For how many positive integers n < 1 0 0 0 can the polynomial f n ( x ) = x 1 0 0 0 + x n + 2 be written as a product of two non-constant polynomials with integer coefficients?
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.
Consider all n of the form 8 ( 2 m + 1 ) , where m is an non-negative integer. Then f 8 ( 2 m + 1 ) ( x ) = x 1 0 0 0 + x 8 ( 2 m + 1 ) + 2 = ( x 1 0 0 0 + 1 ) + ( x 8 ( 2 m + 1 ) + 1 ) . Since x 1 0 0 0 + 1 = ( x 8 ) 1 2 5 + 1 = ( x 8 + 1 ) ( x 9 9 2 − x 9 8 4 + . . . + 1 ) and x 8 ( 2 m + 1 ) + 1 = ( x 8 ) 2 m + 1 + 1 = ( x 8 + 1 ) ( ( x 8 ) 2 m − ( x 8 ) 2 m − 1 + . . . + 1 ) , then x 8 + 1 is a factor of f 8 ( 2 m + 1 ) ( x ) . Therefore, f 8 ( 2 m + 1 ) ( x ) is reducible for all non-negative integers m . Since 1 0 0 0 = 8 ( 2 ⋅ 6 2 + 1 ) , then f n ( x ) is reducible for the 6 2 positive integers n < 1 0 0 0 of the form 8 ( 2 m + 1 ) (corresponding to m = 0 , 1 , . . . , 6 1 ).
Now, we shall show that if f n ( x ) is reducible, then n must be of the form 8 ( 2 m + 1 ) .
Let r be any complex root of f n ( x ) ; so r 1 0 0 0 + r n + 2 = 0 . If ∣ r ∣ ≤ 1 , then taking the modulus of both sides of the equation − 2 = r 1 0 0 0 + r n , we have:
2 = ∣ r 1 0 0 0 + r n ∣ ≤ ∣ r 1 0 0 0 ∣ + ∣ r n ∣ = ∣ r ∣ 1 0 0 0 + ∣ r ∣ n ≤ 1 + 1 = 2
Therefore, equality holds for both the second and fourth lines. For the second line, equality holds iff r 1 0 0 0 = k r n for some positive real number k . For the fourth line, equality holds iff ∣ r ∣ = 1 . Thus, all the roots r i of f n ( x ) must satisfy either: (i) ∣ r i ∣ > 1 , or (ii) ∣ r i ∣ = 1 and r i 1 0 0 0 = k i r i n for some positive real number k i . For case (ii), r i 1 0 0 0 = k i r i n ⇒ r i 1 0 0 0 − n = k i (since r i clearly cannot be zero). This means ∣ k i ∣ = ∣ r i 1 0 0 0 − n ∣ = ∣ r i ∣ 1 0 0 0 − n = 1 1 0 0 0 − n = 1 . Since k i is a positive real number, then k i = 1 and r i 1 0 0 0 = r i n .
