A Plane, A Sphere And A Hyperbola Walked Into A Bar

Algebra Level 4

x , y , x, y, and z z are real numbers such that x + y + z = 0 x+y+z= 0 and x 2 + y 2 + z 2 = 1 x^2 + y^2 + z^2 = 1 . What is the smallest possible value of x y z xyz ?


The answer is -0.136.

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.

9 solutions

Abhishek Sinha
Apr 18, 2014

Substituting for x + y x+y in the second equation, we get x y = z 2 1 / 2 xy=z^2-1/2 . Hence x y z = z 3 z / 2 xyz=z^3-z/2 . Now, it is a simple matter of calculus to find global minima of this function.

EDIT : We are doing a constrained minimization here subject to z 2 / 3 |z|\leq \sqrt{2/3} . Please refer to my reply in response to Calvin's comment.

The global minimum of a cubic polynomial is - \infty . Can you explain your steps clearer?

Calvin Lin Staff - 7 years, 1 month ago

Log in to reply

We need to take into account the fact that the variables are real. First fix z z . Then we have two simultaneous equations x + y = z x+y=-z and x y = z 2 1 / 2 xy= z^2-1/2 . Eliminating y y , we have x 2 + x z + ( z 2 1 / 2 ) = 0 x^2 + xz+ (z^2-1/2)=0 . Since x x is real, the discriminant of this quadratic must be non-negative. Thus we get z 2 / 3 |z|\leq \sqrt{2/3} . Then the problem reduces to minimizing the cubic z 3 z / 2 z^3-z/2 subject to z 2 / 3 |z|\leq \sqrt{2/3} .

Abhishek Sinha - 7 years, 1 month ago

Log in to reply

That's much better. An important step here is using the discriminant (or otherwise) to bound the possible range of values.

A common mistake is to say that since z 2 1 z^2 \leq 1 hence 1 z 1 -1 \leq z \leq 1 . Substituting z = 1 z = -1 in will give us the value of 1.5 -1.5 , but then x x and y y will not be real.

Calvin Lin Staff - 7 years, 1 month ago

Log in to reply

@Calvin Lin The real value of x is 2/sqrt6 and of y is -1/ sqrt6 and z is -1/sqrt6. Thus least value of xyz is 2/sqrt6*1/6=0.136 not -0.136???? Could you please explain why this is wrong.

Ashtik Mahapatra - 7 years, 1 month ago

Log in to reply

@Ashtik Mahapatra We can clearly get x y z = 0 xyz = 0 by settting x = 0 , y = 1 2 , z = 1 2 x = 0, y = \frac{1}{\sqrt{2}}, z = - \frac{1} { \sqrt{2} } , and so the minimum is 0 \leq 0 .

You actually found the maximium value, as opposed to the minimum. Remember that if you use calculus, you have to further check the second order derivative to see what kind of extrema you get (maximum, minimum, inflection).

Calvin Lin Staff - 7 years, 1 month ago

Log in to reply

@Calvin Lin Oh I forgot to do that? Dude I could have got this right easily.

Ashtik Mahapatra - 7 years, 1 month ago

1/3

Muhammad Asad Khan - 7 years, 1 month ago

@Calvin Lin Is this a Brilliant Algebra problem ? Or am I just brooding it out ???

Nishant Sharma - 7 years ago

As per your method, we found z^2=1/6 not 2/3. Another method found that xy+yz+zx=-1/2 & then as we know that AM>=GM. Therefore -1/6>=(xyz)^(2/3). But we found not a real solution. Then again we applied the same application in x^2+y^2+z^2=1 then we found maximum value 27^(-1/2).

