2005/08/27 | 分圆多项式[Cyclotomic Polynomial]
类别(∑〖数学〗) | 评论(0) | 阅读(204) | 发表于 09:56

A polynomial given by

Phi_n(x)==product_(k==1)^n^'(x-zeta_k),(1)

where zeta_k are the roots of unity in C given by

zeta_k=e^(2piik/n)(2)

and k runs over integers relatively prime to n. The prime may be dropped if the product is instead taken over primitive roots of unity, so that

Phi_n(x)==product_(k==1; primitive zeta_k)^n(x-zeta_k).(3)

The notation F_n(x) is also frequently encountered. Dickson et al. (1923) and Apostol (1975) give extensive bibliographies for cyclotomic polynomials.

The cyclotomic polynomial for n>1 can also be defined as

Phi_n(x)==product_(d|n)(1-x^(n/d))^(mu(d))(4)

where mu(d) is the Möbius function and the product is taken over the divisors d of n (Vardi 1991, p. 225).

Cyclotomic polynomial roots

Phi_n(x) is an integer polynomial and an irreducible polynomial with polynomial degree phi(n), where phi(n) is the totient function. Cyclotomic polynomials are returned by the Mathematica command Cyclotomic[n, x]. The roots of cyclotomic polynomials lie on the unit circle in the complex plane, as illustrated above for the first few cyclotomic polynomials.

Cyclotomic

The first few cyclotomic polynomials are

Phi_1(x)=x-1(5)
Phi_2(x)=x+1(6)
Phi_3(x)=x^2+x+1(7)
Phi_4(x)=x^2+1(8)
Phi_5(x)=x^4+x^3+x^2+x+1(9)
Phi_6(x)=x^2-x+1(10)
Phi_7(x)=x^6+x^5+x^4+x^3+x^2+x+1(11)
Phi_8(x)=x^4+1(12)
Phi_9(x)=x^6+x^3+1(13)
Phi_(10)(x)=x^4-x^3+x^2-x+1.(14)

If p is an odd prime, then

Phi_p(x)=(x^p-1)/(x-1)==x^(p-1)+x^(p-2)+...+x+1(15)
Phi_(2p)(x)=(x^(2p)-1)/(x^p-1)(x-1)/(x^2-1)==x^(p-1)-x^(p-2)+...-x+1(16)
Phi_(4p)(x)=(x^(4p)-1)/(x^(2p)-1)(x^2-1)/(x^4-1)==x^(2p-2)-x^(2p-4)+...-x^2+1(17)

(Riesel 1994, p. 306). Similarly, for p again an odd prime,

x^p-1=Phi_1(x)Phi_p(x)(18)
x^(2p)-1=Phi_1(x)Phi_2(x)Phi_p(x)Phi_(2p)(x)(19)
x^(4p)-1=Phi_1(x)Phi_2(x)Phi_4(x)Phi_p(x)Phi_(2p)(x)Phi_(4p)(x).(20)

For the first few remaining values of n,

x-1=Phi_1(x)(21)
x^2-1=Phi_1(x)Phi_2(x)(22)
x^4-1=Phi_1(x)Phi_2(x)Phi_4(x)(23)
x^8-1=Phi_1(x)Phi_2(x)Phi_4(x)Phi_8(x)(24)
x^9-1=Phi_1(x)Phi_3(x)Phi_9(x)(25)
x^(15)-1=Phi_1(x)Phi_3(x)Phi_5(x)Phi_(15)(x)(26)
x^(16)-1=Phi_1(x)Phi_2(x)Phi_4(x)Phi_8(x)Phi_(16)(x)(27)
x^(18)-1=Phi_1(x)Phi_2(x)Phi_3(x)Phi_6(x)Phi_9(x)Phi_(18)(x)(28)

(Riesel 1994, p. 307).

For p a prime relatively prime to n,

F_(np)(x)==(F_n(x^p))/(F_n(x)),(29)

but if p|n,

F_(np)(x)==F_n(x^p)(30)

(Nagell 1951, p. 160).

An explicit equation for Phi_n(x) for squarefree n is given by

Phi_n(x)==sum_(j==0)^(phi(n))a_(nj)z^(phi(n)-j),(31)

where a_(nj) is calculated using the recurrence relation

a_(nj)==-(mu(n))/jsum_(m==0)^(j-1)a_(nm)mu(GCD(n,j-m))phi(GCD(n,j-m)),(32)

with a_(n0)==1, where mu(n) is the Möbius function and GCD(m,n) is the greatest common denominator of m and n.

The polynomial x^n-1 can be factored as

x^n-1==product_(d|n)Phi_d(x).(33)

Furthermore,

x^n+1==(x^(2n)-1)/(x^n-1)==(product_(d|2n)Phi_d(x))/(product_(d|n)Phi_d(x)).(34)

The coefficients of the inverse of the cyclotomic polynomial

1/(1+x+x^2)=1-x+x^3-x^4+x^6-x^7+x^9-x^(10)+...(35)
=sum_(n==0)^(infty)c_nx^n(36)

can also be computed from

c_n=1-2|_1/3(n+2)_|+|_1/3(n+1)_|+|_1/3n_|(37)
=1-3|_1/3(n+2)_|+|_n_|(38)
=2/(sqrt(3))sin[2/3pi(n+1)],(39)

where |_x_| is the floor function.

For p prime,

Phi_p(x)==sum_(k==0)^(p-1)x^k,(40)

i.e., the coefficients are all 1. The first cyclotomic polynomial to have a coefficient other than +/-1 and 0 is Phi_(105)(x), which has coefficients of -2 for x^7 and x^(41). This is true because 105 is the first number to have three distinct odd prime factors, i.e., 105==3.5.7 (McClellan and Rader 1979, Schroeder 1997). The smallest values of n for which Phi_n(x) has one or more coefficients +/-1, +/-2, +/-3, ... are 0, 105, 385, 1365, 1785, 2805, 3135, 6545, 6545, 10465, 10465, 10465, 10465, 10465, 11305, ... (Sloane's A013594).

It appears to be true that, for m,n>1, if Phi_m(x)+Phi_n(x) factors, then the factors contain a cyclotomic polynomial. For example,

Phi_7(x)+Phi_(22)(x)==(x^2+1)(x^8-x^7+2x^4+2)==Phi_4(x)(x^8-x^7+2x^4+2).(41)

This observation has been checked up to m,n==150 (C. Nicol). If m and n are prime, then C_m+C_n is irreducible.

Migotti (1883) showed that coefficients of Phi_(pq)(x) for p and q distinct primes can be only 0, +/-1. Lam and Leung (1996) considered

Phi_(pq)(x)=sum_(k==0)^(pq-1)a_kx^k(42)

for p,q prime. Write the totient function as

phi(pq)==(p-1)(q-1)==rp+sq(43)

and let

0<=k<=(p-1)(q-1),(44)

then

1. a_k==1 iff k==ip+jq for some i in [0,r] and j in [0,s],

2. a_k==-1 iff k+pq==ip+jq for i in [r+1,q-1] and j in [s+1,p-1],

3. otherwise a_k==0.

The number of terms having a_k==1 is (r+1)(s+1), and the number of terms having a_k==-1 is (p-s-1)(q-r-1). Furthermore, assume q>p, then the middle coefficient of Phi_(pq) is (-1)^r.

Resultants of cyclotomic polynomials have been computed by Lehmer (1930), Diederichsen (1940), and Apostol (1970). It is known that rho(Phi_k(x),Phi_n(x))==1 if (m,n)==1, i.e., m and n are relatively prime (Apostol 1975). Apostol (1975) showed that for positive integers m and n and arbitrary nonzero complex numbers a and b,

rho(Phi_m(ax),Phi_n(bx))==b^(phi(m)phi(n))product_(d|n)[Phi_(m/delta)((a^d)/(b^d))]^(mu(n/d)phi(m)/phi(m/delta)),(45)

where delta==GCD(m,d) is the greatest common divisor of m and d, phi(n) is the totient function, mu(n) is the Möbius function, and the product is over the divisors of n. If m and n are distinct primes p and q, then (◇) simplifies to

rho(Phi_q(ax),Phi_p(bx))=={(a^(pq)-b^(pq))/(a^p-b^p)(a-b)/(a^q-b^q)   for a!=b; a^((p-1)(q-1))   for a==b.(46)

The following table gives the resultants rho(Phi_k(x),Phi_n(x)) (Sloane's A054372).

k\n1234567
10
220
3310
42210
551110
6134110
77111110

The numbers of 1s in successive rows of this table are given by 0, 0, 1, 1, 3, 3, 5, 4, 6, 7, 9, ... (Sloane's A075795).

The cyclotomic polynomial Phi_6(x) has the particularly nice Maclaurin series

1/(Phi_6(x))==1+x-x^3-x^4+x^6+x^7-x^9-x^(10)+...,(47)

whose coefficients 1, 0, -1, -1, 0, 1, 1, 0, -1, -1, ... (Sloane's A010892) are given by solving the recurrence equation

a(n)==a(n-1)-a(n-2)(48)

with a(0)==a(1)==1 (Wolfram 2002, p. 128), giving the explicit form

a(n)==2/3sqrt(3)sin[1/3(n+1)pi].(49)

Interestingly, any sequence b(n) satisfying the linear recurrence equation

b(n)==b(n-1)-b(n-2)(50)

can be written as

b(n)==b(0)a(n)+[b(1)-b(0)]a(n-1).(51)
0

评论Comments

日志分类
首页[1408]
∑〖数学〗[349]
Ω〖物理〗[357]
¤〖天文〗[343]
℃〖化学〗[359]