Classical Inequalities addicts can do this. Part 2

Algebra Level 5

Let x , y , z , a , b x,y,z,a,b be all real numbers satisfying 25 x 2 + 16 y 2 + 9 z 2 + 4 a 2 + b 2 = 100 3 25x^2 + 16y^2 + 9z^2 + 4a^2 + b^2 = \dfrac{100}{3} .

Denote f ( 5 x , 4 y , 3 z , 2 a , b ) = ( 5 x + 4 y + 3 z + 2 a ) 2 + ( 4 y + 3 z + 2 a + b ) 2 + ( 3 z + 2 a + b + 5 x ) 2 + ( 2 a + b + 5 x + 4 y ) 2 + ( b + 5 x + 4 y + 3 z ) 2 . \begin{aligned} f(5x,4y,3z,2a,b)& =& (5x+4y+3z+2a)^2 + (4y+3z+2a+b)^2 \\&& +(3z+2a+b+5x)^2 + (2a+b+5x+4y)^2 \\ && + (b+5x+4y+3z)^2 .\end{aligned}

Submit your answer as max ( f ( 5 x , 4 y , 3 z , 2 a , b ) ) min ( f ( 5 x , 4 y , 3 z , 2 a , b ) ) \max(f(5x,4y,3z,2a,b)) - \min(f(5x,4y,3z,2a,b)) .


For more problems like this, try answering this set .


The answer is 500.

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

Christian Daang
Jan 23, 2017

Let's generalize again!

let a 1 + a 2 + . . . + a n = x ( a 1 ) 2 + ( a 2 ) 2 + . . . + ( a n ) 2 = y a_1 + a_2 + ... + a_n = x \\ (a_1)^2 + (a_2)^2 + ... + (a_n)^2 = y

our aim is to minimize f ( a 1 , a 2 , . . . , a n ) = ( a 1 + a 2 + a 3 + . . . + a n 1 ) 2 + ( a 2 + a 3 + a 4 + . . . + a n ) 2 + ( a 3 + a 4 + . . . + a n + a 1 ) 2 + . . . + ( a n + a 1 + a 2 + . . . + a n 2 ) 2 f(a_1 , a_2 , ... , a_n) = (a_1 + a_2 + a_3 + ... + a_{n-1})^2 + (a_2 + a_3 + a_4 + ... + a_n)^2 + (a_3 + a_4 + ... + a_n + a_1)^2 + ... + (a_n + a_1 + a_2 + ... + a_{n-2})^2 .

This can be simplify to: f ( a 1 , a 2 , . . . , a n ) = ( x a n ) 2 + ( x a 1 ) 2 + ( x a 2 ) 2 + . . . + ( x a n 1 ) 2 = n x 2 2 x 2 + y = ( n 2 ) x 2 + y f(a_1 , a_2 , ... , a_n) = (x - a_n)^2 + (x - a_1)^2 + (x - a_2)^2 + ... + (x - a_{n-1})^2 \\ = nx^2 - 2x^2 + y \\ = (n-2)x^2 + y

By cauchy, x 2 n [ ( a 1 ) 2 + ( a 2 ) 2 + . . . + ( a n ) 2 ] x 2 n y \rightarrow x^2 \le n[(a_1)^2 + (a_2)^2 + ... + (a_n)^2] \rightarrow x^2 \le ny

By combining the trivial inequality and x 2 n y x^2 \le ny ,

0 x 2 n y 0 ( n 2 ) x 2 n ( n 2 ) y y f ( a 1 , a 2 , . . . , a n ) [ ( n 1 ) 2 ] y \rightarrow 0 \le x^2 \le ny \\ 0 \le (n-2)x^2 \le n(n-2)y \\ y \le f(a_1 , a_2 , ... , a_n) \le \left[ (n-1)^2 \right] y .

