Integer factorization

Find the last three digits of the sum of all positive integers n < 1000 , n<1000, such that the polynomial f n ( x ) = x 4 + n f_n(x)=x^4+n is a product of two non-constant polynomials with integer coefficients.


The answer is 392.

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.

13 solutions

Ikfi Hanif
May 20, 2014

It is obvious that f n ( x ) f_n(x) must be either in the form of ( x + a ) ( x 3 + b x 2 + c x + d ) o r ( x 2 + a x + b ) ( x 2 + c x + d ) (x+a)(x^3+bx^2+cx+d) or (x^2+ax+b)(x^2+cx+d) ,

with a , b , c a, b, c and d d are integers.

first case

f n ( x ) = ( x + a ) ( x 3 + b x 2 + c x + d ) f_n(x)=(x+a)(x^3+bx^2+cx+d) f n ( x ) = x 4 + ( a + b ) x 3 + ( a b + c ) x 2 + ( a c + d ) x + a d \iff f_n(x)=x^4+(a+b)x^3+(ab+c)x^2+(ac+d)x+ad then, a + b = a b + c = a c + d = 0 a+b=ab+c=ac+d=0

We get, a = b a=-b , c = a 2 c=a^2 , d = a 3 d=-a^3 , n = a d > 0 n=ad>0

n = a d n = a 4 0 n=ad \iff n=-a^4 \leq 0

which is contradiction to the problem.

second case

f n ( x ) = ( x 2 + a x + b ) ( x 2 + c x + d ) f_n(x)= (x^2+ax+b)(x^2+cx+d) f n ( x ) = x 4 + ( a + c ) x 3 + ( b + d + a c ) x 2 + ( b c + a d ) x + b d \iff f_n(x)=x^4+(a+c)x^3+(b+d+ac)x^2+(bc+ad)x+bd then, a + c = b + d + a c = b c + a d = 0 a+c=b+d+ac=bc+ad=0

We get, a = c a=-c , n = b d > 0 n=bd>0 , b + d = a 2 a 0 b+d=a^2 \iff a \neq 0 , b c + a d = c ( b d ) = 0 b = d bc+ad=c(b-d)=0 \iff b=d

2 b = a 2 2b=a^2

We get, b = 2 , 8 , 18 , 32 , b=2, 8, 18, 32, \ldots

n = b 2 < 1000 n=b^2<1000

So n = 4 , 64 , 324 n=4, 64, 324

are the only solutions. So, the result occurs.

The explanation why a 0 a\neq 0 should have been included. This was the only essentially complete solution.

Common mistakes:
1. Missing various cases. Several students forgot to consider the linear case. 2. Assuming that all the coefficients are positive integers (after assigning some signs arbitrarily).

Calvin Lin Staff - 7 years ago
Jubayer Nirjhor
Dec 23, 2013

We set n = 4 y 4 n=4y^4 . By Sophie-Germain identity, we have:

f n ( x ) = x 4 + n = x 4 + 4 y 4 = ( x 2 + 2 x y + 2 y 2 ) ( x 2 2 x y + 2 y 2 ) f_n (x) = x^4+n = x^4+4y^4 = \left(x^2+2xy+2y^2\right)\left(x^2-2xy+2y^2\right)

And it factorizes as two non-constant polynomials with integer coefficients. Now note that:

4 3 4 = 324 < 1000 < 1024 = 4 4 4 4\cdot 3^4 = 324 < 1000 < 1024 =4\cdot 4^4

Hence the answer is simply:

i = 1 3 4 i 4 = 392 \sum_{i=1}^{3} 4\cdot i^4 = \fbox{392}

Can you show that x 4 + n x^4+n can only be factorised when n n is in the form 4 y 4 4y^4 ?

minimario minimario - 7 years, 5 months ago

Log in to reply

An algebraic approach/outline: It's obvious that f f can only be factored into two quadratics(you can't have any rational roots, thus linear,because f ( x ) > 0 f(x)>0 )

Consider the roots of f ( x ) = x 4 + n f(x)=x^4+n or simply x = n 1 4 ( 1 ) 1 4 x=n^{\frac {1}{4}}*(-1)^{\frac {1}{4}} =

1) n 1 4 2 2 ± n 1 4 2 2 i n^{\frac {1}{4}}\frac {\sqrt {2}}{2}\pm n^{\frac {1}{4}}\frac {\sqrt {2}}{2}i

