Inspired by Brian Charlesworth, again

Algebra Level 5

10 x y + 12 x z + 14 y z 10xy+12xz+14yz

Let M M and m m be the maximum and the minimum of the expression above if x 2 + y 2 + z 2 = 1 x^2+y^2+z^2=1 , where x , y x,y and z z are real numbers. Find M 2 + M m + m 2 M^2+Mm+m^2 .

Bonus question : For 2 a x y + 2 b x z + 2 c y z 2axy+2bxz+2cyz , with the same constraint, express M 2 + M m + m 2 M^2+Mm+m^2 in terms of a , b a,b and c c .


Inspiration


The answer is 110.

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.

2 solutions

We'll look first at the "bonus" question.

We have f ( x , y , z ) = 2 a x y + 2 b x z + 2 c y z f(x,y,z) = 2axy + 2bxz + 2cyz subject to the constraint g ( x , y , z ) = x 2 + y 2 + z 2 = 1. g(x,y,z) = x^{2} + y^{2} + z^{2} = 1.

Setting up (a variation of) the Lagrange equations with f λ g = 0 , \bigtriangledown f - \lambda \bigtriangledown g = 0, we have

  • (i) 2 a y + 2 b z + 2 λ x = 0 λ x + a y + b z = 0 , 2ay + 2bz + 2\lambda x = 0 \Longrightarrow -\lambda x + ay + bz = 0,

  • (ii) 2 a x + 2 c z + 2 λ y = 0 a x λ y + c z = 0 , 2ax + 2cz + 2\lambda y = 0 \Longrightarrow ax - \lambda y + cz = 0, and

  • (iii) 2 b x + 2 c y + 2 λ z = 0 b x + c y λ z = 0. 2bx + 2cy + 2\lambda z = 0 \Longrightarrow bx + cy - \lambda z = 0.

In matrix form this system is of the form A < x , y , z > = < 0 , 0 , 0 > A \cdot <x,y,z> = <0,0,0> where

A = [ λ a b a λ c b c λ ] A = \begin{bmatrix}-\lambda&a&b\\a&-\lambda&c\\b&c&-\lambda\end{bmatrix}

Now in order to produce non-zero solutions, A A must be singular, i.e., det ( A ) = 0. \det(A) = 0. We then have

det ( A ) = λ ( λ 2 c 2 ) + a ( b c + a λ ) + b ( a c + b λ ) = 0 λ 3 ( a 2 + b 2 + c 2 ) λ + 2 a b c = 0 , \det(A) = -\lambda(\lambda^{2} - c^{2}) + a(bc + a\lambda) + b(ac + b\lambda) = 0 \Longrightarrow \lambda^{3} - (a^{2} + b^{2} + c^{2})\lambda + 2abc = 0, (P).

Now as we have learned before from Dr. Bretscher by way of Theorem 3.1 of this link , the maximum M = λ 1 M = \lambda_{1} and minimum m = λ 3 m = \lambda_{3} are the greatest and least roots of the polynomial P . P. (Let the "middle" root be λ 2 \lambda_{2} .)

Now by Vieta's we have that

  • λ 1 + λ 2 + λ 3 = 0 , \lambda_{1} + \lambda_{2} + \lambda_{3} = 0,

  • λ 1 λ 2 + λ 1 λ 3 + λ 2 λ 3 = ( a 2 + b 2 + c 2 ) , \lambda_{1} \lambda_{2} + \lambda_{1} \lambda_{3} + \lambda_{2} \lambda_{3} = -(a^{2} + b^{2} + c^{2}), and

  • λ 1 λ 2 λ 3 = 2 a b c . \lambda_{1} \lambda_{2} \lambda_{3} = -2abc.

From this we see that, since λ 2 = λ 1 λ 2 , \lambda_{2} = -\lambda_{1} - \lambda_{2}, the second of these equations becomes

λ 1 ( λ 1 + λ 3 ) + λ 1 λ 3 λ 3 ( λ 1 + λ 3 ) = ( a 2 + b 2 + c 2 ) λ 1 2 + λ 1 λ 3 + λ 3 2 = a 2 + b 2 + c 2 . -\lambda_{1}(\lambda_{1} + \lambda_{3}) + \lambda_{1} \lambda_{3} - \lambda_{3}(\lambda_{1} + \lambda_{3}) = -(a^{2} + b^{2} + c^{2}) \Longrightarrow \lambda_{1}^{2} + \lambda_{1} \lambda_{3} + \lambda_{3}^{2} = a^{2} + b^{2} + c^{2}.

