The Degree Isn't Given Either?

Algebra Level 5

Let a 1 , a 2 , , a n a_1, a_2, \ldots , a_n be the roots of the equation k = 1 n k x k = 0 \displaystyle \sum_{k=1}^n kx^k = 0 . If k = 1 n 1 1 a k = 7 \displaystyle \sum_{k=1}^n \dfrac1{1-a_k} = 7 , find n n .


The answer is 10.

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.

4 solutions

k = 1 n k x k = n k = 1 n ( x a k ) ln ( k = 1 n k x k ) = ln n + k = 1 n ln ( x a k ) = ln ( n k = 1 n ( x a k ) ) \displaystyle \sum_{ k = 1}^n kx^k = n\cdot\prod_{k = 1}^n (x - a_k) \Rightarrow \ln\left(\sum_{ k = 1}^n kx^k \right) = \ln n + \sum_{k =1}^n \ln(x - a_k) = \ln \left( n \cdot \prod_{k = 1}^n (x - a_k) \right) \Rightarrow Derivating with respect to x x k = 1 n 1 x a k = k = 1 n k 2 x k 1 k = 1 n k x k \sum_{k =1}^n \frac{1}{x - a_k} = \frac{\displaystyle \sum_{k =1}^n k^2x^{k -1}}{\displaystyle \sum_{ k = 1}^n kx^k} Substituting x = 1 x = 1 k = 1 n 1 1 a k = k = 1 n k 2 1 k 1 k = 1 n k 1 k = 7 = n ( n + 1 ) ( 2 n + 1 ) 6 n ( n + 1 ) 2 \sum_{k =1}^n \frac{1}{1 - a_k} = \frac{\displaystyle \sum_{k =1}^n k^2\cdot1^{k -1}}{\displaystyle \sum_{ k = 1}^n k\cdot1^k} = 7 = \frac{\frac{n(n +1)(2n +1)}{6}}{\frac{n(n +1)}{2}} \Rightarrow 2 n + 1 3 = 7 n = 10 \frac{2n + 1}{3} = 7 \Rightarrow n = 10

Note .- If p ( x ) = k = 1 n k x k \displaystyle p(x) = \sum_{ k = 1}^n kx^k then x = 1 x = 1 is not a root of p ( x ) p(x) because k = 1 n k 1 k = n ( n + 1 ) 2 0 \sum_{ k = 1}^n k\cdot1^k = \frac{n(n + 1)}{2} \neq 0

Moderator note:

As hinted in the discussion below, can we use the same technique of transforming the polynomial to evaluate k = 1 n 1 ( 1 a k ) 2 \displaystyle \sum_{k=1}^n \dfrac1{(1-a_k)^2} , k = 1 n 1 ( 1 a k ) 3 \displaystyle \sum_{k=1}^n \dfrac1{(1-a_k)^3} , k = 1 n 1 ( 1 a k ) 4 \displaystyle \sum_{k=1}^n \dfrac1{(1-a_k)^4} , \ldots as well? That is an interesting and natural follow up question to ponder on!

Bonus question : After knowing that n = 10 n=10 , can we show that k = 1 10 k x k = 0 \displaystyle \sum_{k=1}^{10} k x^k = 0 has at least one negative root? Or better yet: Can we show that k = 1 10 k x k = 0 \displaystyle \sum_{k=1}^{10} k x^k = 0 has precisely one negative root?

Answer to Challenge master note.-

I think we can continue derivating (use the same technique) to evaluate k = 1 n 1 ( 1 a k ) 2 \displaystyle \sum_{k=1}^n \dfrac1{(1-a_k)^2} , k = 1 n 1 ( 1 a k ) 3 \displaystyle \sum_{k=1}^n \dfrac1{(1-a_k)^3} , k = 1 n 1 ( 1 a k ) 4 \displaystyle \sum_{k=1}^n \dfrac1{(1-a_k)^4} , \ldots ... nevertheless, the derivatives will grow in complexity,... I think we can find shortcuts(easier solutions) if we think a little bit more or other harder solutions that require more knowledge... It depends.. and this is just what I think

Bonus question .-

For doing this question easier to understand for "everybody", I'll use simple calculus...Let p ( x ) = k = 1 10 k x k \displaystyle p(x) = \sum_{k = 1}^{10} kx^{k} then p ( 1 ) = 5 > 0 p(-1) = 5 > 0 and p ( 0.5 ) < 0 p(-0.5) < 0 so effectively,applying intermediate value theorem p ( x ) p(x) has a negative root. Furthemore, p ( x ) p'(x) is a strictly increasing function with lim x p ( x ) = \displaystyle \lim_{x\to -\infty} p'(x) = - \infty and lim x p ( x ) = \displaystyle \lim_{x\to \infty} p'(x) = \infty so effectively p ( x ) p(x) has only two real solutions: x = 0 x = 0 and x x having a negative value.

