• Tidak ada hasil yang ditemukan

pages45-50. 154KB Jun 04 2011 12:06:59 AM

N/A
N/A
Protected

Academic year: 2017

Membagikan "pages45-50. 154KB Jun 04 2011 12:06:59 AM"

Copied!
6
0
0

Teks penuh

(1)

de Bordeaux 17(2005), 45–50

On some equations over finite fields

parIoulia BAOULINA

R´esum´e. Dans ce papier, suivant L. Carlitz, nous consid´erons des ´equations particuli`eres `anvariables sur le corps fini `aq´el´ements. Nous obtenons des formules explicites pour le nombre de solutions de ces ´equations, sous une certaine condition surnetq.

Abstract. In this paper, following L. Carlitz we consider some special equations ofnvariables over the finite field ofqelements. We obtain explicit formulas for the number of solutions of these equations, under a certain restriction onnandq.

1. Introduction and results

Letpbe an odd rational prime,q =ps,s1, andFqbe the finite field of

q elements. In 1954 L. Carlitz [4] proposed the problem of finding explicit formula for the number of solutions inFnq of the equation

(1.1) a1x21+· · ·+anx2n=bx1· · ·xn,

where a1, . . . , an, b ∈ F∗q and n ≥3. He obtained formulas for n = 3 and also for n = 4 and noted that for n 5 it is a difficult problem. The case n = 3, a1 = a2 = a3 = 1, b = 3 (so-called Markoff equation) also

was treated by A. Baragar [2]. In particular, he obtained explicitly the zeta-function of the corresponding hypersurface.

Letgbe a generator of the cyclic groupF∗q. It may be remarked that by multiplying (1.1) by a properly chosen element ofF∗q and also by replacing

xi by kixi for suitable ki ∈F∗q and permuting the variables, the equation (1.1) can be reduced to the form

(1.2) x21+· · ·+x2m+gx2m+1+· · ·+gx2n=cx1· · ·xn,

whereaF∗q and n/2mn. It follows from this that it is sufficient to evaluate the number of solutions of the equation (1.2).

LetNq denote the number of solutions in Fnq of the equation (1.2), and

d = gcd(n2,(q 1)/2). Recently the present author [1] obtained the explicit formulas for Nq in the cases when d= 1 and d= 2. Note that in the case whend= 1, Nq is independent of c.

(2)

Theorem 1.1. Suppose that d > 1 and there is a positive integer l such that2d|(pl+ 1) with l chosen minimal. Then

Nq =qn−1+ 1

2(1 + (−1) n)(

−1)mq(n−2)/2(q1)

+ (1)m+1(q1)n−m

2m−n X

k=0 2|k

2mn k

qk/2

+ (1)((s/2l)−1)(n−1)2n−1q(n−1)/2T,

where

T =  

d1 if m=n andc is a dth power in F∗q,

−1 if m=n andc is not a dth power in F∗q,

0 if m < n.

Theorem 1.2. Suppose that 2 | n, m = n/2, 2d∤ (n2) and there is a positive integer l such that d|(pl+ 1). Then

Nq =qn−1+ (−1)n/2q(n−2)/2(q−1) + (−1)(n−2)/2(q−1)n/2.

2. Preliminary lemmas

Letψbe a nontrivial multiplicative character onFq. We define sumT(ψ) corresponding to characterψ as

T(ψ) = 1

q1

X

x1,...,xn∈Fq

ψ(x21+· · ·+x2m+gx2m+1+· · ·+gx2n) ¯ψ(x1· · ·xn).

(we extendψto all ofFqby settingψ(0) = 0). The Gauss sum correspond-ing toψ is defined as

G(ψ) = X y∈F∗

q

ψ(y) exp(2πiTr(y)/p),

where Tr(y) =y+yp+yp2 +· · ·+yps−1

is the trace ofy from Fq toFp. In the following lemma we have an expression for Nq in terms of sums

(3)

Lemma 2.1. We have

means that the summation is taken over all nontrivial characters ψ

onFq of order dividing d.

The following lemma determines explicitly the values of certain Gauss sums.

Lemma 2.3. Let ψ be a multiplicative character of order δ > 1 on Fq.

Suppose that there is a positive integer l such that δ | (pl+ 1) and 2l | s. Then

G(ψ) = (1)(s/2l)−1+(s/2l)·((pl+1)/δ)√q.

Proof. It is analogous to that of [3, Theorem 11.6.3].