So in the way M M and m m have been designated, we have that M 2 + m M + m 2 = a 2 + b 2 + c 2 . M^{2} + mM + m^{2} = a^{2} + b^{2} + c^{2}.

In this specific case we have that a = 5 , b = 6 a = 5, b = 6 and c = 7 , c = 7, and so

M 2 + m M + m 2 = 5 2 + 6 2 + 7 2 = 110 . M^{2} + mM + m^{2} = 5^{2} + 6^{2}+ 7^{2} = \boxed{110}.

Very nice (+1)! I will carefully read and enjoy your solution in the evening after work.

Otto Bretscher - 5 years, 7 months ago
Otto Bretscher
Nov 16, 2015

Brian has written a fine solution. For the sake of variety, I will phrase my solution in the language of quadratic forms, although this is essentially equivalent to Lagrange multipliers.

The symmetric matrix of our quadratic form (from the "bonus question") is A = [ 0 a b a 0 c b c 0 ] A=\begin{bmatrix}0&a&b\\a&0&c\\b&c&0\end{bmatrix}

The characteristic polynomial is p ( λ ) = λ 3 ( a 2 + b 2 + c 2 ) λ 2 a b c p(\lambda)=\lambda^3-(a^2+b^2+c^2)\lambda-2abc . Since M M and m m are roots of p p , we have M 3 = ( a 2 + b 2 + c 2 ) M + 2 a b c M^3=(a^2+b^2+c^2)M+2abc and m 3 = ( a 2 + b 2 + c 2 ) m + 2 a b c m^3=(a^2+b^2+c^2)m+2abc . Subtracting the two equations gives M 3 m 3 = ( a 2 + b 2 + c 2 ) ( M m ) M^3-m^3=(a^2+b^2+c^2)(M-m) so M 2 + M m + m 2 = M 3 m 3 M m = a 2 + b 2 + c 2 M^2+Mm+m^2=\frac{M^3-m^3}{M-m}=a^2+b^2+c^2 .

In the special case ( a , b , c ) = ( 5 , 6 , 7 ) (a,b,c)=(5,6,7) the answer is M 2 + M m + m 2 = 5 2 + 6 2 + 7 2 = 110 M^2+Mm+m^2=5^2+6^2+7^2=\boxed{110}

Moderator note:

Very nice.

I'm becoming a fan of using quadratic forms to prove quadratic inequalities, as it greatly simplifies the work that is to be done.

I used my TI-84 Plus Silver Edition Calculator to get a VERY close approximation to M. All the choices for x,y and z are decimals (I know non-terminating decimals would get a different Maximum).... x=.5434 , y=.5791521 , z=.6077 with the sum of their squares being 1.000000005. Now with these choices for x, y and z the Maximum of 10xy + 12xz + 14yz is 12.0371128 !! My choices for finding the Minimum were: x=0 , y=1 , z=0 (All perfectly good choices that satisfy the sum of squares condition exactly). This causes the Minimum to be ZERO. Then I saw that MM + Mm + mm would become just MM = 144.8920846 leading me to believe that Minimum, m, must be negative. Is that correct? I better learn this more advanced Math above. Brilliant teaches me more when I get problems incorrect. I was really surprised by the difference from 110 when I entered 145 !! I wish the problem had specified non-negative real numbers for x,y and z.

Update: When allowing for a negative Minimum I got m= -6.99715802 and a final answer of about 109.6267 - I like the Analytical solution better but it is nice to know that my 20-year old calculator still helped and gave me values for x, y and z !!!

Bob Kadylo - 5 years, 6 months ago

Log in to reply

I like your solution sir. Very comprehendable. hahaha sorry Im just being sarcastic your solution is not elegant either... Get out on brilliant.

Rei Allester - 5 years, 5 months ago

Log in to reply

Did you mean "Get out _ of_ brilliant." Perhaps for someone who has solved 3 problems and accumulated 305 points in 9 months, you could tone it down a bit.

Bob Kadylo - 5 years, 5 months ago

Can we specifically find the individual values of M M and m m respectively? If so then how can we find it?

dhiraj agarwalla - 5 years, 6 months ago

Log in to reply

Yes, you can use Cardano's formula for the cubic equation to find M M and m m , but the expression comes out pretty complicated... you can get it from WolframAlpha.

Otto Bretscher - 5 years, 5 months ago

Log in to reply

Thank you :)

dhiraj agarwalla - 5 years, 5 months ago

exquisite. Did it the exact same way. Lagrange Multipliers are way cool

Rei Allester - 5 years, 5 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...