Now suppose that f n ( x ) is reducible. Then f n ( x ) = g n ( x ) ⋅ h n ( x ) for some non-constant polynomials g n ( x ) , h n ( x ) with integer coefficients. Since f n ( x ) is a monic polynomial, then the leading coefficients of g n ( x ) and h n ( x ) must either both be 1 or both be − 1 . Also, 2 = ∣ f n ( 0 ) ∣ = ∣ g n ( 0 ) ∣ ⋅ ∣ h n ( 0 ) ∣ . Since 2 is prime, then ( ∣ g n ( 0 ) ∣ , ∣ h n ( 0 ) ∣ ) = ( 1 , 2 ) or ( 2 , 1 ) . Without loss of generality, we let ∣ g n ( 0 ) ∣ = 1 . g n ( x ) is a non-constant polynomial, so let us consider its roots r 1 , … , r j , where j ≥ 1 . Note that these j roots are also roots of the polynomial f n ( x ) . Since the leading coefficient of g n ( x ) is ± 1 and g n ( 0 ) = ± 1 , then the product of the roots of g n ( x ) is also ± 1 . Thus, ∣ r 1 … r j ∣ = 1 . Since ∣ r 1 ∣ , . . . , ∣ r j ∣ ≥ 1 , we must have ∣ r 1 ∣ = . . . = ∣ r j ∣ = 1 . Since this is so, then r i 1 0 0 0 = r i n for all i = 1 , … , j . Take any of the roots, say r 1 since j ≥ 1 . r 1 is a root of f n ( x ) , so r 1 1 0 0 0 + r 1 n + 2 = 0 . But r 1 1 0 0 0 = r 1 n , so 2 r 1 1 0 0 0 + 2 = 0 . It follows that r 1 1 0 0 0 = r 1 n = − 1 . Let a r g ( r 1 ) be θ . Then 1 0 0 0 θ ≡ n θ ≡ a r g ( − 1 ) ≡ π ( m o d 2 π ) . So we can write 1 0 0 0 θ and n θ as ( 2 p + 1 ) π and ( 2 q + 1 ) π respectively, for some integers p and q . It follows that θ = 1 0 0 0 ( 2 p + 1 ) π = n ( 2 q + 1 ) π , and hence 2 q + 1 = 1 0 0 0 n ( 2 p + 1 ) . Since 2 q + 1 is an integer, then we have 1 0 0 0 ∣ n ( 2 p + 1 ) . So 8 ∣ n ( 2 p + 1 ) . 2 p + 1 is odd, so we must have 8 ∣ n . We now write n = 8 u for some positive integer u (since n is positive). Substituting this back into 2 q + 1 = 1 0 0 0 n ( 2 p + 1 ) , we have 2 q + 1 = 1 2 5 u ( 2 p + 1 ) ⇒ 1 2 5 ( 2 q + 1 ) = u ( 2 p + 1 ) . Clearly, the left hand side of the equation is odd, so u must be odd. So u is in the form 2 m + 1 , where m is an non-negative integer. Combining these results, we have n = 8 ( 2 m + 1 ) for some non-negative integer m .
Therefore, all polynomials f n ( x ) where n is not of the form 8 ( 2 m + 1 ) must be irreducible. So there are exactly 6 2 integers n < 1 0 0 0 such that f n ( x ) is reducible.
This solution is written together with user Anqi L. (in fact it was mostly written by her) who could not publish the solution because she already published 2 on number theory.
Let us define v 2 ( k ) as the largest multiple of 2 dividing k . This is definitely routine, motivated in particular by the constant 2 in the expression f n ( x ) = x 1 0 0 0 + x n + 2 . To further motivate the solution, one may attempt small cases , in other words for n < 1 0 and one easily gets that n = 8 works. Whence, we may conjecture that f n ( x ) is reducible iff v 2 ( 1 0 0 0 ) = v 2 ( n ) .
However, just proving this is very tedious. With a certain intuition, we can first show that we must have x 1 0 0 0 + 1 and x n + 1 sharing factors.Then we show that if x j + 1 and x k + 1 at least a factor, then v 2 ( j ) = v 2 ( k ) . So:
Step 1: f n ( x ) is reducible iff g c d ( x 1 0 0 0 + 1 , x n + 1 ) > 1 .
Proof : We will show the if part first. It is considerably trivial. Suppose g c d ( x 1 0 0 0 + 1 , x n + 1 ) = m for some expression m , since f n ( x ) = ( x 1 0 0 0 + 1 ) + ( x n + 1 ) we must have m ∣ f n ( x ) as desired.