2) n 1 4 2 2 ± n 1 4 2 2 i -n^{\frac {1}{4}}\frac {\sqrt {2}}{2}\pm n^{\frac {1}{4}}\frac {\sqrt {2}}{2}i

Since roots come in as conjugate pairs, we know that 1),2) are roots of the two quadratics: x 2 ± ( 4 n ) 1 4 x + n x^2\pm (4n)^{\frac {1}{4}}x+\sqrt {n}

For it to have integer coefficients, we need n = k 4 4 n=\frac {k^4}{4} for some integer k k .

Xuming Liang - 7 years, 5 months ago

Log in to reply

sad faced....I probably over complicated this....

Xuming Liang - 7 years, 5 months ago
Patrick Corn
May 20, 2014

The polynomial can't have a real root, so both factors would have to be quadratic. WLOG they're monic. After looking at the coefficients of x 3 x^3 and x x , it has to be x 4 + n = ( x 2 a x + b ) ( x 2 + a x + b ) . x^4+n = (x^2-ax+b)(x^2+ax+b). So b 2 = n b^2 = n and a 2 = 2 b a^2 = 2b , so a a and b b are both even and n = 4 ( a / 2 ) 4 . n = 4(a/2)^4. Since n < 1000 , n < 1000, there are only three solutions, namely 4 1 4 , 4 2 4 , 4 3 4 . 4 \cdot 1^4, 4 \cdot 2^4, 4 \cdot 3^4. The sum of these is 392 392 .

"After looking at the coefficients of x 3 x^3 and x x , it has to be x 4 + n = ( x 2 a x + b ) ( x 2 + a x + b ) . x^4+n = (x^2-ax+b)(x^2+ax+b). " Not really. There is a possibility of a=c=0, b=-d, which can only be ruled out by positivity of n.

Calvin Lin Staff - 7 years ago
Low Boon
May 20, 2014

x^4+4y^4 is reducible.

Hence n = 4y^4 <1000. y= 1,2,3 n = 4, 64, 324

Not explained how it is reducible. Not proven that this is the only case possible.

Calvin Lin Staff - 7 years ago
Lester Ilao
May 20, 2014

In order to find n n , we need to find its non-constant factors with integer coefficients. Since x 4 + n x^4+n , where n n is a positive integer, has no middle term we can assume that the middle term with variable x 2 x^2 has been cancelled. By looking for two polynomials which will yield to the expression x 4 + n x^4+n , we get ( x 2 + 2 a + 2 a 2 ) (x^2+2a+2a^2) ( x 2 2 a + 2 a 2 ) (x^2-2a+2a^2) . Simplifying the expression we get: x 4 + 4 a 4 x^4+4a^4 . Therefore n = 4 a 4 n=4a^4 where a a is any positive integer.

When a = 1 a=1 , n = 4 n=4 .

When a = 2 a=2 , n = 64 n=64 .

When a = 3 a=3 , n = 324 n=324 .

When a = 4 a=4 , n = 1024 n=1024 . Since n < 1000 n<1000 , 1024 is not a solution and as a a gets bigger so as the value of n n , therefore, we will only consider a = 1 , 2 , 3 a=1,2,3 .

Now we have the values of n n which are 4, 64, 324.

4 + 64 + 324 = 392 4+64+324=392

" By looking for two polynomials which will yield to the expression x 4 + n x^4+n , we get ( x 2 + 2 a + 2 a 2 ) (x^2+2a+2a^2) ( x 2 2 a + 2 a 2 ) (x^2-2a+2a^2) . Simplifying the expression we get: x 4 + 4 a 4 x^4+4a^4 . Therefore n = 4 a 4 n=4a^4 where a a is any positive integer." The fact that this condition is necessary is not really proven.

Calvin Lin Staff - 7 years ago
Ian Mana
May 20, 2014

If f n ( x ) = x 4 + n f_n (x)=x^4 +n is composed of two non-constant polynomials with integer coefficients, this means that it is composed of two Quadratic Equation.

