Answers are grounded in this paper and cite their sources.
Let q be a prime power and let be a divisor of $q-1$. We classify the permutation polynomials of Fq of polynomial index exactly up to quasi-multiplicative equivalence G(X)≡ uF(vXᵈ)Xq-X, where u,v∈ Fq^* and (d,q-1)=1, and we count the classes. For polynomials vanishing at $0$, the classes are the orbits of an explicit finite group on vectors of values at the -th roots of unity. Through the permutation it induces on the roots of unity, each orbit lies over a double coset of the symmetric group S_ with respect to its translation subgroup and an explicit affine group, and the stabilizers are explicit. For polynomials with nonzero constant term the action is free. Burnside's lemma, Möbius inversion over the coefficient support and congruences along the cycles of affine permutations give an exact formula for the number of classes. For fixed 2 these numbers are quasi-polynomials in s=(q-1)/ of degree at most -1, with period dividing for odd and 2 for even ; for polynomials vanishing at $0$ the degree is exactly -1, and we determine the leading coefficient. For odd prime and for =4 the counts are given in closed form. For index four we also count the binomials and trinomials at each admissible coefficient support. The sums of two binomial counts that occur are expressed through quartic Jacobi sums and Frobenius traces of elliptic curves and of a curve of genus two; one of the elliptic curves is the reduction of the curve Y²=X³+8X²+32X, which has no complex multiplication over Q. The trinomial counts are expressed through quartic cyclotomic numbers and these sums. All these terms cancel when the counts with a fixed residue of the vanishing order modulo s are added.
No takes yet. Share an insight, caveat, or question.
Xiang Fan (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: