Calculus or Algebra?

Algebra Level 5

Given three real numbers x , y x, y and z z such that x y + y z + x z = 5 xy+yz+xz=-5 and x y z = 3 xyz=3 , find the maximum value of x x . If the maximum value of x x is n n , find the value of 2 n 3 2n^3 .


Inspiration .


The answer is 54.

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.

3 solutions

Rishabh Jain
Feb 6, 2016

A l g e b r a i c m e t h o d \huge\color{#302B94}{\boxed{ \color{#007fff}{\mathcal{Algebraic ~method}}}} x y z = 3 y z = 3 x . . . . ( 1 ) xyz=3 \Rightarrow \color{#D61F06}{yz=\dfrac{3}{x}}....(1) x y + z x = 5 y z = 5 3 x . . ( Using 1 ) xy+zx=-5-yz= -5-\dfrac{3}{x}..(\text{Using 1}) y + z = 5 x 3 x 2 . . . . ( 2 ) \Rightarrow \color{#D61F06}{y+z=\dfrac{-5}{x} - \dfrac{3}{x^2}}....(2) Now from (1) and (2) and using ( y z ) 2 0 o r ( y + z ) 2 4 y z \color{#D61F06}{(y-z)^2\geq0~ or ~(y+z)^2\geq 4yz} , we get ( 5 x + 3 ) 2 x 4 12 x \dfrac {(5x+3)^2}{x^4}\geq \dfrac{12}{x} 12 x 3 25 x 2 30 x 9 0 \Rightarrow 12x^3-25x^2-30x-9 \leq 0 ( x 3 ) ( 12 x 2 + 11 x + 3 ) 0 \Rightarrow (x-3)(12x^2+11x+3)\leq 0 Since 12 x 2 + 11 x + 3 > 0 12x^2+11x+3>0 for all real x we can safely right x 3 \large x\leq3 Hence maximum value of x=3 and 2 ( 3 ) 3 = 54 \large 2(3)^3=54

Nice purely algebraic solution! Equation ( 2 ) (2) , though, must'nt it be 5 x 3 x 2 \frac{-5}{x}-\frac{3}{x^2} ...? Also, why are you so experienced when it comes to inequalities?

Manuel Kahayon - 5 years, 4 months ago

Log in to reply

Typo :)... No not a experienced at all..I am just learning to use inequalities ... :- }

Rishabh Jain - 5 years, 4 months ago

Log in to reply

That doesn't seem to be the case, as you already seem to have solutions for many inequality problems.

Speaking of which, have you already heard of this set?

Manuel Kahayon - 5 years, 4 months ago

Log in to reply

@Manuel Kahayon No at this stage it's way above my caliber.. But certainly some day I'll develop my skills to complete that set.. Have you tried that set??

Rishabh Jain - 5 years, 4 months ago

Log in to reply

@Rishabh Jain Uh-huh... Worked on 3 problems for an hour each, got stumped, then viewed solutions. Currently, viewed solutions for 3, solved none.

Manuel Kahayon - 5 years, 4 months ago

Log in to reply

@Manuel Kahayon Needs practise.. Hey I found a mistake in my solution :- { When x attains its maximum value y and z come out to be negative which means that AM-GM cannot be applied to them ??

Rishabh Jain - 5 years, 4 months ago

Log in to reply

@Rishabh Jain For all real numbers x x and y y , ( x + y ) 2 4 x y (x+y)^2 \geq 4xy , since this expands to x 2 2 x y + y 2 0 x^2-2xy+y^2 \geq 0 , or ( x y ) 2 0 (x-y)^2 \geq 0 which is true for all x and y.

Since you used this in your solution, it is correct. AM-GM does not work because if x + y 2 x y \frac{x+y}{2} \geq \sqrt{xy} is used for negative values of x or y, then x y \sqrt{xy} becomes complex, whereareas ( x + y ) 2 4 x y (x+y)^2 \geq 4xy is always true.

