9 times around

Algebra Level 5

Find the number of distinct real solutions of the system of equations { x z z x = 2 3 y x 9 x y = 2 z y 9 y z = 6. \begin{cases} \frac{x}{z}-zx=\frac {2}{3}\\ \frac{y}{x}-9xy=2\\ \frac{z}{y}-9yz=6.\\ \end{cases}


The answer is 6.

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.

4 solutions

Thomas Beuman
Aug 15, 2013

First, let us apply the following change of variables:

a = 3 z x b = 9 x y c = 3 y z a = 3zx \qquad b = 9xy \qquad c = 3yz

The system of equations then transforms into:

b c a = c a b = a b c = 2 \frac{b}{c} - a = \frac{c}{a} - b = \frac{a}{b} - c = 2

We can already see a solution: a = b = c = 1 a=b=c=-1 . However, this corresponds to the two imaginary solutions x = y = ± i 3 x = y = \pm\frac{i}3 , z = ± i z = \pm i . In order for a solution ( a , b , c ) (a,b,c) to give a real solution ( x , y , z ) (x,y,z) , we need to have a b c > 0 abc > 0 . Also note that any solution ( a , b , c ) (a,b,c) corresponds to two solutions: if ( x , y , z ) (x,y,z) is one, then ( x , y , z ) (-x,-y,-z) is the other.

We rewrite the equations:

  1. b = c ( a + 2 ) b = c(a+2)
  2. c = a ( b + 2 ) c = a(b+2)
  3. a = b ( c + 2 ) a = b(c+2)

Entering (1) into (2) gives:

c = a ( c ( a + 2 ) + 2 ) c = a(c(a+2)+2)

Solving for c c :

c = 2 a 1 2 a a 2 c = \frac{2a}{1-2a-a^2}

And thus:

b = 2 a ( a + 2 ) 1 2 a a 2 b = \frac{2a(a+2)}{1-2a-a^2}

Putting the expressions for b b and c c into (3) leads to:

a ( 1 2 a a 2 ) 2 = 2 a ( a + 2 ) ( 2 a + 2 ( 1 2 a a 2 ) ) a(1-2a-a^2)^2 = 2a(a+2)(2a + 2(1-2a-a^2))

This can be expanded and rearranged to:

0 = a ( a 4 + 8 a 3 + 14 a 2 7 ) 0 = a(a^4+8a^3+14a^2-7)

Obviously a = 0 a=0 is no good, and we know that a = 1 a=-1 is not a solution we're interested in, so we can divide by a a and a + 1 a+1 :

0 = a 3 + 7 a 2 + 7 a 7 = f ( a ) 0 = a^3+7a^2+7a-7 = f(a)

This has no rational roots, but from noting that f ( 1000 ) < 0 f(-1000) < 0 , f ( 4 ) > 0 f(-4) > 0 , f ( 0 ) < 0 f(0) < 0 and f ( 1000 ) > 0 f(1000) > 0 we can deduce that there are three real roots, two of them negative and one positive. From the symmetry in a a , b b and c c , it follows that b b and c c must also each be equal to one of these roots. Furthermore, no two of them can be equal, since e.g. a = b a=b implies c = 1 c=-1 , which does not work. Therefore, for all three possible values of a a , the values of b b and c c are precisely the two other roots, and they obey a b c > 0 abc > 0 . We thus have three relevant solutions for ( a , b , c ) (a,b,c) , and thus 6 \boxed{6} real solutions for ( x , y , z ) (x,y,z) .

Moderator note:

This is a correct solution, and well written too! Hint for a more elegant one: tan(2t)=...

In response to the Challenge Master:

I see now! The funny thing is that I did have a good look into trying to apply a substitution featuring trigonometric or hyperbolic functions, but only after I had made my substitution to a a , b b and c c , because that felt so "natural".

