Polynomial prime

What is the smallest degree of a reducible polynomial f f with integer coefficients, such that f ( n ) \lvert f(n) \rvert is a prime number for at least 10 10 different integer values of n ? n?

Details and assumptions

A polynomial with integer coefficients is called reducibl e if it can be written as a product of two non-constant polynomials with integer coefficients.

You may refer to this List of 1000 Primes .


The answer is 8.

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

Mark Hennings
Aug 20, 2013

Suppose that f ( x ) = g ( x ) h ( x ) f(x) = g(x)h(x) where g ( x ) , h ( x ) g(x),h(x) are nonconstant polynomials with integer coefficients, and suppose that f ( x |f(x| is prime for each of ten distinct integer values n 1 , n 2 , , n 10 n_1,n_2,\ldots,n_{10} . Then g ( n j ) × h ( n j ) |g(n_j)|\times|h(n_j)| is prime for 1 j n 1 \le j \le n . Thus, for each 1 j 10 1 \le j \le 10 , either g ( n j ) = ± 1 g(n_j) = \pm1 or h ( n j ) = ± 1 h(n_j) = \pm 1 . Now

  • If p ( x ) p(x) is a polynomial with integer coefficients with three distinct integer zeros, then there exists at most one integer value n n where p ( n ) = 2 p(n)=2 .

  • If p ( x ) p(x) is a polynomial with integer coefficients with four distinct integer zeros, then there exists no value n n for which p ( n ) = 2 p(n) = 2 .

The proofs of these statements is easy. In the first case, let the zeros be a < b < c a<b<c . Then p ( x ) = ( x a ) ( x b ) ( x c ) q ( x ) p(x) = (x-a)(x-b)(x-c)q(x) for some polynomial q ( x ) q(x) with integer coefficients, and ( d a ) ( d b ) ( d c ) 2 |(d-a)(d-b)(d-c)| \ge 2 for any d a , b , c d \neq a,b,c . Moreover ( d a ) ( d b ) ( d c ) = 2 |(d-a)(d-b)(d-c)| = 2 only when either a , d , b , c a,d,b,c or a , b , d , c a,b,d,c are consecutive integers. Thus, for any collection of a , b , c a,b,c , there is at most one possible value of d d for which p ( d ) = 2 p(d)=2 .

In the second case, let the zeros of p ( x ) p(x) be a < b < c < d a<b<c<d . Then p ( x ) = ( x a ) ( x b ) ( x c ) ( x d ) q ( x ) p(x) = (x-a)(x-b)(x-c)(x-d)q(x) for some polynomial q ( x ) q(x) with integer coefficients. Since ( e a ) ( e b ) ( e c ) ( e d ) (e-a)(e-b)(e-c)(e-d) (with e e an integer) is either zero or greater than 2 2 in modulus, we see that p ( e ) p(e) cannot be 2 2 for any integer e e .

Suppose that g ( n j ) = 1 |g(n_j)|=1 a total of k k times, and h ( n j ) = 1 |h(n_j)|=1 a total of 10 k 10-k times. By labelling g ( x ) g(x) and h ( x ) h(x) suitably, we can assume that 0 k 5 0 \le k \le 5 . Then h ( x ) |h(x)| takes the value 1 1 at least five times. Thus we can find ε { 1 , 1 } \varepsilon \in\{-1,1\} such that h ( x ) h(x) takes the value ε \varepsilon at least three times. By the above results, it follows that h ( x ) h(x) takes the value ε -\varepsilon at most once. Thus, in fact, h ( x ) h(x) takes the value ε \varepsilon at least four times, and hence cannot take the value ε -\varepsilon at all. Thus h ( x ) h(x) takes the value ε \varepsilon 10 k 10-k times, and so h ( x ) ε h(x)-\varepsilon has 10 k 10-k zeros, and hence has degree at least 10 k 10-k . The minimum degree of h ( x ) h(x) is 10 k 10-k .

  1. If k = 5 k=5 then, just as in the above argument about h ( x ) h(x) , g ( x ) g(x) has degree at least 5 5 . Thus the smallest degree of f ( x ) = g ( x ) h ( x ) f(x)=g(x)h(x) is 5 + 5 = 10 5+5=10 .

  2. Suppose that k = 4 k=4 . The polynomial g ( x ) = x ( x 3 ) + 1 = x 2 3 x + 1 g(x)=x(x-3)+1=x^2-3x+1 takes the value 1 1 at x = 0 , 3 x=0,3 and the value 1 -1 at x = 1 , 2 x=1,2 . It is not possible for a linear polynomial to hit 1 1 or 1 -1 a total of four times. Thus the smallest degree of g ( x ) g(x) is two, and so the smallest degree of f ( x ) f(x) is 2 + 6 = 8 2+6=8 .