Moving on to the only if part, let us first write f n ( x ) = P ( x ) Q ( x ) for some polynomials P ( x ) and Q ( x ) . Clearly, synthesising the expression f n ( x ) , P ( x ) and Q ( x ) are either monic or else the leading coefficient is − 1 . Also, the product of their constant terms is clearly 2 . Now this directly goes to show that one of the polynomials have a constant of magnitude 1 and we may WLOG assume that it is P ( x ) . Since we have in fact made full use of the coefficients that we are given, it is intuitive in such polynomial problems, to look at the roots . So, trivially, since the product of the roots of P ( x ) is 1 , at least 1 root of P ( x ) is ≤ 1 , and suppose that the root is r . Then we have that f n ( r ) = 0 . When we consider roots, it is also crucial sometimes to involve triangle inequality because we can then consider absolute values of roots to get a contradiction; it is especially important in proving irreducibility. So applying triangle inequality, we get:
∣ r 1 0 0 0 + r n ∣ ≤ ∣ r 1 0 0 0 ∣ + ∣ r n ∣ = ∣ r ∣ 1 0 0 0 + ∣ r ∣ n ≤ 1 1 0 0 0 + 1 n = 2
Remark that this implies: f n ( r ) ≥ 2 − ∣ r 1 0 0 0 + r n ∣ ≥ 0 The equality case that we need holds when r 1 0 0 0 = r n = − 1 , which clearly implies that r is a common root, and the polynomials have to share a common factor. □
Step 2: g c d ( x j + 1 , x k + 1 ) > 1 holds iff v 2 ( j ) = v 2 ( k ) .
Proof : We sort of mimic the proof for step 1. We will show the easier if case first (for a morale booster). Suppose v 2 ( j ) = v 2 ( k ) = l , then clearly x 2 l divides both x j + 1 and x k + 1 as desired.
Now, we show the only if part. Suppose g c d ( x j + 1 , x k + 1 ) > 1 . Then again, using a roots analysis, we know that x j + 1 and x k + 1 share a common root which we will denote as q . Obviously, q j = q k = − 1 . Now, because we want to show that they have the same exponent of 2 , we can always divide them by an odd exponent. Formalising the previous sentence, if p is the smallest number such that q p = − 1 , obviously both j and k are odd multiples of p , so v 2 ( p ) = v 2 ( j ) = v 2 ( k ) as desired. □
So in conclusion, we have shown that in order for f n ( x ) to be irreducible, we need to have v 2 ( n ) = v 2 ( 1 0 0 0 ) = 3 , or stating in words (for ease of counting later), n is an odd multiple of 2 3 = 8 . Trivially, the total number/answer is: ⌊ 2 8 1 0 0 0 ⌋ = 6 2 . (Note that we take floor because n needs to be strictly less than 1 0 0 0 . )
Sorry for the long solution. It contains our motivation, inspiration, inklings, etc. and we believe that it will be more instructive this way :)
How did you figure out in the beginning that x^1000 + x^8 + 2 is reducible?
Log in to reply
Hmm good point. I think zhang must have made an assumption. Perhaps for clarity's sake, we should just remove that sentence. I presume zhang finished everything then just added that sentence ;)
Log in to reply
It is equal to f ( x 8 ) , where f ( u ) = u 1 2 5 + u + 2 . This has a factor of u + 1 by the Remainder Theorem, and hence x 1 0 0 0 + x 8 + 2 has x 8 + 1 as a factor. A similar argument shows that x 1 0 0 0 + x n + 2 is reducible whenever n is an odd multiple of 8 , and these are the full set of reducible polynomials.
Could be streamlined, but I have to say well done nevertheless. :)
A couple little typos, e.g. early in Step 2 it should say x 2 l + 1 rather than x 2 l .
x^1000+x^n+2 if this polynomial has to be written in terms of two polynomials, then the search would be for a common factor.
Now let us consider x^n+2 for no values of n, except 1, this can be written as a product of polynomials having integer coefficient. That's because the roots of x^n+2 are of the form of 2^(1/n) exp(i (2k+1)*pi/n) where k can go from 0 to n-1.
Except n=1 , 2^(1/n) is never an integer. so considering n=1, we are left with this polynomial x^1000+x+2 . This cannot be further reduced because there is no common factor.
