Dickson polynomial

From HandWiki
Revision as of 16:11, 6 February 2024 by TextAI2 (talk | contribs) (url)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)

In mathematics, the Dickson polynomials, denoted Dn(x,α), form a polynomial sequence introduced by L. E. Dickson (1897). They were rediscovered by (Brewer 1961) in his study of Brewer sums and have at times, although rarely, been referred to as Brewer polynomials. Over the complex numbers, Dickson polynomials are essentially equivalent to Chebyshev polynomials with a change of variable, and, in fact, Dickson polynomials are sometimes called Chebyshev polynomials.

Dickson polynomials are generally studied over finite fields, where they sometimes may not be equivalent to Chebyshev polynomials. One of the main reasons for interest in them is that for fixed α, they give many examples of permutation polynomials; polynomials acting as permutations of finite fields.

Definition

First kind

For integer n > 0 and α in a commutative ring R with identity (often chosen to be the finite field Fq = GF(q)) the Dickson polynomials (of the first kind) over R are given by[1]

[math]\displaystyle{ D_n(x,\alpha)=\sum_{i=0}^{\left\lfloor \frac{n}{2} \right\rfloor}\frac{n}{n-i} \binom{n-i}{i} (-\alpha)^i x^{n-2i} \,. }[/math]

The first few Dickson polynomials are

[math]\displaystyle{ \begin{align} D_1(x,\alpha) &= x \\ D_2(x,\alpha) &= x^2 - 2\alpha \\ D_3(x,\alpha) &= x^3 - 3x\alpha \\ D_4(x,\alpha) &= x^4 - 4x^2\alpha + 2\alpha^2 \\ D_5(x,\alpha) &= x^5 - 5x^3\alpha + 5x\alpha^2 \,. \end{align} }[/math]

They may also be generated by the recurrence relation for n ≥ 2,

[math]\displaystyle{ D_n(x,\alpha) = xD_{n-1}(x,\alpha)-\alpha D_{n-2}(x,\alpha) \,, }[/math]

with the initial conditions D0(x,α) = 2 and D1(x,α) = x.

The coefficients are given at several places in the OEIS[2][3][4][5] with minute differences for the first two terms.

Second kind

The Dickson polynomials of the second kind, En(x,α), are defined by

[math]\displaystyle{ E_n(x,\alpha)=\sum_{i=0}^{\left\lfloor \frac{n}{2} \right\rfloor}\binom{n-i}{i} (-\alpha)^i x^{n-2i}. }[/math]

They have not been studied much, and have properties similar to those of Dickson polynomials of the first kind. The first few Dickson polynomials of the second kind are

[math]\displaystyle{ \begin{align} E_0(x,\alpha) &= 1 \\ E_1(x,\alpha) &= x \\ E_2(x,\alpha) &= x^2 - \alpha \\ E_3(x,\alpha) &= x^3 - 2x\alpha \\ E_4(x,\alpha) &= x^4 - 3x^2\alpha + \alpha^2 \,. \end{align} }[/math]

They may also be generated by the recurrence relation for n ≥ 2,

[math]\displaystyle{ E_n(x,\alpha) = xE_{n-1}(x,\alpha)-\alpha E_{n-2}(x,\alpha) \,, }[/math]

with the initial conditions E0(x,α) = 1 and E1(x,α) = x.

The coefficients are also given in the OEIS.[6][7]

Properties

The Dn are the unique monic polynomials satisfying the functional equation

[math]\displaystyle{ D_n\left(u + \frac{\alpha}{u},\alpha\right) = u^n + \left(\frac{\alpha}{u}\right)^n, }[/math]

where αFq and u ≠ 0 ∈ Fq2.[8]

They also satisfy a composition rule,[8]

[math]\displaystyle{ D_{mn}(x,\alpha) = D_m\bigl(D_n(x,\alpha),\alpha^n\bigr) \,= D_n\bigl(D_m(x,\alpha),\alpha^m\bigr) \, . }[/math]

The En also satisfy a functional equation[8]

[math]\displaystyle{ E_n\left(y + \frac{\alpha}{y}, \alpha\right) = \frac{y^{n+1} - \left(\frac{\alpha}{y}\right)^{n+1}}{y - \frac{\alpha}{y}} \,, }[/math]

for y ≠ 0, y2α, with αFq and yFq2.