Manuel Kahayon - 5 years, 4 months ago

Log in to reply

@Manuel Kahayon Ohh.. Right I was just editing my solution to change it from a AM-GM to a Cauchy-Shwarz one(it also works here!!). :D

Rishabh Jain - 5 years, 4 months ago

@Manuel Kahayon I just saw your status.... Do you know Hindi language too?

Rishabh Jain - 5 years, 4 months ago

Log in to reply

@Rishabh Jain Uhh, I'm a filipino, so I only know tagalog and english. I'm currently using brilliant.org in order to gather enough knowledge to beat a previous international olympiad from our country. I want to be able to join international competitions :)

Manuel Kahayon - 5 years, 4 months ago

@Rishabh Jain Rishabh, I have posted a new problem... Since you seem to like inequalities, can you try to solve it...? :)

Manuel Kahayon - 5 years, 4 months ago

Log in to reply

@Manuel Kahayon Great... nice title... I'll try my best(Although chances are low that I'll solve the problem.. ;p

Rishabh Jain - 5 years, 4 months ago

@Manuel Kahayon @Manuel Kahayon I solved one by bashing with Langrange Multipliers.

Abdur Rehman Zahid - 5 years, 2 months ago

exactly the same as mine

P C - 5 years, 4 months ago
Zk Lin
Feb 15, 2016

This proof is more intuitive than it is rigorous or elegant. However, this doesn't make it any less valid!

From x y z = 3 xyz=3 and x y + y z + x z = 5 xy+yz+xz=-5 , we can conclude that only one of x , y x,y or z z is positive whereas another two are negative. Without loss of generality, let x > y z x>y \geq z (ie. x x is positive, y y and z z are negative.)

We obtain the following equations:

x = 3 y z x=\frac{3}{yz}

x = 5 y z y + z x=\frac{-5-yz}{y+z}

Equating both equations yield ( y z ) 2 + 5 y z + 3 ( y + z ) = 0 {(yz)}^{2}+5yz+3(y+z)=0 . Viewing this as a quadratic equation in y z yz , we obtain y z = 5 ± 25 12 ( y + z ) 2 yz=\frac{-5\pm \sqrt{25-12(y+z)}}{2} . Since y z yz is positive, the we replace the ± \pm with + + .

Therefore, y z = 5 + 25 12 ( y + z ) 2 yz=\frac{-5 + \sqrt{25-12(y+z)}}{2} . We see that if ( y + z ) (y+z) is maximized within the allowed constraints, then y z yz is minimized within the allowed constraints. Looking at x = 3 y z x=\frac{3}{yz} again, to maximize x x , minimizing y z yz within the allowed constraints is exactly what we want to do!

Therefore, it is natural for us to seek ways to maximize y + z y+z within the allowed constraints. If we succeed in doing so, y z yz will be minimized, which is what we want.

Note that minimizing ( y + z ) -(y+z) is equivalent to maximizing ( y + z ) (y+z) . Since ( y + z ) < 0 (y+z)<0 , to maximize it, we are bringing it as close to 0 0 as possible. This is similar with minimizing ( y + z ) > 0 -(y+z)>0 .

By AM-GM (note that z y > 0 ) -z \geq -y >0) , we have

y z -y-z

2 ( y ) ( z ) \geq 2\sqrt{(-y)(-z)}

= 2 y z =2\sqrt{yz}

Therefore, min y z = 2 y z \min_{-y-z}=2\sqrt{yz} .

However, since y z 2 y -y-z \geq -2y , we can also say that min y z = 2 y \min_{-y-z}=-2y .

Equating the both gives 2 y z = 2 y 2\sqrt{yz}=-2y , which implies y = z y=z when the expression ( y + z ) -(y+z) attains its minimum.

Substitute y = z y=z back into our original equations gives x y 2 = 3 xy^{2}=3 and 2 x y + y 2 = 5 2xy+y^{2}=-5 .

