Classical Inequalities addicts can do this. Part 1

Algebra Level 5

{ x 1 + x 2 + x 3 + + x n = 2016 x 1 4 + x 2 4 + x 3 4 + + x n 4 = 2016 × 512 \begin{cases}\begin{aligned} x_{1} + x_{2} + x_{3} + \cdots + x_{n} &= 2016 \\\\ x_{1}^{4} + x_{2}^{4} + x_{3}^{4} + \cdots + x_{n}^{4} &= 2016\times 512 \end{aligned}\end{cases}

Let x 1 , x 2 , , x n x_1, x_2, \ldots , x_n be real numbers .

Find the smallest possible value of n n such that there exists real solutions to the equations above.


For more problems like this, try answering this set .


The answer is 252.

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

Christian Daang
Jan 23, 2017

Relevant wiki: Cauchy-Schwarz Inequality

Let's generalize this.

Let x x and y y denote the values

x = x 1 + x 2 + + x n y = x 1 4 + x 2 4 + + x n 4 \begin{aligned} x& =& x_1 + x_2 + \cdots + x_n \\ y& =& x_1 ^4 + x_2^4 + \cdots + x_n ^4 \end{aligned}

By Cauchy-Schwarz inequality (CS), for all sequences of real numbers a i a_i and b i b_i , we have ( i = 1 n a i 2 ) ( i = 1 n b i 2 ) ( i = 1 n a i b i ) 2 \left(\displaystyle \sum_{i=1}^n a_i^2\right)\left( \displaystyle \sum_{i=1}^n b_i^2\right)\ge \left( \displaystyle \sum_{i=1}^n a_ib_i\right)^2 .

Now suppose a i = x i a_i = x_i for i = 1 , 2 , 3 , , n i = 1,2,3, \ldots, n and b 1 = b 2 = = b n = 1 b_1 = b_2 = \cdots = b_n = 1 , then we have

(x_1^2 + x_2 ^2 + \cdots + x_n ^2 ) \left( \underbrace{1 + 1 + 1 + \cdots + 1}_{n \text{ number of 1's}} \right) \geq (x_1 + x_2 + \cdots + x_n)^2 .

Or equivalently, \color{#3D99F6}{x_1^2 + x_2 ^2 + \cdots + x_n ^2 \geq \dfrac{x^2}n} .

Now suppose we apply CS again, but this time, suppose a i = x i 2 a_i = x_i^2 for i = 1 , 2 , 3 , , n i = 1,2,3, \ldots, n and b 1 = b 2 = = b n = 1 b_1 = b_2 = \cdots = b_n = 1 , then we have

(x_1^4 + x_2 ^4 + \cdots + x_n ^4 ) \left( \underbrace{1 + 1 + 1 + \cdots + 1}_{n \text{ number of 1's}} \right) \geq (x_1^2 + x_2 ^2+ \cdots + x_n^2 )^2 .

Or equivalently, y ( x 1 2 + x 2 2 + + x n 2 ) 2 n \color{#3D99F6}{y \geq \dfrac{ (x_1^2 + x_2 ^2+ \cdots + x_n^2 )^2 } {n} } .

Combining these 2 inequalities (highlighted in blue) gives

y x 4 n 3 . y \geq \dfrac{x^4}{n^3}.

In this case, x = 2016 , y = 2016 × 512 x = 2016, y = 2016\times 512 . Upon simplifying, we get n 252 n\geq \boxed{252} , with x 1 = x 2 = = x 252 = 2016 252 = 8 x_1 = x_2 = \cdots = x_{252} = \dfrac{2016}{252} = 8 .

Ya, i did it with Titu's Lemma!

Md Zuhair - 4 years, 4 months ago

Log in to reply

can you kindly show how you apply Titu's Lemma in this problem? :)

Christian Daang - 4 years, 4 months ago

Log in to reply

Ya, Actually the modified form of Titu's Lemma, let me show you the modified form

For any positive integer m m , we have x 1 m a 1 m 1 + x 2 m a 2 m 1 + + x n m a n m 1 ( x 1 + x 2 + + x n ) m ( a 1 + a 2 + + a n ) m 1 . \frac{x_1^m}{a_1^{m-1}}+\frac{x_2^m}{a_2^{m-1}}+\cdots+\frac{x_n^m}{a_n^{m-1}}\ge\frac{(x_1+x_2+\cdots+x_n)^m}{(a_1+a_2+\cdots+a_n)^{m-1}}.

Md Zuhair - 4 years, 4 months ago

Log in to reply

@Md Zuhair This is wrong. A condition for Titu's lemma to work is to have x 1 , x 2 , , x n x_1, x_2, \ldots , x_n to be strictly positive. But this problem didn't specify that they are all positive numbers.

So you have only shown that the answer is 252 if you made the assumption that x 1 , x 2 , , x n x_1, x_2,\ldots , x_n are positive.

Pi Han Goh - 4 years, 4 months ago

Log in to reply

@Pi Han Goh Yes, Its correct, So how have you done it?

Md Zuhair - 4 years, 4 months ago

@Pi Han Goh So what if some of them are negative? Can you help?

Steven Jim - 4 years ago

Log in to reply

@Steven Jim See the solution posted by the author.

Pi Han Goh - 4 years ago

Log in to reply

@Pi Han Goh I read it. What I mean is that can we use Titu's lemma here?

Steven Jim - 4 years ago

Log in to reply

@Steven Jim No, we can't.

Pi Han Goh - 4 years ago

Log in to reply

@Pi Han Goh Yep sorry. My bad.

Md Zuhair - 4 years ago

Use power mean inequality and you will get your answer in one step.

Aaron Jerry Ninan - 4 years, 4 months ago

Log in to reply

Did the same!

Thomas Jacob - 4 years, 4 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...