Then,

m i n ( f ( a 1 , a 2 , . . . , a n ) ) = y m a x ( f ( a 1 , a 2 , . . . , a n ) ) = [ ( n 1 ) 2 ] y \rightarrow min(f(a_1 , a_2 , ... , a_n)) = y \\ max(f(a_1 , a_2 , ... , a_n)) = \left[ (n-1)^2 \right] y .

now, letting n = 5 a 1 = 5 x a 2 = 4 y a 3 = 3 z a 4 = 2 a a 5 = b y = ( 5 x ) 2 + ( 4 y ) 2 + ( 3 z ) 2 + ( 2 a ) 2 + b 2 = 100 3 n = 5 \\ a_1 = 5x \\ a_2 = 4y \\ a_3 = 3z \\ a_4 = 2a \\ a_5 = b \\ y = (5x)^2 + (4y)^2 + (3z)^2 + (2a)^2 + b^2 = \cfrac{100}{3}

f ( 5 x , 4 y , 3 z , 2 a , b ) = ( 5 x + 4 y + 3 z + 2 a ) 2 + ( 4 y + 3 z + 2 a + b ) 2 + ( 3 z + 2 a + b + 5 x ) 2 + ( 2 a + b + 5 x + 4 y ) 2 + ( b + 5 x + 4 y + 3 z ) 2 m i n ( f ( 5 x , 4 y , 3 z , 2 a , b ) ) = y = 100 3 m a x ( f ( 5 x , 4 y , 3 z , 2 a , b ) ) = [ ( n 1 ) 2 ] y = [ ( 5 1 ) 2 ] [ 100 3 ] = 1600 3 \rightarrow f(5x,4y,3z,2a,b) = (5x+4y+3z+2a)^2 + (4y+3z+2a+b)^2 + (3z+2a+b+5x)^2 + (2a+b+5x+4y)^2 + (b+5x+4y+3z)^2 \\ min(f(5x,4y,3z,2a,b)) = y = \cfrac{100}{3} \\ max(f(5x,4y,3z,2a,b)) = [(n-1)^2]y = [(5-1)^2] \left[ \cfrac{100}{3} \right] = \cfrac{1600}{3}

m a x ( f ( 5 x , 4 y , 3 z , 2 a , b ) ) m i n ( f ( 5 x , 4 y , 3 z , 2 a , b ) ) = 1600 3 100 3 = 500 \therefore max(f(5x,4y,3z,2a,b)) - min(f(5x,4y,3z,2a,b)) = \cfrac{1600}{3} - \cfrac{100}{3} = \boxed{500}

A solution using matrices.