AKSHAY KUMAR - 7 years, 1 month ago

Here's what I did. Since x + y + z = 0 x+y+z = 0 x 3 + y 3 + z 3 = 3 x y z { x }^{ 3 }+{ y }^{ 3 }+{ z }^{ 3 }=3xyz Multiplying both sides by 2, we get ( x + y ) ( x 2 + y 2 x y ) + ( y + z ) ( y 2 + z 2 y z ) + ( z + x ) ( z 2 + x 2 z x ) = 6 x y z (x+y)({ x }^{ 2 }+{ y }^{ 2 }-xy) + (y+z)({ y }^{ 2 }+{ z }^{ 2 }-yz) + (z+x)({ z }^{ 2 }+{ x }^{ 2 }-zx) = 6xyz From the given conditions, x + y = z x + y = -z y + z = x y + z = -x z + x = y z + x = -y and x 2 + y 2 = 1 z 2 { x }^{ 2 }+{ y }^{ 2 }=1-{ z }^{ 2 } y 2 + z 2 = 1 x 2 { y }^{ 2 }+{ z }^{ 2 }=1-{ x }^{ 2 } z 2 + x 2 = 1 y 2 { z }^{ 2 }+{ x }^{ 2 }=1-{ y }^{ 2 } Substituting ( z ) ( 1 z 2 x y ) + ( x ) ( 1 x 2 y z ) + ( y ) ( 1 y 2 z x ) = 6 x y z (-z)(1-{ z }^{ 2 }-xy) + (-x)(1-{ x }^{ 2 }-yz) + (-y)(1-{ y }^{ 2 }-zx) = 6xyz Simplifying ( x 2 + y 2 + z 2 ) ( x + y + z ) + 3 x y z = 6 x y z ({ x }^{ 2 }+{ y }^{ 2 }+{ z }^{ 2 }) - (x + y + z) +3xyz = 6xyz 3 x y z = 1 3xyz =1 x y z = 1 3 xyz = \boxed{\frac { 1 }{ 3 }} I checked my working thrice and found no mistake. Please tell me why this method is not applicable. Thank you.

Vineeth Chelur - 7 years, 1 month ago

Log in to reply

What you did is a common mistake in algebra... Its called repeated substitution which gives more answers than expected. For example:-

If x = x 2 1 x = x^{2} - 1 , you cannot substitute again x 2 1 x^{2} - 1 in place of x, as it makes the equation, x = ( x 2 1 ) 2 1 x = (x^{2} - 1)^{2} - 1 , which gives more solutions than expected.

Abhishek Kanthed - 7 years, 1 month ago

Log in to reply

@Abhishek Kanthed While your reasoning is correct, it doesn't apply to this scenario. It does give more solutions than expected, but it will still include the original set of solutions.

@Vineeth Chelur Check your expansion of ( z ) ( 1 z 2 x y ) + (-z)(1-z^2-xy) + \ldots . The first term should be ( x 3 + y 3 + z 3 (x^3+y^3+z^3 , and you will get 0 = 0 0 = 0 which is a true statement.

Calvin Lin Staff - 7 years, 1 month ago

Log in to reply

@Calvin Lin Mind skips the tiniest of things sometimes and many times too. Thanks for rectifying my mistake. I feel like a fool for making such a stupid mistake

Vineeth Chelur - 7 years, 1 month ago

Oh. I had no idea. Thank you for your reply.

Vineeth Chelur - 7 years, 1 month ago
Daniel Liu
Apr 22, 2014

Suppose we have a cubic with roots x , y , z x,y,z . Let this cubic be x 3 + a x 2 + b x + c x^3+ax^2+bx+c . Since x + y + z = 0 x+y+z=0 , we see that a = 0 a=0 . Thus the cubic can be expressed as x 3 + b x + c x^3+bx+c .

Now, by reverse-Newton's Sums, we see that x 2 + y 2 + z 2 + 2 b = 0 x^2+y^2+z^2+2b=0 ; thus x 2 + y 2 + z 2 = 1 = 2 b x^2+y^2+z^2=1=-2b and b = 1 2 b=-\dfrac{1}{2} . Thus, our cubic has reduced to x 3 1 2 x + c = 0 x^3-\dfrac{1}{2}x+c=0 . We want to maximize c c in order to minimize x y z xyz .

Drawing a simple graph tells us that the maximum c c can be while x , y , z x,y,z are still real numbers is when two of the roots are equal. Let this multiplicity-2 root be r r , and let the remaining root be R R . Since r + r + R = 0 r+r+R=0 , we see that R = 2 r R=-2r .

Thus x 3 1 2 x + c = ( x r ) 2 ( x + 2 r ) x^3-\dfrac{1}{2}x+c=(x-r)^2(x+2r) . Expanding this and matching coefficients gives that r = 1 6 r=\sqrt{\dfrac{1}{6}} , and that c = 2 r 3 c=2r^3 . Plugging the value of r r in and solving for c c gives c = 1 3 1 6 0.136 c=\dfrac{1}{3}\sqrt{\dfrac{1}{6}}\approx \boxed{-0.136}

@Calvin Lin see any holes in my solution?

Daniel Liu - 7 years, 1 month ago

Log in to reply

Other than the horrendous abuse of notation for x x ? Grin.

And "minimum x y z xyz " as opposed to maximum?

Calvin Lin Staff - 7 years, 1 month ago
Adrian Neacșu
Apr 18, 2014

I don't think that that's the global minimum.

Here's what I did.

I thought I could use Vieta's formula for a polynomial with degree 3 and solutions x x , y y and z z and dominant coefficient 1.

We know S 1 = x + y + z , S 2 = x y + x z + y z , S 3 = x y z {S}_{1}=x+y+z, {S}_{2}=xy+xz+yz, {S}_{3}=xyz .

We have S 1 = 0 {S}_{1}=0 .

We can get S 2 {S}_{2} if we calculate ( x + y + z ) 2 = x 2 + y 2 + z 2 + 2 S 2 {(x+y+z)}^2={x}^{2}+{y}^{2}+{z}^{2}+2{S}_{2} .

Then S 2 = 0.5 {S}_{2}=-0.5

So the polynomial is f ( X ) = X 3 0.5 X S 3 f(X)={X}^{3}-0.5X-{S}_{3} (I used X since x is already used.)

We know that x x is a root for f f which means that f ( x ) = 0 f(x)=0 .

Which gives us x 3 0.5 x S 3 = 0 {x}^{3}-0.5x-{S}_{3}=0 .

Then S 3 = x 3 0.5 x {S}_{3}={x}^{3}-0.5x .

I calculated the derivative of S 3 {S}_{3} and got two critic points ± 6 6 \pm \frac { \sqrt { 6 } }{ 6 } .

And got a minimum and a maximum, but they're not global.

If you don't think it's the global minimum, can you find a set of values which leads to a smaller x y z xyz ?

Calvin Lin Staff - 7 years, 1 month ago

Log in to reply

No, x should be between in [ 2 3 , 2 3 ] [-\sqrt { \frac { 2 }{ 3 } }, \sqrt { \frac { 2 }{ 3 } } ] so that y and z should be real.

Adrian Neacșu - 7 years, 1 month ago

Log in to reply

Yes, that's the condition that you missed out in your proof.

Another approach of your setup would be to take the condition that a cubic with 3 real roots has non-negative discriminant, which gives us that

4 × 1 × ( 0.5 ) 3 27 × 1 2 × S 3 2 0 S 3 2 1 54 . -4 \times 1 \times (-0.5)^3 - 27 \times 1^2 \times S_3^2 \geq 0 \Rightarrow S_3 ^2 \leq \frac{1}{54}.

It then remains to check that 1 54 - \sqrt{ \frac{1}{54} } can indeed be achieved.

Calvin Lin Staff - 7 years, 1 month ago
Dieuler Oliveira
Jul 30, 2014

Let's find a cubic polynomial such that x , y x, y and z z are their roots (the idea here is to use the discriminant, Δ \Delta , of a cubic polynomial to force it to have only real solutions). Since, x + y + z = 0 x+y+z=0 , it has the following shape: x 3 + p x + q = 0 x^{3}+px+q=0 Δ = q 2 + 4 p 3 27 0 \Delta=q^{2}+\frac{4p^{3}}{27}\leq{0} Now, using Newton's Theorem, we have: S 2 + p . S 0 + q . S 1 = 0 S_{2}+p.S_{0}+q.S_{-1}=0 1 + 3 p + q . S 1 = 0... ( 1 ) 1+3p+q.S_{-1}=0...(1) Let's find S 1 S_{-1} , by using its reciprocal polynomial and Vieta's Formula: q x 3 + p x 2 + 1 = 0 qx^{3}+px^{2}+1=0 then S 1 = p q . . . ( 2 ) S_{-1}=\frac{-p}{q}...(2) Substituting ( 2 ) (2) in ( 1 ) (1) : 1 + 3 p + q . p q = 0 1+3p+q.\frac{-p}{q}=0 p = 1 2 \Rightarrow p=\frac{-1}{2} Therefore, this is the polynomial: x 3 1 2 x + q = 0 x^{3}-\frac{1}{2}x+q=0 Δ = q 2 + 4 ( 1 2 ) 3 27 0 \Delta=q^{2}+\frac{4\left(\frac{-1}{2}\right)^{3}}{27}\leq0 q 2 1 54 q^{2}\leq\frac{1}{54} 0.136 1 3 6 q 1 3 6 0.136 \Rightarrow-0.136 \approx -\frac{1}{3\sqrt{6}}\leq{q}\leq \frac{1}{3\sqrt{6}} \approx 0.136 q m i n 0.136 q_{min} \approx -0.136 m i n { x y z } 0.136 min\{xyz\} \approx \boxed{-0.136}

Aritra Jana
Oct 10, 2014

from given: x y + y z + z x = 1 2 xy+yz+zx=-\frac{1}{2}

we consider a monic cubic polynomial with roots= x , y , z x,y,z

so, from the given,and by vieta's relations, our polynomial P ( a ) = a 3 a 2 x y z P(a)=a^{3}-\frac{a}{2}-xyz

now, P ( a ) = 3 a 2 1 2 P'(a)=3a^{2}-\frac{1}{2}

easily observable that P ( a ) P'(a) is negative between ( 1 6 , 1 6 ) (-\frac{1}{\sqrt{6}},\frac{1}{\sqrt{6}}) and is positive elsewhere

the negative derivative of P ( x ) P(x) signifies it's decreasing nature and the left and right extremities of the set ( 1 6 , 1 6 ) (-\frac{1}{\sqrt{6}},\frac{1}{\sqrt{6}}) are the values at which P ( x ) P(x) attains it's local maximum and minimum respectively.

NOW If the polynomial is to have 3 real roots, it should cut the abscissa thrice, thus making the local minimum 0 ≤0 (or else, the curve wont bulge down into the 4th quadrant in this case)

So, we calculate the local minimum or, P ( 1 6 ) P(\frac{1}{\sqrt{6}})

which gives the value 1 3 6 x y z -\frac{1}{3\sqrt{6}}-xyz

By our arguments, this is 0 ≤0

thus, giving x y z 1 3 6 = 0.136 xyz≥-\frac{1}{3\sqrt{6}}=\boxed{-0.136}

An alternative method for solving this is to use a Lagrange Multiplier with two constraints. Once the gradients are found, multiplied by some constant (hence the name, "Lagrange Multiplier") and set equal to the gradient of the function to be minimized (i.e. xyz), the problem becomes a system of five equations that can be solved through substitution.

James Shi
May 27, 2014

Using the UVW method (http://www.artofproblemsolving.com/Forum/viewtopic.php?t=278791)

x y z xyz is minimized when one of them equals 0 0 or two of them are equal.

If one of them is 0 0 , x y z = 0 xyz = 0 .

If two of them are equal, you have 2 x + y = 0 2x + y = 0 and 2 x 2 + y 2 = 1 2x^2 + y^2 = 1 . Substituting y = 2 x y = -2x into the second equation gives 6 x 2 = 1 6x^2 = 1 . You want to find the minimum value of x 2 y x^2 y , which can be rewritten as y 6 \frac{y}{6} . When x x is positive, y y is negative, and vice versa. You want the negative value of y y to minimize y 6 \frac{y}{6} , so x = 6 6 x = \frac{\sqrt{6}}{6} and y = 6 3 y = -\frac{\sqrt{6}}{3} so the minimum value of x y z xyz is 6 18 0.136 \boxed{-\frac{\sqrt{6}}{18} \approx -0.136} .