  3. Suppose that k = 3 k=3 . Then there exists ε { 1 , 1 } \varepsilon \in \{-1,1\} such that g ( x ) g(x) takes the value ε \varepsilon at least twice, and hence g ( x ) ε g(x)-\varepsilon has at least two zeros, and hence g ( x ) g(x) has degree at least two. Thus the smallest degree of f ( x ) f(x) is 2 + 7 = 9 2+7=9 .

  4. Suppose that k 2 k\le2 . Then, since g ( x ) g(x) is non constant, g ( x ) g(x) has degree at least one. Thus the smallest degree of f ( x ) f(x) is 1 + 10 k = 11 k 9 1+10-k=11-k \ge 9 .

Our best bet, therefore, is to choose g ( x ) = x 2 3 x + 1 g(x) = x^2 - 3x + 1 and hunt for a suitable degree 6 6 polynomial h ( x ) h(x) . It happens that g ( x ) g(x) is helpful, since it achieves a lot of primes. In particular g ( 3 ) = g ( 6 ) = 19 g(-3)=g(6)=19 , g ( 2 ) = g ( 5 ) = 11 g(-2)=g(5)=11 , g ( 1 ) = g ( 4 ) = 5 g(-1)=g(4)=5 are all prime. If we consider h ( x ) = 1 + ( x + 3 ) ( x + 2 ) ( x + 1 ) ( x 4 ) ( x 5 ) ( x 6 ) h(x) = 1 + (x+3)(x+2)(x+1)(x-4)(x-5)(x-6) then h ( x ) h(x) takes the value 1 1 at 3 , 2 , 1 , 4 , 5 , 6 -3,-2,-1,4,5,6 , and moreover h ( 0 ) = h ( 3 ) = 719 h(0) = h(3) = -719 , h ( 1 ) = h ( 2 ) = 1439 h(1) = h(2) = -1439 are all negative primes. Thus we deduce that f ( x ) |f(x)| is prime for x = 3 , 2 , 1 , 0 , 1 , 2 , 3 , 4 , 5 , 6 x=-3,-2,-1,0,1,2,3,4,5,6 , and we are done. Thus the smallest possible degree of f ( x ) f(x) is 8 8 .

Moderator note:

Excellent solution: complete and clear. Brilliant!

Excellent solution and presentation! Incidentally, the example we had in mind was essentially the same (up to a horizontal shift). In particular, it also used the primes 719 and 1439. I don't know if there are any smaller examples.

Alexander Borisov - 7 years, 9 months ago

Log in to reply

Smaller, no (but it depends on what you mean by "smaller"). Other, yes. If a polynomial p ( x ) p(x) has two integer zeros, then p ( x ) p(x) never attains the value 2 2 for integer x x unless its two zeros are either consecutive or have a difference of 3 3 . The only ways to hit 1 1 twice and 1 -1 twice are by achieving 1 , 1 1 , 1 1,-1-1,1 or 1 , 1 , 1 , 1 -1,1,1,-1 by consecutive integers. Thus g ( x ) = ± ( x 2 3 x + 1 ) g(x) = \pm(x^2-3x+1) or a translate thereof.

Working with g ( x ) = x 2 3 x + 1 g(x) = x^2 - 3x + 1 we need to have h ( x ) = ( x a ) ( x b ) ( x c ) ( x d ) ( x e ) ( x f ) ± 1 h(x) \; = \; (x-a)(x-b)(x-c)(x-d)(x-e)(x-f) \pm 1 where a , b , c , d , e , f a,b,c,d,e,f are distinct integers where g ( x ) g(x) is prime, and we want h ( x ) h(x) to be prime at x = 0 , 1 , 2 , 3 x=0,1,2,3 . There are lots of candidate values here, since g ( 4 ) = g ( 1 ) = 5 g ( 5 ) = g ( 2 ) = 11 g ( 6 ) = g ( 3 ) = 19 g ( 7 ) = g ( 4 ) = 29 g ( 8 ) = g ( 5 ) = 41 g ( 10 ) = g ( 7 ) = 71 g ( 11 ) = g ( 8 ) = 89 g ( 12 ) = g ( 9 ) = 109 g ( 13 ) = g ( 10 ) = 131 \begin{array}{ccc} g(4)=g(-1)=5 & g(5)=g(-2)=11 & g(6)=g(-3)=19 \\ g(7)=g(-4)=29 & g(8)=g(-5)=41 & g(10)=g(-7)=71 \\ g(11)=g(-8)=89 & g(12)=g(-9)=109 & g(13)=g(-10)=131 \end{array} to name a few.