The Dickson polynomial y = Dn is a solution of the ordinary differential equation

[math]\displaystyle{ \left(x^2-4\alpha\right)y'' + xy' - n^2y=0 \,, }[/math]

and the Dickson polynomial y = En is a solution of the differential equation

[math]\displaystyle{ \left(x^2-4\alpha\right)y'' + 3xy' - n(n+2)y=0 \,. }[/math]

Their ordinary generating functions are

[math]\displaystyle{ \begin{align} \sum_n D_n(x,\alpha)z^n &= \frac{2-xz}{1-xz+\alpha z^2} \\ \sum_n E_n(x,\alpha)z^n &= \frac{1}{1-xz+\alpha z^2} \,. \end{align} }[/math]

Links to other polynomials

By the recurrence relation above, Dickson polynomials are Lucas sequences. Specifically, for α = −1, the Dickson polynomials of the first kind are Fibonacci polynomials, and Dickson polynomials of the second kind are Lucas polynomials.

By the composition rule above, when α is idempotent, composition of Dickson polynomials of the first kind is commutative.

  • The Dickson polynomials with parameter α = 0 give monomials.

[math]\displaystyle{ D_n(x,0) = x^n \, . }[/math]

  • The Dickson polynomials with parameter α = 1 are related to Chebyshev polynomials Tn(x) = cos (n arccos x) of the first kind by[1]

[math]\displaystyle{ D_n(2x, 1) = 2T_n(x) \,. }[/math]

  • Since the Dickson polynomial Dn(x,α) can be defined over rings with additional idempotents, Dn(x,α) is often not related to a Chebyshev polynomial.

Permutation polynomials and Dickson polynomials

A permutation polynomial (for a given finite field) is one that acts as a permutation of the elements of the finite field.

The Dickson polynomial Dn(x, α) (considered as a function of x with α fixed) is a permutation polynomial for the field with q elements if and only if n is coprime to q2 − 1.[9]

(Fried 1970) proved that any integral polynomial that is a permutation polynomial for infinitely many prime fields is a composition of Dickson polynomials and linear polynomials (with rational coefficients). This assertion has become known as Schur's conjecture, although in fact Schur did not make this conjecture. Since Fried's paper contained numerous errors, a corrected account was given by (Turnwald 1995), and subsequently (Müller 1997) gave a simpler proof along the lines of an argument due to Schur.

Further, (Müller 1997) proved that any permutation polynomial over the finite field Fq whose degree is simultaneously coprime to q and less than q1/4 must be a composition of Dickson polynomials and linear polynomials.

Generalization

Dickson polynomials of both kinds over finite fields can be thought of as initial members of a sequence of generalized Dickson polynomials referred to as Dickson polynomials of the (k + 1)th kind.[10] Specifically, for α ≠ 0 ∈ Fq with q = pe for some prime p and any integers n ≥ 0 and 0 ≤ k < p, the nth Dickson polynomial of the (k + 1)th kind over Fq, denoted by Dn,k(x,α), is defined by[11]

[math]\displaystyle{ D_{0,k}(x,\alpha) = 2 - k }[/math]

and

[math]\displaystyle{ D_{n,k}(x,\alpha)=\sum_{i=0}^{\left\lfloor \frac{n}{2} \right\rfloor}\frac{n - ki}{n-i}\binom{n-i}{i} (-\alpha)^i x^{n-2i} \,. }[/math]

Dn,0(x,α) = Dn(x,α) and Dn,1(x,α) = En(x,α), showing that this definition unifies and generalizes the original polynomials of Dickson.

The significant properties of the Dickson polynomials also generalize:[12]

  • Recurrence relation: For n ≥ 2,
[math]\displaystyle{ D_{n,k}(x,\alpha) = xD_{n-1,k}(x,\alpha)-\alpha D_{n-2,k}(x,\alpha)\,, }[/math]
with the initial conditions D0,k(x,α) = 2 − k and D1,k(x,α) = x.
  • Functional equation:
[math]\displaystyle{ D_{n,k}\left(y + \alpha y^{-1}, \alpha\right) = \frac{y^{2n} +k\alpha y^{2n-2} + \cdots +k\alpha^{n-1}y^2 + \alpha^n}{y^n} = \frac{y^{2n} + {\alpha}^n}{y^n} + \left(\frac{k\alpha}{y^n} \right) \frac{y^{2n} - {\alpha}^{n-1}y^2}{y^2 - \alpha} \,, }[/math]
where y ≠ 0, y2α.
  • Generating function:
[math]\displaystyle{ \sum_{n=0}^{\infty} D_{n,k}(x,\alpha)z^n = \frac{2 - k + (k-1)xz}{1 - xz + \alpha z^2} \,. }[/math]

Notes

References