Guillermo Templado - 4 years, 10 months ago

Log in to reply

Nice explanation!! Mr Guillermo

Pi Han Goh - 4 years, 10 months ago

Very neat solution! Thank youuuu

Pi Han Goh - 4 years, 10 months ago

Log in to reply

I made a small typo, I'm going to solve it

Guillermo Templado - 4 years, 10 months ago

Oh, one small thing: you should show that none of a 1 , a 2 , , a n a_1, a_2, \ldots , a_n is equal to 1.

Pi Han Goh - 4 years, 10 months ago

Log in to reply

Good reply!,I'll try to solve it...

Guillermo Templado - 4 years, 10 months ago

Log in to reply

Yep. I see that you have added the explanation, which is just remainder factor theorem . An unnatural approach is to use Descartes' rule of signs , and show that f ( x ) = k = 1 n k x k \displaystyle f(x) = \sum_{k=1}^n k x^k has no positive roots.

Pi Han Goh - 4 years, 10 months ago

Challenge Master Note : Can we evaluate k = 1 n 1 ( 1 a k ) 2 \displaystyle \sum_{k=1}^n \dfrac1{(1-a_k)^2} , k = 1 n 1 ( 1 a k ) 3 \displaystyle \sum_{k=1}^n \dfrac1{(1-a_k)^3} , k = 1 n 1 ( 1 a k ) 4 \displaystyle \sum_{k=1}^n \dfrac1{(1-a_k)^4} , \ldots as well?

Pi Han Goh - 4 years, 10 months ago

Log in to reply

I think so...But, right now, I'm not sure... I remember Otto Bretcher did a poblem with k = 1 n 1 ( 1 a k ) 2 \displaystyle \sum_{k =1}^n \frac{1}{(1 - a_k)^2} . Let me to think it..

Guillermo Templado - 4 years, 10 months ago

Log in to reply

Post that as a problem!!!! =D =D

Pi Han Goh - 4 years, 10 months ago
Calvin Lin Staff
Aug 16, 2016

Since a i a_i are the roots to the equation k x k = 0 \sum k x^k = 0 , using the transformation y = 1 1 x y = \frac{1}{1-x} or x = 1 1 y = y 1 y x = 1 - \frac{1}{y} = \frac{y-1}{y} , we get that 1 1 a i \frac{1}{1- a_i} are the roots to

k x k = 0 k ( y 1 y ) k = 0 k y n k ( y 1 ) k = 0 \sum k x^k = 0 \Leftrightarrow \sum k \left( \frac{ y-1} {y} \right)^k = 0 \Leftrightarrow \sum k y^{n-k} (y-1)^k = 0

Now, we want to find 1 1 a i \sum \frac{1}{1-a_i} , which are the sum of the roots to the equation in y y . Hence, we need to find the negative coefficient of y n 1 y^{n-1} divided by the coefficient of y n y^{n} , which is just:

k × ( k 1 ) × ( 1 ) 1 k × 1 = k 2 k = n ( n + 1 ) ( 2 n + 1 ) 6 n ( n + 1 ) 2 = 2 n + 1 3 \frac{ - \sum k \times { k \choose 1 } \times (-1)^1 } { \sum k \times 1 } = \frac{ \sum k^2 }{ \sum k } = \frac{ \frac{n(n+1)(2n+1) } {6} } { \frac{n(n+1) } { 2} } = \frac{2n+1}{3}

Very beautiful approach sir. I was thinking whether this method of substitution of y = 1 1 x y=\dfrac{1}{1-x} could also work for y = 1 ( 1 x ) n y=\dfrac{1}{(1-x)^{n}} for any n N n \in \mathbb{N} . Moreover how about y = a ( a x ) n y=\dfrac{a}{(a-x)^{n}} ?

Tapas Mazumdar - 4 years, 8 months ago

Log in to reply

Try it. What do you get?