If we define "smaller" by wanting to make the sum of the moduli of the four primes achieved by h ( x ) h(x) at 0 , 1 , 2 , 3 0,1,2,3 as small as possible, i.e. to minimize H ( h ) = h ( 0 ) + h ( 1 ) + h ( 2 ) + h ( 3 ) H(h) \;= \; |h(0)| + |h(1)| + |h(2)| + |h(3)| then the above solution is the smallest. It is clear that the value of H ( h ) H(h) is made smaller by making a , b , c , d , e , f a,b,c,d,e,f as close to 0 , 1 , 2 , 3 0,1,2,3 as possible. Thus the smallest possible values of H ( h ) H(h) occur when a , b , c , d , e , f , 0 , 1 , 2 , 3 a,b,c,d,e,f,0,1,2,3 are consecutive (in some order). Testing shows us that the smallest value of H ( h ) H(h) occurs when the consecutive order is a , b , c , 0 , 1 , 2 , 3 , d , e , f a,b,c,0,1,2,3,d,e,f (and we are in the case discussed above). Moreover, of all the possible consecutive orderings for a , b , c , d , e , f , 0 , 1 , 2 , 3 a,b,c,d,e,f,0,1,2,3 , only the a , b , c , 0 , 1 , 2 , 3 , d , e , f a,b,c,0,1,2,3,d,e,f ordering coupled with the + + option out of ± \pm for h ( x ) h(x) generates four primes at 0 , 1 , 2 , 3 0,1,2,3 .

There are other possible solutions, however. We can choose a , b , c , d , e , f a,b,c,d,e,f to be 5 , 3 , 1 , 4 , 6 , 8 -5,-3,-1,4,6,8 , and take the + + option, obtaining 2879 -2879 and 5039 -5039 as the prime values obtained by h ( x ) h(x) . Or we can choose a , b , c , d , e , f a,b,c,d,e,f to be 5 , 4 , 1 , 4 , 7 , 8 -5,-4,-1,4,7,8 , and take the - option, obtaining 4481 -4481 and 7561 -7561 instead.

Mark Hennings - 7 years, 9 months ago

Log in to reply

Thank you! I also felt that this example was the smallest, in every reasonable sense, but did not get around to prove it. Note that the construction can be generalized slightly: h ( x ) = n ( x a ) ( x b ) ( x c ) ( x d ) ( x e ) ( x f ) ± 1 h(x) = n\cdot (x-a)(x-b)(x-c)(x-d)(x-e)(x-f) \pm 1 But clearly the smallest examples must correspond to n = 1. n=1.

Alexander Borisov - 7 years, 9 months ago

Excuse me, but unless I'm delusional, your statement "If a polynomial p ( x ) p(x) has two integer zeros, then p ( x ) p(x) never attains the value 2 2 for integer x x unless its two zeros are either consecutive or have a difference of 3 3 " is incorrect; take p ( x ) = ( x 5 ) ( x 7 ) ( x 2 5 x 8 ) p(x)=(x-5)(x-7)(x^2-5x-8) for instance. Its two integer zeros are 5 5 and 7 7 and yet p ( 6 ) = 2 p(6)=2 .

Ryan Soedjak - 7 years, 9 months ago

Log in to reply

@Ryan Soedjak Your example is correct. There was a word omitted from my previous post. It should have said that if a polynomial takes the value 0 0 twice, then it cannot take the value 2 2 twice unless its zeros are either consecutive or three apart from each other. The fact the only ways to hit 1 1 and 1 -1 twice each are 1 , 1 , 1 , 1 1,-1,-1,1 and 1 , 1 , 1 , 1 -1,1,1,-1 is then an immediate consequence.

Mark Hennings - 7 years, 9 months ago

Well done. I had established that g(x) = x(x-3)+1 and h(x) = 1+(x+a)...(x+f) was the smallest possible format, but I couldn't actually find values for a,b,c,d,e,f that gave primes in all cases. I did find an exact solution with degree 9 too.

Matt McNabb - 7 years, 9 months ago

I lost it at " By the above results, it follows that h takes the value - epsilon at most once." I can get why h takes this value at most twice, but why at most once?

Gabriel Romon - 7 years, 4 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...