Similarly if we consider the roots of x^1000+k, then it will be k^(1/1000)* exp(i (2m+1) pi/1000) where m can go from 0 to 999. Now the modulus of the roots are |k^(1/1000)| * | exp(i (2m+1) pi/1000) | the last term is always 1. Hence the modulus of the roots are determined by | k^(1/1000) | . Now in order to reduce this to a polynomials having integer coefficients k must be equal to 1 or 10^(1000) .
therefore the above polynomial can be written as a sum of two individual polynomials such as (x^1000+1)+(x^n+1)
Now lets factorize x^1000+1. One thing to notice here is , 1000=2^3*5^3 .
so the factorization of x^1000+1 will yield (1+x^8) (x^32-x^24+x^16-x^8+1) (x^160-x^120+x^80-x^40+1)*(x^800-x^600+x^400-x^200+1)
so the x^n+1 must have a common factor x^8+1. This is possible only when n=8m where m is any odd number between 1 to 123. so total number of odd number that exists between this limit is 62. Hence n can have 62 values of the form 8m (m=1,3,5,7,.....)
The key thing to notice in this solution is x^n+1 can only be factorized to polynomials having integer coefficient is when n is odd. This follows from the fact that only when n is odd, we get one real root as 1.
I owe my thanks to Zhang Lulu for helping with the Latex.
Let us define v 2 ( k ) as the largest multiple of 2 dividing k . This is definitely routine, motivated in particular by the constant 2 in the expression f n ( x ) = x 1 0 0 0 + x n + 2 . To further motivate the solution, one may attempt small cases , in other words for n < 1 0 and one easily gets that n = 8 works. This, for those curious, is by the remainder theorem. Whence, we may conjecture that f n ( x ) is reducible iff v 2 ( 1 0 0 0 ) = v 2 ( n ) .
However, just proving this is very tedious. With a certain intuition, we can first show that we must have x 1 0 0 0 + 1 and x n + 1 sharing factors.Then we show that if x j + 1 and x k + 1 at least a factor, then v 2 ( j ) = v 2 ( k ) . So:
Step 1: f n ( x ) is reducible iff g c d ( x 1 0 0 0 + 1 , x n + 1 ) > 1 .
Proof : We will show the if part first. It is considerably trivial. Suppose g c d ( x 1 0 0 0 + 1 , x n + 1 ) = m for some expression m , since f n ( x ) = ( x 1 0 0 0 + 1 ) + ( x n + 1 ) we must have m ∣ f n ( x ) as desired.
Moving on to the only if part, let us first write f n ( x ) = P ( x ) Q ( x ) for some polynomials P ( x ) and Q ( x ) . Clearly, synthesising the expression f n ( x ) , P ( x ) and Q ( x ) are either monic or else the leading coefficient is − 1 . Also, the product of their constant terms is clearly 2 . Now this directly goes to show that one of the polynomials have a constant of magnitude 1 and we may WLOG assume that it is P ( x ) . Since we have in fact made full use of the coefficients that we are given, it is intuitive in such polynomial problems, to look at the roots . So, trivially, since the product of the roots of P ( x ) is 1 , at least 1 root of P ( x ) is ≤ 1 , and suppose that the root is r . Then we have that f n ( r ) = 0 . When we consider roots, it is also crucial sometimes to involve triangle inequality because we can then consider absolute values of roots to get a contradiction; it is especially important in proving irreducibility. So applying triangle inequality, we get:
∣ r 1 0 0 0 + r n ∣ ≤ ∣ r 1 0 0 0 ∣ + ∣ r n ∣ = ∣ r ∣ 1 0 0 0 + ∣ r ∣ n ≤ 1 1 0 0 0 + 1 n = 2
Remark that this implies: f n ( r ) ≥ 2 − ∣ r 1 0 0 0 + r n ∣ ≥ 0 The equality case that we need holds when r 1 0 0 0 = r n = − 1 , which clearly implies that g c d ( x 1 0 0 0 − 1 , x n − 1 ) = r . □
Step 2: g c d ( x j + 1 , x k + 1 ) > 1 holds iff v 2 ( j ) = v 2 ( k ) .
Proof : We sort of mimic the proof for step 1. We will show the easier if case first (for a morale booster). Suppose v 2 ( j ) = v 2 ( k ) = l , then clearly x 2 l + 1 divides both x j + 1 and x k + 1 as desired.
Now, we show the only if part. Suppose g c d ( x j + 1 , x k + 1 ) > 1 . Then again, using a roots analysis, we know that x j + 1 and x k + 1 share a common root which we will denote as q . Obviously, q j = q k = − 1 . Now, because we want to show that they have the same exponent of 2 , we can always divide them by an odd exponent. Formalising the previous sentence, if p is the smallest number such that q p = − 1 , obviously both j and k are odd multiples of p , so v 2 ( p ) = v 2 ( j ) = v 2 ( k ) as desired. □
We have reduced the problem to a mere counting exercise, IMHO. Because we have that f n ( x ) is irreducible ⇔ v 2 ( n ) = v 2 ( 1 0 0 0 ) = 3 . For ease of counting, note that we need only find the number of odd multiples of 2 3 = 8 . The total number is: ⌊ 2 8 1 0 0 0 ⌋ = 6 2 (Note that we take floor because n needs to be strictly less than 1 0 0 0 )
∣ r 1 0 0 0 + r n ∣ ≤ ∣ r 1 0 0 0 ∣ + ∣ r n ∣ …
The equality case that we need holds when r 1 0 0 0 = r n = − 1
That's similar to what I did, but I just used some complex number cases
Suppose f n = g ⋅ h , where g and h have integer coefficients. Because the leading terms get multiplied, the leading coefficients of g and h are ± 1 . Multiplying both polynomials by − 1 if necessary, we can assume that the leading coefficients are 1 (that is, g and h are monic). Similarly, the constant terms of g and h multiply to 2 . Possibly switching the two factors, we can assume that g ( 0 ) = ± 1 and h ( 0 ) = ± 2 .
Lemma 1. Suppose t is any complex root of f n . Then ∣ t ∣ ≥ 1 . Moreover, ∣ t ∣ = 1 if and only if t 1 0 0 0 = t n = − 1 .
Proof. Applying triangle inequality to − 2 = t 1 0 0 0 + t n , we get 2 = ∣ − 2 ∣ ≤ ∣ t 1 0 0 0 ∣ + ∣ t n ∣ = ∣ t ∣ 1 0 0 0 + ∣ t ∣ n So if ∣ t ∣ < 1 , we get a contradiction. If ∣ t ∣ = 1 , then the triangle inequality is an equality which happens if and only if the summands "line up", that is t 1 0 0 0 = t n = − 1 .
By the Vieta's formula, the product of all roots of g is ± 1 . By the above lemma, this can only happen if all of them have absolute value 1 , and in this case they all must satisfy the condition t 1 0 0 0 = t n = − 1 .
Lemma 2. For any given positive integer n < 1 0 0 0 , the system of equations t 1 0 0 0 = t n = − 1 has a solution if and only if n = 1 6 m + 8 for some integer m .
Proof. In one direction, if n = 1 6 m + 8 , denote g cd ( n , 1 0 0 0 ) by k . It is divisible by 8 , but not by 1 6 . Then consider t to be any root of x k + 1 . Since t k = − 1 and 1 0 0 0 / k and n / k are odd, t 1 0 0 0 = t n = − 1 .
In the other direction, suppose k = g cd ( n , 1 0 0 0 ) , then k = 1 0 0 0 a + n b for some integers a and b . So t k = ( − 1 ) a ⋅ ( − 1 ) b = ± 1 . If 1 0 0 0 / k is even, this would imply t 1 0 0 0 = ( ± 1 ) 1 0 0 0 / k = 1 . So 1 0 0 0 / k is odd, thus k is divisible by 8 , but not by 1 6 . Now if n / k is even, then ( − 1 ) n = ( ± 1 ) n / k = 1 , a contradiction. Therefore, n / k is odd, which implies n / 8 is odd, so n = 1 6 m + 8 .
Lemma 2, together with the discussion above it, prove that n = 1 6 m + 8 is a necessary condition for f n to be a product of two integer polynomials. In the other direction, if n = 1 6 m + 8 and k = g cd ( n , 1 0 0 0 ) , then one can take g ( x ) = x k + 1 : f n ( x ) = ( x n + 1 ) + ( x 1 0 0 0 + 1 ) = ( x k + 1 ) ( ( x n − k − x n − 2 k + . . . − x k + 1 ) + ( x 1 0 0 0 − k − . . . + 1 ) )
Finally, to count the number of n , note that 0 ≤ m ≤ 6 1 , so the answer is 6 2 .
