Back To Basics II

Algebra Level 5

x , y , x,y, and z z are real numbers satisfying the equation x 2 + y 2 + z 2 = 1. x^{2}+y^{2}+z^{2}=1.

The maximum value of 3 x y 2 z 3xy^{2}z can be expressed in the form a b \dfrac{a}{b} for co-prime positive integers a a and b b .

Find a + b a+b .


Bonus:

Can you generalize the maximum value of x m y n z p x^{m}y^{n}z^{p} for positive integers m , n , p m,n,p given the same constraints?


The answer is 11.

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.

1 solution

Brandon Monsen
Dec 5, 2015

Since the constraints are squares of numbers, and we aren't subtracting or otherwise doing anything to the expression where a negative sign would matter, we can assume that x , y , z x,y,z are all positive.

Rewrite our constraint as

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

By AM-GM Inequality , we get that

x 2 + 1 2 y 2 + 1 2 y 2 + z 2 4 1 4 x 2 y 4 z 2 4 \frac{x^{2}+\frac{1}{2}y^{2}+\frac{1}{2}y^{2}+z^{2}}{4} \geq \sqrt[\Large{4}]{\frac{1}{4}x^{2}y^{4}z^{2}}

2 4 x y 2 z \frac{\sqrt{2}}{4} \geq \sqrt{xy^{2}z}

x y 2 z 1 8 xy^{2}z \leq \frac{1}{8}

Therefore

3 x y 2 z 3 8 3xy^{2}z \leq \frac{3}{8} .

And we know that equality can occur because by AM-GM once again, we get that

x 2 = 1 2 y 2 = z 2 x^{2}=\frac{1}{2}y^{2}=z^{2}

And by our constraint we get

4 x 2 = 1 2 y 2 = 1 4 z 2 = 1 4x^{2}=1 \\ 2y^{2}=1 \\ 4z^{2}=1

All of which have solutions over the real numbers, therefore equality must be obtained.

And our answer is 3 + 8 = 11 3+8=\boxed{11}


For the generalization:

Rewrite our constraint as

i = 1 m x 2 m + j = 1 n y 2 n + k = 1 p z 2 p = 1 \sum _{ i=1 }^{ m }{ \frac { { x }^{ 2 } }{ m } } +\sum _{ j=1 }^{ n }{ \frac { { y }^{ 2 } }{ n } } +\sum _{ k=1 }^{ p }{ \frac { { z }^{ 2 } }{ p } } =1

And so, by AM-GM, we get

i = 1 m x 2 m + j = 1 n y 2 n + k = 1 p z 2 p m + n + p x 2 m y 2 n z 2 p m m n n p p m + n + p \frac{\sum _{ i=1 }^{ m }{ \frac { { x }^{ 2 } }{ m } } +\sum _{ j=1 }^{ n }{ \frac { { y }^{ 2 } }{ n } } +\sum _{ k=1 }^{ p }{ \frac { { z }^{ 2 } }{ p } }}{m+n+p} \geq \sqrt[\large{m+n+p}]{\frac{x^{2m}y^{2n}z^{2p}}{m^{m}n^{n}p^{p}}}

1 m + n + p x 2 m y 2 n z 2 p m m n n p p m + n + p \frac{1}{m+n+p} \geq \sqrt[\large{m+n+p}]{\frac{x^{2m}y^{2n}z^{2p}}{m^{m}n^{n}p^{p}}}

m m n n p p ( m + n + p ) m + n + p x 2 m y 2 n z 2 p \frac{m^{m}n^{n}p^{p}}{(m+n+p)^{m+n+p}} \geq x^{2m}y^{2n}z^{2p}

And finally...

x m y n z p m m n n p p ( m + n + p ) m + n + p x^{m}y^{n}z^{p} \leq \sqrt{\frac{m^{m}n^{n}p^{p}}{(m+n+p)^{m+n+p}}}

Does this still work if m , n , p m,n,p are positive real numbers (and x , y , z x,y,z are positive as well)?

Otto Bretscher - 5 years, 6 months ago

Log in to reply

I reckon it won't, since you cant have a fraction repeated 7.8 7.8 times. I mis-worded the bonus so I'll fix that now.

Brandon Monsen - 5 years, 6 months ago

Log in to reply

But does your result still hold when m , n , p m,n,p aren't integers?

Otto Bretscher - 5 years, 6 months ago

Log in to reply

@Otto Bretscher From the tests I just did it will not, but once I get home later on tonight I'll see what I can do about fixing that.

Brandon Monsen - 5 years, 6 months ago

Log in to reply

@Brandon Monsen It should work... I did the problem with Lagrange multipliers, and there we are not restricted to integer values of m , n m,n and p p .

Otto Bretscher - 5 years, 6 months ago

Log in to reply

@Otto Bretscher And you got that exact maximum value when doing so? That's interesting. I assumed it would fall apart due to the fractional part in the summation.

Brandon Monsen - 5 years, 6 months ago

Log in to reply

@Brandon Monsen Lagrange multipliers, solved for λ \lambda , give you m x m 2 y n z p mx^{m-2}y^nz^p = n x m y n 2 z p =nx^my^{n-2}z^p = p x m y n z p 2 =px^my^nz^{p-2} so x 2 m = y 2 n = z 2 p \frac{x^2}{m}=\frac{y^2}{n}=\frac{z^2}{p} ... same condition.

Otto Bretscher - 5 years, 6 months ago

Log in to reply

@Otto Bretscher Oh ok awesome! Thanks for confirming!

Brandon Monsen - 5 years, 6 months ago

Log in to reply

@Brandon Monsen You can also do AM-GM with fractional weights, of course... just write m x 2 m + n y 2 n + p z 2 p = 1 m\frac{x^2}{m}+n\frac{y^2}{n}+p\frac{z^2}{p}=1 instead of the sums.

Otto Bretscher - 5 years, 6 months ago

To complete your solution, you must show that equality can occur.

Jon Haussmann - 5 years, 6 months ago

Log in to reply

True, Let me fix that

Brandon Monsen - 5 years, 6 months ago

you cant apply am gm on all real numbers

Dev Sharma - 5 years, 6 months ago

Log in to reply

Oh sorry, you are applying it on squares which is positive

Dev Sharma - 5 years, 6 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...