Lets say the two equation is g ( x ) = x 2 + a x + b g(x)=x^2+ax+b and h ( x ) = x 2 + c x + d h(x)=x^2+cx+d . If we multiply the two equation, g ( x ) g(x) and h ( x ) h(x) it will make an equation which is f n ( x ) = x 4 + ( a + c ) x 3 + ( b + d + a c ) x 2 + ( a d + b c ) x + b d f_n (x)=x^4+(a+c)x^3+(b+d+ac)x^2+(ad+bc)x+bd and also a + c = 0 a+c=0 b + d + a c = 0 b+d+ac=0 a d + b c = 0 ad+bc=0 b d = n bd=n .

by a + c = 0 a+c=0 we can conclude that a = c a=-c . By substituting a = c a=-c to a d + b c = 0 ad+bc=0 we will have, b c d c = c ( b d ) = 0 bc-dc=c(b-d)=0 and by that we will have 2 possible outcomes b d = 0 b-d=0 and c = 0 c=0 .

By the first possible outcome b d = 0 b-d=0 we can conclude that b = d b=d and b d = b 2 = n bd=b^2=n and n < 1000 n<1000 so b < 31 b<31 . if we subtitute b = d b=d and a = c a=-c to b + d + a c = 0 b+d+ac=0 , we will have 2 b c 2 = 0 2b-c^2=0 , so c = s q r t 2 b c=sqrt{2b} this means 2 b 2b is a perfect square to make c c an integer. The only values for b b is 2 , 8 , 18 {2,8,18} which give n n have the values 4 , 64 , 324 {4,64,324}

By the second possible outcome c = 0 c=0 , if substituted to b + d + a c = 0 b+d+ac=0 we will have b + d = 0 b+d=0 so b = d b=-d but it cannot be because n must be positive and by b = d b=-d this makes b d bd to be negative so we disregard c = 0 c=0 as a possible out come.

So the only values for n n are 4 , 64 , 324 {4,64,324} , the sum of them is 4 + 64 + 324 = 392 4+64+324 = 392 . The last three digits is 392 392

" this means that it is composed of two Quadratic Equation." Not equations, but polynomials; also might be degree one and three. Good otherwise.

Calvin Lin Staff - 7 years ago
Sambit Senapati
May 20, 2014

According to the question, f(x)=g(x)*h(x) where g(x) and h(x) are non-constant polynomials.

It is obvious that the leading coefficients of g(x) and h(x) must be 1 or -1 each.

If for some n x 4 + n x^4+n can be expressed as a product of non-constant polynomials g(x) and h(x) with leading coefficients -1 then x 4 + n x^4+n can also be expressed as a product of 2 non-constant polynomials having leading coefficients 1(This can be achieved by simply multiplying g(x) and h(x) each by -1).

So, we are left with 2 cases.

C a s e 1 Case 1

Let x 4 + n = ( x + a ) ( x 3 + b x 2 + c x + n / a ) x^4+n=(x+a)(x^3+bx^2+cx+n/a) where a,b,c are integers and a n a|n .

Comparing coefficients on both sides we can see that this is not possible for +ve integer n.

C a s e 2 Case 2

Let x 4 + n = ( x 2 + a x + b ) ( x 2 + c x + n / b ) x^4+n=(x^2+ax+b)(x^2+cx+n/b) where a,b,c are integers and b|n.

Comparing coefficients we get b = a 2 2 b=\frac{a^2}{2} c = a c=-a n = a 4 4 n=\frac{a^4}{4}

This means n must be even.

Setting a 4 4 < 1000 \frac{a^4}{4}<1000 we get a=2,4,6.

Therefore the required answer is 392 \boxed{392} .

"This means n must be even." I think, "a must be even" is meant. It looks like the case a=c=0 (ultimately to be discarded) was missed. The details of the argument on comparing coefficients are missing.

Calvin Lin Staff - 7 years ago
Kevin Chang
May 20, 2014

We can express x^4+n as (x^4+2 sqrt(n) x^2+n)-2 sqrt(n) x^2=(x^2+sqrt(n))^2-(sqrt 4 x)^2=(x^2+sqrt 4 x+sqrt(n)(x^2-sqrt 4 *x+sqrt(n)). For the coefficients of these factors to be integers, we must have sqrt(n) be an integer, so n must be a perfect square. We let n=k^2. The fourth root of 4n=4k^2 is equal to sqrt(2k). Thus 2k must be a perfect square as well. We let k=2m^2 and have n=4m^4. When m=1, 2, 3, 4, we have n=4, 64, 324, 1024, respectively. Thus, the answer is 4+64+324=392.

It is not proven that other factorizations are impossible.

Calvin Lin Staff - 7 years ago
Danny He
May 20, 2014

For a number k 4 c , f k c 4 ( x ) = x 4 + k c 4 \sqrt[4]{k}c, \; f_{kc^4}(x) = x^4 + kc^4

x 4 + k c 4 = ( x 2 + a x c + k c 2 ) ( x 2 a x c + k c 2 ) x^4 + kc^4 = (x^2+axc+\sqrt{k}c^2)(x^2-axc+\sqrt{k}c^2)

From this we see that a 2 = 2 k a^2 = 2\sqrt{k}

k c 4 Z + { k , c } Z + s o a Z + kc^4 \in \mathbb{Z}^+ \Rightarrow \left \{k,c \right \} \subset \mathbb{Z}^+ \; so \; a \in \mathbb{Z}^+

k c 4 < 1000 k < 1000 k 31 kc^4 < 1000 \Rightarrow k<1000 \Rightarrow \sqrt{k} \leq 31

a 2 62 a n d 2 k = a 2 , a \therefore a^2 \leq 62 \; and \; \because 2\sqrt{k} = a^2, \; a is even.

*The square of any even number is even and the square of any odd number is odd, hence all even square numbers are squares of even numbers.

a { 2 , 4 , 6 } a 2 2 { 2 , 8 , 18 } a \in \left \{2,4,6 \right \} \Rightarrow \frac {a^2}{2} \in \left \{2,8,18 \right \}

So k { 2 , 8 , 18 } \sqrt{k} \in \left \{2,8,18 \right \} hence k { 4 , 64 , 324 } k \in \left \{4,64,324 \right \}

c = 1 n = k , n { 4 , 64 , 324 } c =1 \Rightarrow n=k,\; n \in \left \{4,64,324 \right \}

c = 2 n = 16 k , n { 64 , 1024 , 5184 } c = 2 \Rightarrow n = 16k, \; n \in \left \{ 64, 1024, 5184 \right \}

c = 3 n = 81 k , n { 324 , 5184 , 26244 } c = 3 \Rightarrow n =81k, \; n \in \left \{ 324,5184,26244 \right \}

Obviously 1 c 3 1 \leq c \leq 3 So the only possible values of n n are 4 , 64 4,64 and 324 324

4 + 64 + 324 = 392 4+64+324 =392

"For a number k 4 c , f k c 4 ( x ) = x 4 + k c 4 \sqrt[4]{k}c, \; f_{kc^4}(x) = x^4 + kc^4

x 4 + k c 4 = ( x 2 + a x c + k c 2 ) ( x 2 a x c + k c 2 ) x^4 + kc^4 = (x^2+axc+\sqrt{k}c^2)(x^2-axc+\sqrt{k}c^2)

From this we see that a 2 = 2 k a^2 = 2\sqrt{k} " Other factorizations may be possible, but were not considered.

Calvin Lin Staff - 7 years ago
Alexander Xue
Dec 24, 2013

Note that if the polynomial f f will always be greater than 0 for n > 0 n > 0 . Thus, the polynomial has no real roots, which means that the polynomial is the product of two degree 2 polynomials (product of degree 1 and degree 3 is eliminated).

Let the degree 2 polynomials be x 2 + a x + b x^2+ax+b and x 2 + c x + d x^2+cx+d . Note that these polynomials have leading coefficient 1, which is because the leading coefficient of f f is 1 and that the two polynomials must have integer coefficients. Multiplying, we obtain x 4 + ( a + c ) x 3 + ( b + a c + d ) x 2 + ( b c + a d ) x + b d x^4 + (a+c)x^3 + (b+ac+d)x^2 + (bc+ad)x + bd .

Then we obtain the equations: a + c = 0 b + a c + d = 0 b c + a d = 0 b d = n \\ a+c=0 \\ b+ac+d = 0 \\ bc+ad = 0 \\ bd=n \\ .

