How many polynomials P of degree 4 with real coefficients satisfy P ( x 2 ) = P ( x ) P ( − x ) ?
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.
There are a lot of Pascal Triangle numbers in the coefficients of the 10 different solutions. Is this a coincidence or is there a relationship somewhere?
Log in to reply
I guess this has to do with the decomposition in factors. For example, the solutions which are product of degree 1 polynomials will have directly Pascal triangle numbers: x 4 , x 3 ( x − 1 ) , x 2 ( x − 1 ) 2 , x ( x − 1 ) 3 and ( x − 1 ) 4 ) will give you the first five rows of Pascal triangle. Up to degree 4, the polynomial which come in the decomposition are always pretty nice, so they present only numbers which all are Pascal's triangle.
But if you would look at even higher degree polynomials, it would start to be tough to get a pattern. Just check the list of all odd cyclotomic polynomials in the wiki page. You'll see these guys are nastier than they look. For example, their coefficients are not always ± 1 . In fact it can be much higher that the degree of the polynomial
Perhaps a good example is the cyclotomic polynomial for the 1 181 895th root of unity. It satisfies the said property, yet one of its coefficient is 1 4 1 0 2 7 7 3 = 4 7 ⋅ 6 1 ⋅ 4 9 1 9 . It's the first time a coefficient exceeds the degree. Furthermore, this is (as far as I know) not a number in Pascal's triangle.
I don't know that I'd call generating and then solving a system of equations 'brute force.' We aren't checking every possible solution, rather directly applying algebra. It's a bit of work, but not very difficult.
Log in to reply
I agree, but solving this system of equations by hand is still pretty annoying (to be honest, I first used a computer, then tried to solve it by hand). As far as I know (that is, the English-speaking mathematicians I heard), "brute force" does not mean that it requires a lot of force, nor that it is very brutal, just that it's a fairly uninspired and technical procedure. But this meaning is surely very different in other communities (engineering, computer science, physics, applied maths, etc...)
Including the factor of ( − 1 ) de g P in the ∗ property may be convenient, but it doesn't change the number of solutions for any degree. (If f ( x ) is a solution to P ( x 2 ) = ± P ( x ) ⋅ P ( − x ) , then − f ( x ) is a solution to P ( x 2 ) = ∓ P ( x ) ⋅ P ( − x ) .)
So there are two linear solutions to the original equation: − x and − x + 1 . These are just the opposites of the two linear functions you used for ∗ . In general, all even degree polynomial solutions of the original equation would have leading coefficient + 1 , and all odd degree polynomial solutions would have leading coefficient − 1 .
Log in to reply
Sharifibarak1@gmail.com
Correct, I thought the leading coefficient being 1 to be so natural, that I did not consider it could be otherwise.... Habits!
Problem Loading...
Note Loading...
Set Loading...
There are (at least) two solutions. Let me call them "Factoring" and "Brute force". In both case we need to assume that the coefficients of the polynomial are reals (see PS at the end). Also both solutions do require (in my opinion, tedious) case checking. The "factoring" solution can [probably] be improved to a case-checking-free solution.
"Factoring" solution
This property is conserved by factoring the polynomials, i.e. if two polynomials have this property, then so do their product. In fact, it is more convenient to look at the property ( ∗ ) : P ( x 2 ) = P ( x ) ⋅ P ( − x ) ⋅ ( − 1 ) de g P (otherwise there are no degree 1 solutions).
Next, note that if a is a root of P , then a 2 is also a root of P . This means a 4 , a 8 , a 1 6 , … are also roots of P . Since P has finite degree, this implies that all roots are either 0 or a root of unity. Also the leading coefficient of the polynomial has to be 1. Note that roots of unity are solutions to the cyclotomic polynomials, so if we assume the coefficients of P to be integers, this would reduce the case checking a lot.
So let's start by looking at degree 1 polynomials with property ( ∗ ) . The only possible roots here are − 1 , 0 and 1 . The actual polynomials satisfying property ( ∗ ) are x and x − 1 (the only excluded member here is x + 1 ).
For degree 2, we only need to look at roots which do not occur from a degree 1 polynomial. This would mean any root of unity might pop up (because ( x − ζ ) ( x − ζ ˉ ) has real coefficients and degree 2). Let a be a n th root of unity (which is not a first, second, or third root of unity) and a root of P , then a 2 as well as the complex conjugate of a should also be roots of P . [Complex conjugates need to be there since P has real coefficients.] This makes P of degree at least 3. And implies only the cyclotomic polynomial for the third root x 2 + x + 1 may satisty ( ∗ ) . And then you may check it does indeed have the property.