Until someone posts an "official" solution, let me give one:

We can rewrite the equations to:

3 x = 2 z 1 z 2 3 y = 2 ( 3 x ) 1 ( 3 x ) 2 z = 2 ( 3 y ) 1 ( 3 y ) 2 3x = \frac{2z}{1-z^2} \\ 3y = \frac{2(3x)}{1-(3x)^2} \\ z = \frac{2(3y)}{1-(3y)^2}

Note that this has introduced the invalid solution x = y = z = 0 x=y=z=0 . Substitution time:

3 x = tan α 3 y = tan β z = tan γ 3x = \tan \alpha \qquad 3y = \tan \beta \qquad z = \tan \gamma

This gives:

tan α = tan 2 γ α 2 γ (mod π ) tan β = tan 2 α β 2 α (mod π ) tan γ = tan 2 β γ 2 β (mod π ) \tan \alpha = \tan 2\gamma \iff \alpha \equiv 2\gamma \text{ (mod } \pi \text{)} \\ \tan \beta = \tan 2\alpha \iff \beta \equiv 2\alpha \text{ (mod } \pi \text{)} \\ \tan \gamma = \tan 2\beta \iff \gamma \equiv 2\beta \text{ (mod } \pi \text{)}

Combining all three:

α 8 α (mod π ) α = k π 7 , k Z \alpha \equiv 8\alpha \text{ (mod } \pi \text{)} \iff \alpha = \frac{k\pi}7, \ k \in \mathbb{Z}

Note that these solutions provide only 7 different solutions for x = 1 3 tan α x = \frac13 \tan \alpha , e.g. for 0 k 6 0 \leq k \leq 6 , and that k = 0 k=0 yields the forbidden solution x = 0 x=0 . This leaves 6 solutions, which can all be easily verified to give genuine solutions for ( x , y , z ) (x,y,z) .

Thomas Beuman - 7 years, 10 months ago

Log in to reply

Perfect!

Alexander Borisov - 7 years, 10 months ago

Not a particularly elegant solution I admit, but at the time of writing this is the only one, so it's better than nothing. I've tried to think of something smart, even after solving it, but I failed (so far).

Thomas Beuman - 7 years, 10 months ago
Leo Lai
May 20, 2014

Let r = 3 x , s = 3 y , t = z r=3x,\ s=3y,\ t=z , then the original system becomes { r t r t = 2 s r s r = 2 t s t s = 2 \left\{\begin{array}{l} &\frac{r}{t}-rt=2\\ &\frac{s}{r}-sr=2\\ &\frac{t}{s}-ts=2 \end{array}\right. Now, let t = tan θ t=\tan\theta , where θ [ 0 , π ) \theta\in[0,\pi) . Substitute into the first equation and simplify to get r = 2 t 1 t 2 = 2 tan θ 1 tan 2 θ = tan 2 θ r=\frac{2t}{1-t^2}=\frac{2\tan\theta}{1-\tan^2\theta}=\tan2\theta Similarly, s = tan 4 θ s=\tan4\theta and t = tan 8 θ t=\tan8\theta . Now, tan 8 θ = tan θ sin 8 θ cos θ = cos 8 θ sin θ sin 7 θ = 0 \tan8\theta=\tan\theta\Leftrightarrow \sin8\theta\cos\theta=\cos8\theta\sin\theta\Leftrightarrow\sin7\theta=0 The equation has 7 solutions, but the solution θ = 0 \theta=0 gives r = s = t = 0 r=s=t=0 , which makes the equations undefined, so there are 6 solutions.

"The equation has 7 solutions" is not quite correct. Probably meant 7 solutions from -\pi/2 to \pi/2. Otherwise, a correct solution

Calvin Lin Staff - 7 years ago
Sambit Senapati
May 20, 2014

Let 3 x = tan A , 3 y = tan B , z = t a n C 3x=\tan A, 3y=\tan B, z=\ tanC

From the 1st, 2nd and 3rd equations respectively we get the following 3 equations:

tan A = tan 2 C ( 1 ) tan B = tan 2 A ( 2 ) tan C = tan 2 B ( 3 ) \begin{aligned} \tan A & = \tan2C &-(1)\\ \tan B &= \tan 2A & -(2)\\ \tan C& =\tan 2B & -(3)\\ \end{aligned}

