2 \large x=\frac{2z^2}{1+z^2}, \quad y=\frac{2x^2}{1+x^2}, \quad z=\frac{2y^2}{1+y^2}

Excluding the trivial solution ( 0 , 0 , 0 ) (0,0,0) , find the number of triplets of real numbers ( x , y , z ) (x,y,z) that satisfy the system of equations above.

0 1 2 Infinitely Many None Of these choices

This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try refreshing the page, (b) enabling javascript if it is disabled on your browser and, finally, (c) loading the non-javascript version of this page . We're sorry about the hassle.

5 solutions

here is another solution

please download the image and then zoom

This solution makes a common misconception, which is forgetting to check the conditions under which a theorem holds. In order to apply AM-GM, what must be true about the terms?

Do you see how to fix this solution?

Calvin Lin Staff - 4 years, 7 months ago

Log in to reply

All the terms should be non negative and real ?

A Former Brilliant Member - 4 years, 7 months ago

Log in to reply

Right. We don't have that as yet. Can you show why they must be non-negative?

Calvin Lin Staff - 4 years, 7 months ago

Log in to reply

@Calvin Lin yes because x , y ,z are real and square of every real number is positive so 1 and x squared are positive

A Former Brilliant Member - 4 years, 7 months ago

Log in to reply

@A Former Brilliant Member Yes. This is much better. Note that without x , y , z 0 x , y , z \geq 0 , we have other equality conditions like ( 1 , 1 , 1 ) (1, -1, -1 ) .

Calvin Lin Staff - 4 years, 7 months ago

Log in to reply

@Calvin Lin ok thank you calvin sir , could you tell me how to do this problem? https://brilliant.org/problems/rmo-delhi-2016-3/

A Former Brilliant Member - 4 years, 7 months ago

Log in to reply

@A Former Brilliant Member This is awesome!

Laquita Jackson - 4 years, 7 months ago

Log in to reply

@Laquita Jackson thank you mariah

A Former Brilliant Member - 4 years, 7 months ago

@A Former Brilliant Member Whenever you ask for help, do not just leave it at that, which implies you are demanding an answer without putting in any effort

Please explain what you’ve tried (E.g. Titu's lemma) and where you got stuck (The resulting inequality is in the wrong direction). This will allow others to understand what you know/are aware of.

Calvin Lin Staff - 4 years, 7 months ago

Log in to reply

@Calvin Lin i have tried my level best to solve it

like spent a lot of time on it

i didn't get it i tried Titus lemma and some of the other inequalities---(cauchy, AM GM, chebyshev)

the comment is 3 days old

so I don't remember exactly what i did

but i tried to solve it using everything i knew

till date I have solved many inequalities on brilliant

the ones where you have to directly apply a classical inequality are easy

the ones where you can't do this are tough

i am struggling with these inequalities(where you can't apply a classical inequality directly

every problem is different with a unique solution

i am finding it hard to solve these inequalities

because there's something different in every problem

so my question is how do you master these inequality questions

A Former Brilliant Member - 4 years, 7 months ago

This is exactly how I solved. :)

Adeeb Zaman - 4 years, 7 months ago
Shaun Leong
Oct 22, 2016

x = 2 2 1 + z 2 x=2-\frac{2}{1+z^2} y = 2 2 1 + x 2 y=2-\frac{2}{1+x^2} z = 2 2 1 + y 2 z=2-\frac{2}{1+y^2} The minimum value of any of x , y , z x,y,z from the equations above ioccurs when the denominator is minimised, or when they are 0 0 .

WLOG let 0 z y x 2 0 \leq z \leq y \leq x \leq 2 . Note that the other case of 0 z x y 2 0 \leq z \leq x \leq y \leq 2 can be handled similarly. x y z x \geq y \geq z 2 1 + z 2 2 1 + x 2 2 1 + y 2 \Rightarrow -\frac{2}{1+z^2} \geq -\frac{2}{1+x^2} \geq -\frac{2}{1+y^2} 1 + z 2 1 + x 2 1 + y 2 \Rightarrow 1+z^2 \geq 1+x^2 \geq 1+y^2 z y x \Rightarrow z \geq y \geq x The last step followed since x , y , z x,y,z are all positive. Since x z x \geq z and z x z \leq x , we have x = y = z x=y=z x = 2 x 2 1 + x 2 \Rightarrow x=\frac{2x^2}{1+x^2} Excluding the trivial solution, 1 + x 2 = 2 x 1+x^2=2x ( x 1 ) 2 = 0 \Rightarrow (x-1)^2=0 x = y = z = 1 \Rightarrow x=y=z=1 Hence there is only 1 \boxed{1} solution.

Good start! Unfortunately, there are some issues:

  1. Because the equations are not symmetric, we cannot do "WLOG z y x z \leq y \leq x ". If we had z x y z \leq x \leq y , then we cannot relabel the terms (since the expressions are only cyclic) to get the desired z y x z \leq y \leq x ".