We are given that i = 1 5 x i 2 = r 2 \sum_{i=1}^5 x_i^2=r^2 , where x 1 = 5 x , x 2 = 4 y , x 3 = 3 z , x 4 = 2 a , x 5 = b , r 2 = 100 3 x_1=5x,\ x_2=4y,\ x_3=3z,\ x_4=2a,x_5=b,\ r^2=\frac{100}{3} . Let, x = [ x 1 x 2 x n ] T \mathbf{x}=[x_1\ x_2\ \cdots\ x_n]^T , where n = 5 n=5 . The function f f can be then expressed as i n ( x x i ) 2 \sum_i^n(x-x_i)^2 , where x = i = 1 n x i x=\sum_{i=1}^n x_i . This can be further expressed as f = x T ( i = 1 n ( 1 i e i ) ( 1 i e i ) ) x f=\mathbf{x}^T\left(\sum_{i=1}^n (\mathbf{1}_i-\mathbf{e}_i)(\mathbf{1}_i-\mathbf{e}_i)\right)\mathbf{x} where 1 = [ 1 1 1 1 ] T R n \mathbf{1}=[1 \ 1\ 1\cdots\ 1]^T\in \mathbb{R}^n , and e i = [ 0 0 0 1 0 0 ] T \mathbf{e}_i=[0\ 0\ \cdots\ 0\ 1\ 0\ \cdots\ 0]^T , with the 0 0 at the i i th coordinate. Clearly, i = 1 n ( 1 i e i ) ( 1 i e i ) = n 1 1 T 1 i = 1 n e i T i = 1 n e i i = 1 n 1 T + i = 1 n e i e i T = ( n 2 ) 1 1 T + I \sum_{i=1}^n (\mathbf{1}_i-\mathbf{e}_i)(\mathbf{1}_i-\mathbf{e}_i)=n\mathbf{1}\mathbf{1}^T-\mathbf{1}\sum_{i=1}^n \mathbf{e}_i^T-\sum_{i=1}^n\mathbf{e}_i\sum_{i=1}^n \mathbf{1}^T+\sum_{i=1}^n\mathbf{e}_i\mathbf{e}_i^T=(n-2)\mathbf{1}\mathbf{1}^T+\mathbf{I} where I \mathbf{I} is the identity matrix. Thus, f ( x ) = ( n 2 ) ( x T 1 ) 2 + x T x = ( n 2 ) ( i = 1 n x i ) 2 + r 2 f(\mathbf{x})=(n-2)(\mathbf{x}^T\mathbf{1})^2+\mathbf{x}^T\mathbf{x}=(n-2)\left(\sum_{i=1}^n x_i\right)^2+r^2 Now, it can be seen that the minimum value of ( i = 1 n x i ) 2 (\sum_{i=1}^n x_i)^2 is 0 0 as it can be achieved as follows: If n n is even, then by choosing n / 2 n/2 entries each equal to + r / 2 n +r/{\sqrt{2n}} , and n / 2 n/2 entries r / 2 n -r/{\sqrt{2n}} , and if n n is odd, then by choosing ( n 1 ) / 2 (n-1)/2 entries, r / 2 n r/\sqrt{2n} , ( n 1 ) / 2 (n-1)/2 entries r / 2 n -r/\sqrt{2n} and one entry 0 0 . The maximum value of ( i = 1 n x i ) 2 (\sum_{i=1}^n x_i)^2 is obtained from Cauchy-Scwartz to be n r 2 nr^2 , which is achieved when all the n n entries are equal to r / n r/\sqrt{n} . Thus, max f min f = ( n 2 ) n r 2 0 = n ( n 2 ) r 2 \max f-\min f=(n-2)nr^2-0=n(n-2)r^2 , which, for the given question, evaluates to 500 \boxed{500} .

Abhishek Sinha
Feb 4, 2017

Define the matrix A = ( 1 1 1 1 0 0 1 1 1 1 1 0 1 1 1 1 1 0 1 1 1 1 1 0 1 ) A= \begin{pmatrix} 1 & 1 & 1 & 1 & 0 \\ 0 & 1 & 1 & 1 &1 \\ 1 & 0 & 1 & 1 &1 \\ 1 & 1 & 0 & 1 & 1\\ 1 & 1 & 1 &0 &1 \end{pmatrix} and the vector r = ( 5 x 4 y 3 z 2 a b ) . \vec{r} = \begin{pmatrix} 5x \\ 4y \\ 3z \\ 2a \\ b \end{pmatrix}. Then we can write f ( r ) = A r 2 = r A A r f(\vec{r})= ||A\vec{r}||^2= \vec{r}'A'A\vec{r} with the constraint, r r = 100 3 . \vec{r}'\vec{r}= \frac{100}{3}. Using the Courant-Fischer theorem , we immediately have for non-zero r \vec{r} max f ( r ) = λ max ( A A ) × 100 3 = 16 × 100 3 \max f(\vec{r}) = \lambda_{\max}(A'A) \times \frac{100}{3}= 16 \times \frac{100}{3} and, min f ( r ) = λ min ( A A ) × 100 3 = 1 × 100 3 , \min f(\vec{r})= \lambda_{\min}(A'A) \times \frac{100}{3}= 1\times \frac{100}{3}, where λ max ( A A ) \lambda_{\max}(A'A) and \lambda_\min(A'A) denote the largest and smallest eigenvalue of the real symmetric matrix A A A'A . Hence, the answer follows.

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...