From the first and third equations, we can deduce that a ( b d ) = 0 a(b-d) = 0 (multiplying the first equation by b and subtracting the third equation from this new one). If a = 0 a=0 , then a + c = 0 c = 0 a+c = 0 \implies c=0 , and b + a c + d = 0 b = d b+ac+d = 0 \implies b=-d . However, this is impossible because b d = b 2 = n bd=-b^2=n , where n > 0 n > 0 . Thus b d = 0 b = d b-d = 0 \implies b=d .

Now, consider the first and second original equations. We have a + c = 0 a = c a+c=0 \implies a=-c , and that b + a c + d = 0 2 b a 2 = 0 2 b = a 2 b+ac+d = 0 \implies 2b-a^2 = 0 \implies 2b=a^2 . Finding all such b b and summing up over the squares will obtain the answer, because all the equations will be met in this process. Luckily, this is not a terribly long casework, as b = a 2 / 2 b=a^2/2 , which means a is even because b 2 b^2 is an integer. Furthermore, b 2 < 1000 b < 32 b^2<1000 \implies b<32 , so b = 2 2 / 2 , 4 2 / 2 , and 6 2 / 2 b=2^2/2, 4^2/2, \text{and } 6^2/2 ( 8 2 / 2 = 32 8^2/2 =32 ). We have b = 2 , 8 , 18 b=2, 8, 18 . Then the answer is 2 2 + 8 2 + 1 8 2 = 392 2^2+8^2+18^2 = \boxed{392} .

Bong Man
Dec 23, 2013

we can write fn(x)=x^4+n as fn(x)=x^4+4y^4 or fn(x)=x^4+(y^4)/4 for integer n<1000

for n=4.y^4

y=1 => n=4.1^4=4

y=2 => n=4.2^4=64

y=3 => n=4.3^4=324

y=4 => n=4.4^4=1024 outer limit

and

for n=(1/4)y^4

y=1 => n=(1/4)1^4=1/4 outer limit

y=2 => n=(1/4).2^4=4

y=4 => n=(1/4).4^4=64

y=4 => n=(1/4).6^4=324

y=8 => n=(1/4).8^4=1024 outer limit

so sum of all n= 4+64+324=392

Abhishek Bakshi
Dec 22, 2013

I know this is cheating but the cheating might be justified. We know that the expression (a^4 + 4 b^4) can be factorized as [(a + b)^2 + b^2] * [(a - b)^2 + b^2]. Considering this as the only factorization possible for positive n, we get n=4 b^4 for "b" belonging to positive integers. now n<1000. therefore maximum value of b is 3. so we get n = 4, 64 and 324, the sum of which is 392.

Why do you think this is cheating ?

Happy Melodies - 7 years, 5 months ago

Log in to reply

i think it is cheating because it may not be the only possible factorization.

Abhishek Bakshi - 7 years, 5 months ago

Probably because he assumed that there was no other factorization.

Akshaj Kadaveru - 7 years, 5 months ago

Yes i used the Sophie-Germain identity as well and only considered the cases when n is in the form of 4b^4. Just wondering if anyone can show that n must be of the form 4b^4. This is something that is lacking in most solutions so far?

Song Kai Tan - 7 years, 5 months ago
Sagnik Saha
Dec 23, 2013

We know that polynomials of the form x 4 + 4 y 4 x^{4} + 4y^{4} can be factored into two polynomials with integer coefficients, namely, ( x 2 + 2 x y + 2 y 2 ) ( x 2 2 x y + 2 y 2 ) (x^{2} + 2xy + 2y^{2}) (x^{2} - 2xy + 2y^{2} ) .

Thus n can be 4 × 1 4 4 \times 1^{4} , 4 × 2 4 4 \times 2^{4} and 4 × 3 4 4 \times 3^{4} as 4 × 4 4 4 \times 4^{4} exceeds 1000.

Hence the required answer is 4 + 64 + 324 4 + 64 + 324 = 392 \boxed{392}

How do you know that the polynomial n 4 + n n^4 + n of another form (not x 4 + 4 y 4 x^4 + 4y^4 ) cannot be factored into two non-constant polynomial?

Happy Melodies - 7 years, 5 months ago

Log in to reply

I believe that this is because n n has to be variable-free.We can use the Sophie-Germain identity,which is the only way to factorize a binomial starting with x 4 x^4 .

Rahul Saha - 7 years, 5 months ago

Log in to reply

Alright, I got it :) but do you mean constant-free?

Happy Melodies - 7 years, 5 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...