  2. You should explain fully why the terms are positive. Currently, the only explanation is that "Without loss of generality, they are bigger than 0", which isn't a valid argument.

Do you see how to fix these issues?

Calvin Lin Staff - 4 years, 7 months ago

Log in to reply

To minimise let's say x x from the first equation we know that z z mist be 0 0 , since z 2 z^2 is always non-negative and appears in the denominator. Same follows for y , z y,z .

I guess for the first concern we could consider both cases?

Shaun Leong - 4 years, 7 months ago

Log in to reply

Right, for the first concern, you just have to explain that "in the other case, we have a similar chain of inequalities, but in the opposite direction". That is the "proper WLOG" to explain.

Calvin Lin Staff - 4 years, 7 months ago
Abhishek Sinha
Oct 28, 2016

Observe that, if any one of the variables is zero, it implies that all other variables are zero. Thus, without any loss of generality, we can focus on solutions in which none of the variables is zero. In this case, we can take reciprocal of the given three equations and sum them up to obtain 1 x + 1 y + 1 z = 3 2 + 1 2 ( 1 x 2 + 1 y 2 + 1 z 2 ) \frac{1}{x}+ \frac{1}{y}+\frac{1}{z}= \frac{3}{2} + \frac{1}{2}( \frac{1}{x^2} + \frac{1}{y^2} + \frac{1}{z^2}) Rearranging the above, we obtain ( 1 x 1 ) 2 + ( 1 y 1 ) 2 + ( 1 z 1 ) 2 = 0 (\frac{1}{x}-1)^2+ (\frac{1}{y}-1)^2+ (\frac{1}{z}-1)^2=0 Which implies that the only other real solution is x = y = z = 1 x=y=z=1

Chew-Seong Cheong
Oct 18, 2016

As the three equations are similar, we assume the solution to be x = y = z = t x=y=z=t . Then we have:

2 t 2 1 + t 2 = t 2 t 2 1 + t 2 t = 0 t ( 2 t 1 + t 2 1 ) = 0 \begin{aligned} \frac {2t^2}{1+t^2} & = t \\ \frac {2t^2}{1+t^2} - t & = 0 \\ t\left(\frac {2t}{1+t^2}-1\right) & = 0 \end{aligned}

{ t = 0 ( x , y , z ) = ( 0 , 0 , 0 ) 2 t 1 t 2 = 0 t = 1 ( x , y , z ) = ( 1 , 1 , 1 ) \implies \begin{cases} t = 0 & \implies (x,y,z) = (0,0,0) \\ 2t - 1 - t^2 = 0 & \implies t = 1 \implies (x,y,z) = (1,1,1) \end{cases}

Therefore, the number of solutions other than (0,0,0) is 1 \boxed{1} .

Can you explain how you concluded that x 1 x \leq 1 ? For example, with z = 2 z = 2 , we have x = 2 × 4 1 + 4 > 1 x = \frac{ 2 \times 4 } { 1 + 4 } > 1\ .

You seem to be saying that if f ( z ) g ( z ) f(z) \leq g(z) and min g ( x ) = K \min g(x) = K , then we must have f ( z ) K g ( z ) f(z) \leq K \leq g(z) . This is not true, say with f ( z ) = z 2 , g ( z ) = z 2 + 1 f(z) = z^2, g(z) = z^2 + 1 .

Calvin Lin Staff - 4 years, 7 months ago

Log in to reply

Yes, guess I was trying to fit in a solution.

Chew-Seong Cheong - 4 years, 7 months ago

I have changed the solution.

Chew-Seong Cheong - 4 years, 7 months ago

Log in to reply

Unfortunately, we cannot just assume that x = y = z = t x = y = z = t . Even though the equations are cyclic, there could be a solution that isn't cyclic.

I believe that there are complex solutions to this system in which all of the terms are distinct (and we should account for that, and also update the question accordingly).

Calvin Lin Staff - 4 years, 7 months ago
James Wilson
Oct 16, 2017

All I did was start by observing that x , y , z 0 x,y,z \geq 0 because the right hand sides of the equations must all be 0 \geq 0 . Now assume all the variables are strictly positive (for obvious reasons). Then I observed that if x = 1 x=1 , then the rest of the variables must equal 1 (and this is an if and only if relationship because of the symmetry involved in the three equations). Now I consider x 1 x\neq 1 , which I write equivalently as ( x 1 ) 2 > 0 (x-1)^2>0 . This, in turn implies x > y x>y (obtained by adding 2 x 2x to both sides of the inequality, dividing by 1 + x 2 1+x^2 , multiplying by x x , and substituting y y for 2 x 2 1 + x 2 \frac{2x^2}{1+x^2} ). However, if x 1 x \neq 1 , then y 1 y\neq 1 , and by similar reasoning, we can conclude y > x y>x . This is a contradiction. Therefore, there are no other solutions.

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...