Akif Khan
May 3, 2014

We can solve it by converting the given set of equations into a cubic polynomial. This is what i did.

F i r s t l e t x , y , z b e t h r e e r e a l r o o t s o f s o m e c u b i c e q u a t i o n F ( t ) . S o t h e s u m o f r o o t s w i l l b e x + y + z , a n d p r o d u c t o f r o o t s w i l l b e x y z N o w a p p l y i n g t h e i d e n t i t y ( x + y + z ) 2 = x 2 + y 2 + z 2 + 2 ( x y + y z + z x ) w e g e t , ( x y + y z + z x ) = 0.5 S o t h e c u b i c e q u a t i o n w i l l b e t 3 + 0 t 2 0.5 t + d = 0 ( w h e r e d i s t h e c o n s t a n t a n d i s e q u a l t o t h e p r o d u c t o f t h e r o o t s x y z ) S o l v i n g t h e a b o v e e q u a t i o n b y s o l u t i o n o f r a d i c a l s . P u t t = p + q S o w e g e t p 3 + q 3 + ( p + q ) ( 3 p q 0.5 ) + d = 0 N o w p u t t i n g , 3 p q 0.5 = 0 w e g e t , p = 1 6 q . p u t t i n g t h i s v a l u e i n a b o u v e e q n . W e g e t q 6 + ( d ) q 3 + 1 216 = 0. P u t t i n g q 3 = u s o t h e e q u a t i o n b e c o m e s a q u a d r a t i c i n u . u 2 + d u + 1 216 = 0. s i n c e x , y , z a r e r e a l s o a b o v e q u a d r a t i c s h o u l d a l s o h a v e r e a l r o o t s w h i c h c a n o n l y b e i f d 2 \sfrac 4216 0 O r f o r t h e m i n i m u m v a l u e o f d , d = ± 2 6 3 / 2 First\quad let\quad x,y,z\quad be\quad three\quad real\quad roots\quad of\quad some\quad cubic\quad equation\quad F(t).\\ So\quad the\quad sum\quad of\quad roots\quad will\quad be\quad x+y+z\quad ,\quad and\quad product\quad of\quad roots\quad will\quad be\quad xyz\\ Now\quad applying\quad the\quad identity\quad \\ \Rightarrow \quad (x+y+z)^{ 2 }={ \quad x }^{ 2 }\quad +\quad { y }^{ 2 }\quad +\quad { z }^{ 2 }\quad +\quad 2(xy\quad +\quad yz\quad +zx)\\ we\quad get,\quad (xy\quad +\quad yz\quad +zx)\quad =\quad -0.5\\ So\quad the\quad cubic\quad equation\quad will\quad be\quad { t }^{ 3 }\quad +\quad 0{ t }^{ 2 }\quad -0.5t\quad +\quad d\quad =\quad 0\\ (where\quad d\quad is\quad the\quad constant\quad and\quad is\quad equal\quad to\quad the\quad product\quad of\quad the\quad roots\quad xyz)\\ Solving\quad the\quad above\quad equation\quad by\quad solution\quad of\quad radicals.\\ Put\quad t\quad =\quad p+q\\ \Rightarrow \quad \quad So\quad we\quad get\quad { p }^{ 3\quad }+\quad { q }^{ 3 }\quad +\quad (p+q)(3pq-0.5)\quad +d\quad =\quad 0\\ \quad \qquad Now\quad putting,\quad 3pq-0.5=0\\ we\quad get,\quad p=\quad \frac { 1 }{ 6q } .\quad putting\quad this\quad value\quad in\quad abouve\quad eqn.\quad We\quad get\\ \Rightarrow \quad { q }^{ 6 }\quad +\quad { (d)q }^{ 3 }\quad +\quad \frac { 1 }{ 216 } \quad =\quad 0.\quad Putting\quad { q }^{ 3 }=u\quad so\quad the\quad equation\quad becomes\quad a\quad quadratic\quad in\quad u.\\ { \Rightarrow \quad u }^{ 2 }\quad +\quad du\quad +\quad \frac { 1 }{ 216 } \quad =\quad 0.\quad \\ since\quad x,y,z\quad are\quad real\quad so\quad above\quad quadratic\quad should\quad also\quad have\quad real\quad roots\quad which\quad can\quad only\quad be\quad if\quad { d }^{ 2 }-\sfrac { 4 }{ 216 } \quad \ge \quad 0\\ Or\quad for\quad the\quad minimum\quad value\quad of\quad d,\quad d=\pm \quad \frac { 2 }{ 6^{ 3/2 } }