Calvin Lin Staff - 4 years, 8 months ago

Let f ( x ) = 1 n i x i = n 1 n ( x a i ) f(x) = \sum_1^n ix^i =n\prod_1^n (x-a_i)

We know, 0 n x i = x n + 1 1 x 1 d d x 0 n x i = 0 n i x i 1 \sum_0^n x^i = \dfrac{x^{n+1}-1}{x-1}\\\dfrac{d}{dx}\sum_0^n x^i = \sum_0^n ix^{i-1}

So, f f becomes, f ( x ) = x d d x x n + 1 1 x 1 = x ( n x n + 1 ( n + 1 ) x n + 1 ) ( x 1 ) 2 \begin{aligned}f(x) &= x \dfrac{d}{dx}\dfrac{x^{n+1}-1}{x-1}\\&= \dfrac{x\left(nx^{n+1}-(n+1)x^n + 1\right)}{(x-1)^2}\end{aligned}

We know 1 1 is not a root. So, we can use limit value to get f ( 1 ) f(1) or use sum of A.P formula. f ( 1 ) = lim x 1 x ( n x n + 1 ( n + 1 ) x n + 1 ) ( x 1 ) 2 = 0 n i = n ( n + 1 ) 2 f(1) = \lim_{x\to 1}\dfrac{x\left(nx^{n+1}-(n+1)x^n + 1\right)}{(x-1)^2} = \sum_0^n i = \dfrac{n(n+1)}2

f ( 1 ) = 0 n i 2 = n ( n + 1 ) ( 2 n + 1 ) 6 f'(1) = \sum_0^n i^2 = \dfrac{n(n+1)(2n+1)}6

Using formula, 1 n 1 1 a i = f ( 1 ) f ( 1 ) \displaystyle \sum_1^n \dfrac1{1-a_i} = \dfrac{f'(1)}{f(1)} , we get 7 = 2 n + 1 3 7=\dfrac{2n+1}3

So, n = 10 \boxed{n = 10}


Proof for the above formula,

Suppose f ( x ) = k 1 n ( x a i ) f(x) = k\prod_1^n (x-a_i)

Differentiating once, f ( x ) = 1 n f ( x ) x x i = f ( x ) 1 n 1 x x i f'(x) = \sum_1^n \dfrac{f(x)}{x-x_i} =f(x)\sum_1^n \dfrac1{x-x_i}

So, 1 n 1 x a i = f ( x ) f ( x ) \boxed{\displaystyle \sum_1^n \dfrac1{x-a_i} = \dfrac{f'(x)}{f(x)}}

VERY NICEEE!!!

Pi Han Goh - 4 years, 8 months ago

Log in to reply

Thank you :D !!!

Kishore S. Shenoy - 4 years, 8 months ago
James Wilson
Dec 5, 2017

The polynomial k = 1 n k ( 1 + x ) k k = 1 n k \frac{\sum_{k=1}^n k(1+x)^k}{\sum_{k=1}^n k} has a k 1 a_k-1 as its roots. Since the constant term is 1 1 , the polynomial can be factored as ( 1 x a 1 1 ) ( 1 x a 2 1 ) . . . ( 1 x a n 1 ) \Big(1-\frac{x}{a_1-1}\Big)\Big(1-\frac{x}{a_2-1}\Big)...(1-\frac{x}{a_n-1}\Big) . If one expands this product, then she finds that the coefficient of the linear term is equal to k = 1 n 1 1 a k \sum_{k=1}^n \frac{1}{1-a_k} . (One can also note that the coefficient of the linear term is the same as the coefficient of x n 1 x^{n-1} .) This means that the coefficient of the linear term must equal 7 7 . The coefficient of the linear term also has a representation in terms of n n . If one applies the binomial theorem to ( 1 + x ) k (1+x)^k , she can determine the coefficient of the linear term as k 2 k \frac{\sum k^2}{\sum k} . (I'll leave that to the reader.) Hence, n ( n + 1 ) ( 2 n + 1 ) 6 n ( n + 1 ) 2 = 7 2 n + 1 3 = 7 n = 10 \frac{\frac{n(n+1)(2n+1)}{6}}{\frac{n(n+1)}{2}}=7\Rightarrow \frac{2n+1}{3}=7\Rightarrow n=10 .

The polynomial k = 1 n k ( 1 + x ) k k = 1 n k \frac{\sum_{k=1}^n k(1+x)^k}{\sum_{k=1}^n k} has roots 1 1 a k \frac{-1}{1-a_k} .

How do you know that this is true?

Pi Han Goh - 3 years, 6 months ago

Log in to reply

I made a typo! Let me fix that.

James Wilson - 3 years, 6 months ago

Log in to reply

I'll add some more details to my solution some time soon.

James Wilson - 3 years, 6 months ago

Ok, hopefully that makes more sense now.

James Wilson - 3 years, 6 months ago

Log in to reply

Ah got it. Nice work, Wilson !

Pi Han Goh - 3 years, 6 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...