A shift and reduce

For how many positive integers N < 1000 N<1000 , does there exist a polynomial f ( x ) f(x) with integer coefficients such that all of the following conditions are true:

  1. The degree of f f is at most 6. 6.

  2. The sum of absolute values of coefficients of f f is at most 7. 7.

  3. The polynomial f ( x ) + N f(x)+N is reducible over the integers.

Details and assumptions

A polynomial that is reducible over the integers can be expressed as a product of two non-constant polynomials with integer coefficients.


The answer is 999.

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

Mark Hennings
Sep 16, 2013

Any integer N N betwen 1 1 and 1 2 ( 3 7 1 ) = 1093 \tfrac{1}{2}(3^7-1) = 1093 can be written uniquely in fhe form N = j = 0 6 a j 3 j N \; = \; \sum_{j=0}^6 a_j 3^j where each a j a_j is one of 1 1 , 0 0 or 1 -1 . Thus the polynomial f ( x ) = j = 0 6 a j x j f(x) \; = \; -\sum_{j=0}^6 a_j x^j has degree at most 6 6 . Since a j { 1 , 0 , 1 } a_j \in \{-1,0,1\} for all j j , it is clear that the sum of the moduli of the coefficients of f ( x ) f(x) is at most 7 7 . Finally f ( 3 ) + N = 0 f(3) + N = 0 , and hence x 3 x-3 is a factor of f ( x ) + N f(x) + N .

Provided that N 5 N \ge 5 the above decomposition for N N requires the presence of 3 2 3^2 or a higher power of 3 3 , and so this polynomial f ( x ) + N f(x) + N is at least quadratic, and hence reducible. The polynomial f ( x ) + N f(x)+N is linear or constant (so zero) when 1 N 4 1 \le N \le 4 , since 2 = 3 1 2=3-1 , 4 = 3 + 1 4=3+1 . However, putting f ( x ) = x 2 N f(x) = x^2-N enables us to find a suitable polynomial f ( x ) f(x) such that f ( x ) + N = x 2 f(x)+N=x^2 is reducible when 1 N 4 1 \le N \le 4 .

Thus the condition can be satisfied for all integers 1 N 999 1 \le N \le 999 .

Moderator note:

Nicely done! Indeed, for very small N N an ad hoc argument is required.

I'm glad it turned out that every case 1-1000 could be done by 3 3 . It would have been some ugly counting if the limits of 6 6 and 7 7 were a little bit smaller; we might have had to look at the case of breaking f ( x ) + N f(x)+N down into a quadratic and a cubic, etc.

I accidentally continued halfway through typing in my solution, but probably saved time in the long run as it is basically identical to yours (although I was going to give an algorithm for exactly how to find the breakdown for any particular N N )

Matt McNabb - 7 years, 8 months ago

Log in to reply

Of course the breakdown for N N is just the ternary expansion for N + 1 2 ( 3 7 1 ) N +\tfrac12(3^7-1) , with every digit reduced by 1 1 .

Mark Hennings - 7 years, 8 months ago
Patrick Corn
Sep 15, 2013

Every integer can be written uniquely as a linear combination of powers of 3 with coefficients 1 , 0 , 1 -1, 0, 1 . (This follows by induction using the fact that if x x is between 3 n 3^n and 3 n + 1 3^{n+1} , then exactly one of 3 n + 1 x 3^{n+1}-x and x 3 n x -3^n is less than 3 n 3^n .)

If N N is a positive integer less than 1094 1094 , we can use this to construct a polynomial f f with degree at most six such that f ( 3 ) = N f(3) = -N . So f ( x ) + N f(x) + N will have a root, hence will be reducible over the integers if its degree is at least 2.

But the only N N for which its degree is less than 2 are 0 , 1 , 2 , 3 , 4 0, 1, 2, 3, 4 , and these are handled separately by the polynomials f ( x ) = x 2 + ( N 1 ) x f(x) = -x^2 + (N-1)x .

Moderator note:

Nice solution, but with a little gap: the argument does not work for very small N , N, when the resulting polynomial has degree less than 2. 2. See Mark H.'s solution, which is very similar, but more detailed, and, in particular, notices this little issue.

Did the answer get edited after the Challenge Master note?

Daniel Chiu - 7 years, 8 months ago

Log in to reply

he fixed it

Cuong Doan - 7 years, 8 months ago

Log in to reply

Actually I had the argument for small N in there the whole time. I didn't even know I could edit my solutions.

Patrick Corn - 7 years, 8 months ago

Short and simple!

Zi Song Yeoh - 7 years, 8 months ago

so the answer is?

Intan Sartika - 7 years, 9 months ago

Log in to reply

You should be able to figure out the answer from his solution

Greg Wu - 7 years, 8 months ago

Log in to reply

999

Zi Song Yeoh - 7 years, 8 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...