Milly Choochoo
Apr 21, 2014

This was too easy with my method of logic.

First, imagine/draw out the two functions x 2 + y 2 + z 2 = 1 x^2 + y^2 + z^2 = 1 and x + y + z = 0 x + y + z = 0 , and look at their intersections, which looks like the picture in the problem description.

Now if you imagine the function f ( x , y , z ) = x y z f(x,y,z) = xyz , with f ( x , y , z ) f(x,y,z) represented as, let's say, the temperature at the spot ( x , y , x ) (x,y,x) , you can tell that the maximum and minimum temperatures will occur along lines that make equal angles with all of the axes. (Think of it like spears of maximum and minimum values jutting straight out of the origin into the octants - 'maximum' spears in the first, third, sixth, and eighth octants, and 'minimum' spears in the second, fourth, fifth, and seventh octants).

Put the original intersection of those two functions into that temperature-graph of f ( x , y , z ) f(x,y,z) . It is logical to make the assumption now that the highest and lowest 'temperatures' will occur at the tips of that tilted-circle (the intersection of the two 3d graphs) in the third and fifth octants (the tips of the tilted circle in the second and fourth octants have a z z value of zero, so forget those!). The reason for this logic is that those points on the tilted circle of intersection are closest to the lines of maximum values for the function f ( x , y , z ) f(x,y,z) .

So, with that assumption, the minimum will occur when x = y x = y :

x + x + z = 0 x = z 2 x + x + z = 0 \rightarrow x = -\frac{z}{2}

x 2 + x 2 + z 2 = 1 z 2 2 + z 2 = 1 x^2 + x^2 + z^2 = 1 \rightarrow \frac{z^2}{2} + z^2 = 1

z = ± 2 3 \rightarrow z = \pm \sqrt{\frac{2}{3}}

We're looking for the minimum value, so

z = 2 3 z = -\sqrt{\frac{2}{3}}

x y z = ( 1 6 ) ( 1 6 ) ( 2 3 ) xyz = (\frac{1}{\sqrt{6}})(\frac{1}{\sqrt{6}})(-\sqrt{\frac{2}{3}})

= 2 6 3 = \boxed{ -\frac{\sqrt{2} }{6 \sqrt{3}}}

You need to be careful with statements like "you can tell that the maximum and minimum temperatures will occur along lines that make equal angles with all of the axes." and "It is logical to make the assumption now that the highest and lowest 'temperatures' will occur at the tips of that tilted-circle ".

They do not need to be true for general functions, even if they are symmetric / cyclic.

Calvin Lin Staff - 7 years, 1 month ago

Log in to reply

This is very true. The reason why I felt it was appropriate here was because the function f ( x , y , z ) = x y z f(x,y,z)=xyz is quite simply and well behaved (obviously).

Excuse this very late response. I was just going through old problems that I had done.

Milly Choochoo - 6 years, 12 months ago

which is also equal to 1 54 \boxed { - \frac{1}{\sqrt{54}} }

Milly Choochoo - 7 years, 1 month ago

You get the polynomial h(x)= x^3 - x \frac{1}{2} +a and you want to find the maximal a for which h(x) has real roots, this will give the answer -a. By slowly moving the graphic we find that h(x) must have a double root, so we solve the sistem and find -a= - \frac{\sqrt{6}}{18} but plugging this in the answer box doesn't work :(

Musca Paul - 7 years ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...