Integer Sum of Powers

If A B A\neq B and A n + B n A^n + B^n is an integer for all positive integers n , n, must A A and B B both be integers?


Yes No

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

Michael Mendrin
May 16, 2018

Let A = x + y A = x + \sqrt{y} and B = x y B= x-\sqrt{y} and find out what happens with different n n . The odd powers of the radical term cancel out.

Now it does seem restrictive that A , B A, B can only have the form m ± n m \pm \sqrt{n} where m , n m, n are integers

Hint: Quadratic formula gets you part of the way. Getting rid of the 2 in the denominator can be slightly tricky.

Calvin Lin Staff - 3 years ago

Log in to reply

Ah, getting rid of 2 in the denominator is tricky, because it cannot be done!

As an example, with a = 1 + 5 2 , b = 1 5 2 a = \frac{1 + \sqrt{5}}{2}, b = \frac{ 1 - \sqrt{5} } {2} results in an integer a n + b n a^n + b^n . Guess I fooled you (and myself) again.

Calvin Lin Staff - 3 years ago

Log in to reply

We could use use a = 2 + 20 a = 2+\sqrt{20} , b = 2 20 b = 2-\sqrt{20} and not have the pesky 2 2 as the denominator.

Michael Mendrin - 3 years ago

Log in to reply

@Michael Mendrin But the phrasing used is "can only". IE It seems like you were trying to classify all potential counterexamples (which I was).

Calvin Lin Staff - 3 years ago

Log in to reply

@Calvin Lin right... right

Michael Mendrin - 3 years ago
Calvin Lin Staff
May 16, 2018

[This is not a complete solution.]

Hint: First show that a , b a, b are of the form Z ± Z 2 \frac {\mathbb{Z} \pm \sqrt{ \mathbb{Z}}}{2} . Why?
Hint: Counterexamples are pretty easy after that.


Classifying all possible counterexamples

[This part is definitely incomplete. It's just my initial thoughts.]

Claim: The counter examples are of the form p ± q 2 \frac{ p \pm \sqrt{q} } { 2} , where any odd-divisor of q q is equal to 1 ( m o d 4 ) \pmod 4 .

Let a + b = x , a 2 + b 2 = y a+b = x, a^2 + b^2 = y , then a b = x 2 y 2 , a b = 2 y x 2 ab = \frac{x^2 - y } { 2}, |a-b| = \sqrt{ 2y-x^2} and a , b = x ± 2 y x 2 2 a, b = \frac{ x \pm \sqrt{2y-x^2} } { 2} .
a n + b n = ( a n 1 + b n 1 ) ( a + b ) ( a b ) ( a n 2 + b n 2 ) a^n + b^n = ( a^{n-1} + b^{n-1} ) (a+b) - (ab) ( a^{n-2} + b^{n-2} ) .
Hence, by induction, if a + b a+b and a b ab are integers, then so are a n + b n a^n + b^n .
If a b = x 2 y 2 ab = \frac{ x^2 - y } { 2} is not an integer, then it is a half integer. In order for a n + b n a^n +b^n to always be an integer, it must always be even.

John Ross
Jun 4, 2018

The explicit formula for the Lucas numbers (just like the Fibbonacci numbers except starting with 1,3 instead of 1,1) is L ( n ) = ( 1 + 5 2 ) n + ( 1 5 2 ) n L(n)=(\frac{1+\sqrt5}{2})^n+(\frac{1-\sqrt5}{2})^n

Dong kwan Yoo
May 18, 2018

For all A, B satisfying ( A+B = m, AB = n : don't need to be integer )

Let A^n + B^n = S_n

Because A, B are solutions of " X^2 - mX + n = 0 " ,

S n+2 = m S n+1 - n S_n

So, if S 1 and S 2 are integers , all 'S_n's are integer, too.

While I understand what you're going for, unfortunately, it isn't exactly expressed in your solution.

What you currently have is "If A,B are not-integers that satisfsy A+B , AB are integers, then A^n+B^n will always be an integer".
As such, you need to demonstrate that the conditional is true (or at least offer up an explanation of why it is true).

Calvin Lin Staff - 3 years ago

Log in to reply

If A, B are solution of X^2 - mx + n = 0 : ( for example, x^2 - 3x +1 = 0 )

A^n+2 + B^n+2 = m ( A^n+1 + b^n+1) - n ( A^n + B^n )

So, if A+B and A^2 + B^2 are integers, by induction, A^n + B^n is also integer.

( for example, A+B = 3, AB = 1 : A, B : not integer,

but, A^2 + B^2 = 7 : integer,

A^3 + B^3 = 3 ( A^2 + B^2 ) - (A+B) : integer,

A^4 + B^4 = 3 ( A^3 + B^3 ) - (A^2 + B^2 ) : integer,

...

A^n + B^n : integer.

Dong kwan Yoo - 3 years ago

Log in to reply

Right, so what you're missing from the original solution is just saying

( for example, x^2 - 3x +1 = 0 )

That demonstrates that the conditional is true.

Calvin Lin Staff - 3 years ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...