It is interesting to note that a polynomial of the form {(x)^2^p x m}+x^n+2 can be factored as: {(x)^2^p + 1}R(x), where R(x) is a non constant polynomial, and there are (floor m/2) irreducible solutions for value of n (i.e., for (floor m/2) no. of values of n, the polynomial can be factored into 2 terms). Hence, {(x)^1000}+x^n +2 = {(x)^2^3 x 125}Q(x), and there are (floor 125/2)= 62 irreducible solutions for value of n( the polynomial can be factored into 2 non constant polynomials for 62 values of n) Note: Degree of Q(x) = 1000-8 =992.
You can find a pattern using the PARI code: for(n=1,999,print(factor(x^1000+x^k+2))). It turns out that every integer that is congruent to 8 modulo 16 is factorizable over Q[x].
Hmm nice observation, but proving it would mean otherwise
Problem Loading...
Note Loading...
Set Loading...
For an integer k , define v 2 ( k ) as the greatest integer r such that 2 r divides k . For example, v 2 ( 1 0 0 0 ) = 3 and v 2 ( 7 ) = 0 .
We claim f n ( x ) is reducible in integers if and only if v 2 ( n ) = v 2 ( 1 0 0 0 ) . We divide our proof into two parts.
Part 1 : f n ( x ) is reducible, if and only if x n + 1 and x 1 0 0 0 + 1 have a common factor.
First, we show the only if part. Suppose f n ( x ) = P ( x ) Q ( x ) , where P ( x ) and Q ( x ) are non-constant integer polynomials. First, we know that P ( x ) and Q ( x ) have leading coefficients ± 1 . Second, the product of the constant terms of P ( x ) and Q ( x ) is 2 . This means that one of the constant terms has magnitude 1 . Without loss of generality, let the constant term of P ( x ) have magnitude 1 . Then, the product of the roots of P ( x ) has magnitude 1 , so P ( x ) at least one root with magnitude less then or equal to 1 . Call this root r . Then, f n ( r ) = 0 . However, by the triangle inequality,
∣ r 1 0 0 0 + r n ∣ ≤ ∣ r 1 0 0 0 ∣ + ∣ r n ∣ = ∣ r ∣ 1 0 0 0 + ∣ r ∣ n ≤ 1 + 1 = 2
Thus, ∣ f n ( r ) ∣ ≥ 2 − ∣ r 1 0 0 0 + r n ∣ ≥ 0 . Equality holds if and only if r 1 0 0 0 = r n = − 1 . However, then r is a common root of x n + 1 and x 1 0 0 0 + 1 , so the two polynomials must have a common factor.
To show the if part, suppose x n + 1 and x 1 0 0 0 + 1 have a common factor. Then, f n ( x ) is the sum of the two polynomials so their common factor clearly divides f n ( x ) .
Part 2 : x a + 1 and x b + 1 have a common factor if and only if v 2 ( a ) = v 2 ( b ) .
First, we show the only if part. Suppose x a + 1 and x b + 1 have a common factor. Then, they have a common root, r . Then, r a = r b = − 1 Let k be the smallest integer such that r k = − 1 . Then, a and b are odd multiples of k . Then, v 2 ( k ) = v 2 ( a ) and v 2 ( k ) = v 2 ( b ) . Thus, v 2 ( a ) = v 2 ( b ) .
To show the if part, suppose v 2 ( a ) = v 2 ( b ) =q. Then, clearly x 2 q + 1 divides both x a + 1 and x b + 1 .
Thus, we have shown that f n ( x ) is reducible in integers if and only if v 2 ( n ) = v 2 ( 1 0 0 0 ) . Thus, we want to find all n under 1 0 0 0 such that v 2 ( n ) = 3 . However, this is all n under 1 0 0 0 that are odd multiples of 8 and we can easily verify that there are 6 2 of them.