(4)

Lemma 2.4. Let ψbe a character of orderδ onFq, where δ >1 andδ |d.

Suppose that there is a positive integer l such that 2δ |(pl+ 1) and 2l |s. Then

T(ψ) =

(1)((s/2l)−1)(n−1)2n−1q(n−1)/2 if m=n,

0 if m < n.

Proof. Letλbe a character with the same conditions as in Lemma 2.2. Ifδ

is odd then the order of ¯λis equalδ and the order of ¯λη is equal 2δ. Since 2δ|(pl+ 1) and 2l|s, by Lemma 2.3, it follows that

G(¯λ) = (1)(s/2l)−1√q

(2.1)

and

G(¯λη) = (1)(s/2l)−1+(s/2l)·((pl+1)/2δ)√q.

(2.2)

If δ is even then ¯λ and ¯λη are the characters of order 2δ. Then similar reasoning yields

(2.3) G(¯λ) =G(¯λη) = (1)(s/2l)−1+(s/2l)·((pl+1)/2δ)√q.

In any case G(¯λ)2 = G(¯λη)2. Therefore, by Lemma 2.2, T(ψ) = 0 for

m < n.

Now suppose thatm=n. Since (pl+ 1)/δ is even, it follows that

(2.4) G(ψ) = (1)(s/2l)−1√q.

Ifδ is odd thenn+ ((n2)/δ) is even, and from (2.1), (2.2), (2.4) and Lemma 2.2 we obtain

T(ψ) = 1 2q (−1)

(s/2l)−1√q

·(1)((s/2l)−1)nqn/2

×h1 + (1)(s/2l)·((pl+1)/2δ)n+1(1)(s/2l)·((pl+1)/2δ)ni

= (1)((s/2l)−1)(n−1)2n−1q(n−1)/2,

and therefore lemma is established in this case.

Ifδ is even thenn is even, and (2.3), (2.4) and Lemma 2.2 imply

T(ψ) = 1 2q (−1)

(s/2l)−1√q·(1)((s/2l)−1+(s/2l)·((pl+1)/2δ))n

2nqn/2

= (1)((s/2l)−1)(n−1)2n−1q(n−1)/2.

(5)

3. Proof of the theorems

Thus, from (3.1) and the well-known relation

(6)

References

[1] I. Baoulina, On the Problem of Explicit Evaluation of the Number of Solutions of the Equationa1x2

1+· · ·+anxn2 =bx1· · ·xn in a Finite Field. In Current Trends in Number Theory, Edited by S. D. Adhikari, S. A. Katre and B. Ramakrishnan, Hindustan Book Agency, New Delhi, 2002, 27–37.

[2] A. Baragar, The Markoff Equation and Equations of Hurwitz. Ph. D. Thesis, Brown University, 1991.

[3] B. C. Berndt, R. J. Evans, K. S. Williams, Gauss and Jacobi Sums. Wiley-Interscience, New York, 1998.

[4] L. Carlitz,Certain special equations in a finite field. Monatsh. Math.58(1954), 5–12.

IouliaBaoulina

The Institute of Mathematical Sciences CIT Campus, Taramani

Chennai 600113, India

Referensi

Dokumen terkait

Pour chacun de ces algorithmes nous d´eterminons la borne sup´erieure optimale de la valeur minimale des coefficients d’approximation pris en nombres cons´ecutifs appropri´es..

On en déduit des informations sur la taille des valeurs au bord de la bande critique de fonctions L de puis- sances symétriques dans certaines sous-familles.. Dans une deux-

Nous prouvons d’abord une version du th´eor`eme de Runge valide pour des vari´et´es de dimension sup´erieu- re, g´en´eralisant une version uniforme du th´eor`eme de Runge due `..

Une application de cette formule donne une formule pour le nombre de repr´esentations d’un nombre entier positif comme ´etant la somme de douze nombres

Le r´esultat principal de cet article est une description explicite de la structure des sous-groupes de ramification du grou- pe de Galois d’un corps local de dimension 2 modulo

En combinant des concepts de th´eorie additive des nom- bres avec des r´esultats sur les d´eveloppements binaires et les s´eries partielles, nous ´etablissons de nouvelles bornes

qui est le produit de deux nombres premiers distincts et tel que le nombre de classes du n -i`eme corps cyclotomique soit ´egal ` a 1.. Nous ´etablissons certaines

Pour les syst`emes de num´eration abstraits construits sur des langages r´eguliers exponentiels (comme par exemple, ceux provenant des substitutions), nous montrons que l’ensemble