So, A = n π + 2 C n I A=n\pi+2C \quad \forall n \in \mathbb{I} -(4)

B = m π + 2 n π + 4 C m I B=m\pi+2n\pi+4C \quad \forall m \in \mathbb{I} -(5)

C = k π + ( 2 m + 4 n ) π + 8 C k I C=k\pi+(2m+4n)\pi+8C \quad \forall k \in \mathbb{I}

C = ( k + 2 m + 4 n ) π 7 \implies C=\frac{-(k+2m+4n)\pi}{7}

Substituting the value of C obtained above in (4) and (5) we get,

B = ( 4 k + m + 2 n ) π 7 B=\frac{-(4k+m+2n)\pi}{7}

A = ( 2 k + 4 m + n ) π 7 A=\frac{-(2k+4m+n)\pi}{7}

We set A , B , C ( 0 , π ) A,B,C \in (0,\pi) . We are allowed to do so since all values of x x can be represented by using tan θ \tan \theta for θ ( 0 , π ) \theta \in(0,\pi)

Suppose we have got a solution for x. Observe (2). As value of x is fixed the value of y is also fixed(as (2) becomes a linear equation in y). Now observe (3). As the value of y is fixed the value of z is fixed (same reason as the previous one). So, we have now proved that for a particular x we get a unique y and for a particular pair of x and y we get a unique z. -(6)

This gives us 7 solutions, of the form x = tan k π 7 x = \tan \frac {k \pi}{7} for k = 0 k=0 to 6. However, we must exclude the case when k = 0 k=0 , as that leads to x = 0 x=0 . Hence, there are 6 solutions in all.

At the start, you should start that x , y , z x, y, z are non-zero from the conditions, which allows us to multiply throughout. Otherwise, it is not true that x z 2 x = 2 z 3 x z z x = 2 3 x - z^2x = \frac {2z}{3} \Leftrightarrow \frac {x}{z} - zx = \frac {2}{3} .

This solution can be simplified slightly by working modulo π \pi , to show that 8 A 4 B 2 C A ( m o d π ) 8A \equiv 4B \equiv 2C \equiv A \pmod{\pi} , hence 7 A 0 ( m o d π ) 7A \equiv 0 \pmod{\pi} .

Calvin Lin Staff - 7 years ago
Saurav Shakya
May 20, 2014

x/z - zx = 2/3 => x = 2/3 { z/(1-z^2)}

y/x - 9xy =2 => y = 2 { x/(1-9x^2)}

z/y - 9yz =6 => z = 6 {y/(1-9y^2)} Now, Let 3x = tana , 3y = tanb and z=tanc we get, tana = tan2c=>a =2c +n1* 180 tanb = tan2a=>b =2a+n2* 180 tanc = tan2b=>c =2b+n3 180 where n1,n2,n3 any integer, Also, c = 2(2a+n2 *180) +n3 180 c =4a + 2n2 180+n3 180 c =4(2c+n1 180)+2n2 180+n3 180 c =8c + (4n1+2n2+n3) 180 7c= - (4n1+2n2+n3)*180 c= - (4n1 +2n2 +n3)/7 *180 Since n1 , n2 and n3 can be any integers (4n1 +2n2+n3) can also be any integer. Thus, c= n/7 *180 , where n is any integer Again, z=tanc = tan(n/7 * 180) Since the period of tan(theta) is 180. we have 7 different solutions for z out of which one is 0. But z cannot be 0 so we reject it. Now, from every other non zero solutions of z we can find one non zero x for each z as x = 2/3 { z/(1-z^2)} and from every this non zero x values we can find one y for each x as y = 2 { x/(1-9x^2)}. Thus, we have 6 set of distinct real solutions.

Notation very hard to read; uses 180 instead of \pi. The solution is fine otherwise.

Calvin Lin Staff - 7 years ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...