For degree 3, we need to check that the fourth root does not satisfy the property. Indeed, I did not show that two polynomials which do not have the property may not result in a polynomial which does (exercise?). The fourth root has a degree two polynomial ( x 2 + 1 ), which can be pumped up to degree 3 as ( x 2 + 1 ) ( x + 1 ) . You may check it does not have the property. The next candidate is the fifth root of unity. But the same argument as above shows that a polynomial which has property ( ∗ ) and has a n th root of unity as a solution has degree at least 4.
In degree 4, it is also clear that a polynomial which has property ( ∗ ) and admits a n th root of unity with n ≥ 7 as a solution won't work: a , a 2 and a 4 as well as their conjugates form at least 5 distinct elements; so it cannot have degree 4. So there only remains the fifth and sixth root of unity. The 5th root of unity: x 4 + x 3 + x 2 + x + 1 has the property ( ∗ ) .
For the sixth root, we need to check that x 2 − x + 1 times some other polynomial do not have the property. The possible choices to multiply with are: ( x 2 − x + 1 ) ( x 2 + 1 ) and ( x 2 − x + 1 ) ( x + 1 ) 2 . It turns out the property fails for these two.
So all in all we need to build product of all the polynomials with property ( ∗ ) so that the result has degree 4. Our list of primitive polynomials is x , x − 1 , x 2 + x + 1 , x 4 + x 3 + x 2 + x + 1 . So we have: x 4 , x 3 ( x − 1 ) , x 2 ( x − 1 ) 2 , x ( x − 1 ) 3 , ( x − 1 ) 4 , ( x 2 + x + 1 ) 2 , ( x 2 + x + 1 ) x 2 , ( x 2 + x + 1 ) x ( x − 1 ) , ( x 2 + x + 1 ) ( x − 1 ) 2 , and x 4 + x 3 + x 2 + x + 1 So we have in total 1 0 solutions.
Side note: If we knew that P has integer coefficients, then we could focus on the cyclotomic polynomials (there is a list of cyclotomic polynomials on Wikipedia). For example, in degree two, we need only to check the polynomials x 2 − x + 1 , x 2 + 1 and x 2 + x + 1 . In degree 4, we need only to check x 4 + x 3 + x 2 + x + 1 , x 4 + 1 , x 4 − x 3 + x 2 − x + 1 and x 4 − x 2 + 1 . This would make the case checking easier.
Alternatively, here are two improvements which would make the case checking disappear:
Improvement 1: Show that a polynomial has property ( ∗ ) if and only each of its factors has it.
Improvement 2: Show that the cyclotomic polynomials with property ( ∗ ) are exactly the odd ones (those of the third, 5th, 7th, 9th, 11th, ... roots of unity)!
Number of polynomials with property ( ∗ ) : if the improvements above are correct, then the number of polynomials of degree n with this property is given by the power of x n in the power series expansion of 1 − x 1 ∏ i = 1 ∞ 1 − x ϕ ( 2 i + 1 ) 1 (you probably need to truncate the product before making the expansion). This gives the sequence 1,2,4,6,10,14,22,30,44,58,... which is not in the OEIS.
"Brute force" solution
Write P ( x ) = a x 4 + b x 3 + c x 2 + d x + e . Then P ( x 2 ) = a x 8 + b x 6 + c x 4 + d x 2 + e while P ( x ) ⋅ P ( − x ) = a 2 x 8 + ( − b 2 + 2 a c ) x 6 + ( c 2 − 2 b d + 2 a e ) x 4 + ( − d 2 + 2 c e ) x 2 + e 2 This yields five equations: a b c d e = a 2 = − b 2 + 2 a c = c 2 − 2 b d + 2 a e = − d 2 + 2 c e = e 2 The first equation a = a 2 has possible solutions a = 0 and a = 1 . But a = 0 would mean the polynomial does not have degree 4. So a = 1 .
The last equation e = e 2 has two possible solutions e = 0 and e = 1 . This splits the problem in two cases:
PS: there are also 6 extra complex solutions b = ( − ı ⋅ 7 + 1 ) / 2 , b = ( ı ⋅ 7 + 1 ) / 2 , b = − ( ı ⋅ 1 5 + 1 ) / 2 , b = ( ı ⋅ 1 5 − 1 ) / 2 , b = − ( ı ⋅ 7 + 1 ) / 2 , b = ( ı ⋅ 7 − 1 ) / 2 , c = − ( ı ⋅ 7 + 1 ) / 2 , c = ( ı ⋅ 7 − 1 ) / 2 , c = − 2 , c = − 2 , c = − 1 , c = − 1 , d = − 1 , d = − 1 , d = − 8 / ( ı ⋅ 1 5 + 1 ) , d = 8 / ( ı ⋅ 1 5 − 1 ) , d = − 4 / ( ı ⋅ 7 + 1 ) , d = 4 / ( ı ⋅ 7 − 1 ) , e = 0 e = 0 e = 1 e = 1 e = 1 e = 1