From the first equation, we get y = 3 x y=-\sqrt{\frac{3}{x}} , y 2 = 3 x y^{2}=\frac{3}{x} . Substitute this into the second equation gives the following cubic equation in terms of x x :

12 x 3 25 x 2 30 x 9 = 0 12x^{3}-25x^{2}-30x-9=0

( x 3 ) ( 12 x 2 + 11 x + 3 ) = 0 (x-3)(12x^{2}+11x+3)=0 .

The only real solution is x = 3 x=3 . We check that x = 3 , y = z = 1 x=3,y=z=-1 indeed satisfies the two equations above, hence the solution.

Moderator note:

This can be greatly simplified. You have the good idea of looking at the discriminant of a quadratic equation with real roots. In particular, consider the quadratic in y y : ( z 2 ) y 2 + ( 5 z + 3 ) y + 3 z = 0 (z^2) y^2 + (5z + 3) y + 3z = 0 .

Since this has a real root y y , we get that the discriminant is non-negative, or that Δ = ( 5 z + 3 ) 2 4 × z 2 × 3 z 0 \Delta = ( 5z + 3 ) ^2 - 4 \times z^2 \times 3z \geq 0 . Simplifying this, we get ( z 3 ) ( 12 z 2 + 11 z + 3 ) 0 - ( z - 3 ) ( 12z^2 + 11z + 3 ) \geq 0 . It follows that the maximum value of z z is 3, and we are done (replace z z with x x ).

Manuel Kahayon
Feb 6, 2016

Let x , y x, y and z z be the roots of a cubic polynomial function P ( m ) P(m) . Then, by Vieta's formulas, P ( m ) = m 3 + k m 2 + ( x y + y z + x z ) m x y z P(m) = m^3+km^2+(xy+yz+xz)m-xyz , which gives us P ( m ) = m 3 + k m 2 5 m 3 P(m) = m^3+km^2-5m-3 for some k k .

Next, observe the graph of P ( m ) P(m) . We notice that the graph rises to the right and falls to the left, meaning its local maxima is to the left of its local minima. Also, as with all cubic polynomial functions, P ( m ) P(m) has three real roots if and only if its local maxima is greater than or equal to zero, and when its local minima is less than or equal to zero.

Now, obviously, x , y x, y and z z must be real numbers. Since all real numbers become nonnegative when squared, it is implied that in polynomial P ( m ) = m 3 + k m 2 5 m 3 P(m) = m^3+km^2-5m-3 , m 2 m^2 is always positive, and so, this means that for all m m except for 0, the value of P ( m ) P(m) increases as k increases, and decreases as k decreases, as shown in the figure below.

This also implies that as k decreases, the value of the local maxima also decreases, but, the value of the third root also increases, as shown in the illustration below.

So, in short, we have to find the minimum value of k for which the polynomial P ( m ) P(m) has local maxima exactly equal to zero. But, when that happens, notice that two of the roots of P ( m ) P(m) will be equal, as the maxima becomes tangent to the x-axis. Since the roots of P ( m ) P(m) are x , y x, y and z z , this implies that y = z y=z to get the maximum value of x. Substituting,

2 x y + y 2 = 5 2xy+y^2 = -5

x y 2 = 3 xy^2=3

Then, solving the systems, we get x = 3 x=3 , and 2 x 3 = 54 2x^3 = \boxed {54}

That was my first instinct, to let y = z y=z since I have already seen a similar problem like this before.

Reineir Duran - 5 years, 4 months ago

Log in to reply

Reinier? Ikaw ba yan? Friend mo ako sa facebook kasama ni manuel sales... Congrats nga pala sa mtap metrobank... Sana magkita tayo sa national... Sabi sakin ni manuel nagkita raw kayo sa mtap kahapon ahh...

Manuel Kahayon - 5 years, 4 months ago

Log in to reply

Yep, it's me! Thanks and congratulations too. I am not expecting too much that I would go to the nationals but I have a strong desire of getting there.

Reineir Duran - 5 years, 4 months ago

Aw man I got x=3 and totally spaced out on the final answer being in the form 2*x^3.

Tristan Goodman - 2 years, 3 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...