Abstract
In this paper, we use the estimate for trigonometric sums and the properties of the congruence equations to study the computational problem of one kind sixth power mean of the three-term exponential sums. As a conclusion, we give an exact computational formula for it.
1 Introduction
Let q ≥ 3 be a positive integer, m and n are any integers. Then, the three-term exponential sum C(m, n, k; q) can be defined as follows:
where k ≥ 3 is a given integer and e(y) = e2πiy.
Many studies have been concerned with the properties of related exponential sums; we refer the readers to [1]-[10]. A series of interesting results have been obtained. For example, Xiancun Du and Di Han [5] proved the identity that for any integers k ≥ 3, we have
where the constant C(k, p) is defined as follows:
Especially, if k = 6, then one has the following identity
Wenpeng Zhang and Di Han [11] obtained the identity
if prime p > 3 satisfying (3, p − 1) = 1.
Yahui Yu and Wenpeng Zhang [12] considered the sixth power mean of the three-term exponential sums
and proved the that for any integer k ≥ 3, one has the identity
where χ be the Dirichlet character mod p.
However, regarding the sixth power mean of the three-term exponential sums:
it seems that it has not been studied before or at least we have not found any relevant results. The problem is meaningful for it can reflect the upper bound approximation of C(m, n, k; p). It is easy to find that the mean value (1) is the best possible. Thus we believe that (2) has similar asymptotic properties to (1). In fact, we can use the analytic method and the qualities of the congruence equation to propose an exact computational formula for (2) with a special integer k = 3. That is to say, we shall prove the following identity:
Theorem
Let p > 3 be a prime with (3, p − 1) = 1. Then we have
From this theorem we may immediately deduce the following estimate:
Corollary
Let p > 3 be a prime with (3, p − 1)=1. Then for any integers m and n, we have the upper bound estimate
Let p > 3 be an odd prime. For any fixed integers h ≥ 4 and k > 3 with (k, p − 1) = 1, whether there exist two computational formulae for the mean value
are two open problems, we will further study.
2 Some simple lemmas
In this section, we will put up several simple Lemmas which are necessary to prove the theorem. Hereunder, we will use many properties of trigonometric sums and congruence equation, which can be found in references [6] and [13], so we will not repeat them here. First we have the following:
Lemma 2.1
For any prime p > 3 with (3, p − 1)= 1, we have the computational formula
Proof
For any integer n with (n, p) = 1, from [1] or [6] we know that
where is
This proves Lemma 2.1.□
Lemma 2.2
For any prime p > 3 with (3, p − 1)= 1, we have the computational formula
Proof
From the trigonometric identity
we have
On the other hand, from (3) and Lemma 2.1 we also have
From (5) and (6) we may immediately deduce
This proves Lemma 2.2.□
3 Proof of the theorem
In this section, we will give the proof of our theorem. Note that (3, p − 1) = 1, from (4) and the properties of reduced residue system mod p we have
Now we compute the values of U and V in (7) respectively. It is clear that from Lemma 2.2 we have
To compute the value of U, we note that the identity
so the condition
is equivalent to
or
or
For all integers 1 ≤ a, b, c, d, e ≤p − 1, we compute the number of the solutions in (9). We separate the solutions in (9) into three cases:
: a = 1, 2 ≤ b, c ≤ p − 1; b = 1, 2 ≤ a, c ≤ p − 1; c = 1, 2 ≤ a, b ≤ p − 1.
: a = b = 1, 2 ≤ c ≤ p − 1; a = c = 1, 2 ≤ b ≤ p − 1; b = c = 1, 2 ≤ a ≤ p − 1.
: a = b = c = 1.
In the case (C), the equation (9) becomes d + e ≡ 2 mod p and d2 + e2 ≡ 2 mod p, so d = e = 1. That is, in this case, the equation (9) has only one solution (a, b, c, d, e) = (1,1,1,1,1).
In the case (B), if a = b = 1 and 2 ≤ c ≤ p − 1, then the equation (9) becomes 2 ≤ c ≤ p − 1, d + e ≡ c + 1 mod p, d2 + e2 ≡ c2 + 1 mod p and de ≡ c mod p, or (d−1)(e−1)≡ 0 mod p and (d2−1)(e2−1) ≡ 0 mod p, 2 ≤ c ≤ p − 1. In this case, the number of the solutions of the congruence equation is 2(p−2). So in the case (B), the number of all solutions of the congruence equation (9) is 3 × 2(p−2) = 6(p−2).
It is clear that the number of the solutions of the congruence equation d + e ≡ b + c mod p, d2 + e2 ≡ b2 + c2 mod p, 1 ≤ b, c, d, e ≤ p − 1 is (p−1)(2p−3). In fact, this congruence equation is equivalent to the congruence equation d + e ≡ b + 1 mod p, d2 + e2 ≡ b2 + 1 mod p, 1 ≤ b, c, d, e ≤ p − 1. So from (B) we know that the number of the solutions is (p − 1)(2p−3). In the case (A), the congruence equation (9) becomes 2 ≤ b, c ≤ p − 1, d + e ≡ b + c mod p, d2 + e2≡ b2 + c2 mod p. So from (B) and (C) we know that the number of the solutions of the congruence equation (9) in case (A) is 3 × [(p − 1)(2p − 3)− 4(p − 2) − 1].
Combining three cases (A), (B) and (C), we deduce that the number of all solutions of the equation (9) is
That is, for any prime p with (3, p − 1) = 1, the value of U is
Combining (7), (8) and (10) we may immediately deduce the identity
This completes the proof of our theorem.
Acknowledgement
The authors would like to thank the referees for their very helpful and detailed comments, which have significantly improved the presentation of this paper.
This work is supported by N. S. F. (11371291), P. N. S. F. (2013JM1017), N. S. B. R. (2016JQ1013) and S. R. P. F. (16JK1373) of P. R. China.
References
[1] Apostol T. M., An extension of the Lehmers’ picturesque exponential sums, Math. Comp., 1993, 61, 25-28.10.1090/S0025-5718-1993-1182242-8Search in Google Scholar
[2] Berndt B. C., On Gaussian sums and other exponential sums with periodic coefficients, Duke Math. J., 1973, 40, 145-156.10.1215/S0012-7094-73-04014-3Search in Google Scholar
[3] Cochrane T. and Zheng Z., Upper bounds on a two-term exponential sums, Sci. China (Series A), 2001, 44, 1003-1015.10.1007/BF02878976Search in Google Scholar
[4] Cochrane T. and Pinner C., Using Stepanov’s method for exponential sums involving rational functions, J. Number Theory, 2006, 116, 270-292.10.1016/j.jnt.2005.04.001Search in Google Scholar
[5] Du X., Han D., On the fourth power mean of the three-term exponential sums, J. NWU, 2013, 43, 541-544.Search in Google Scholar
[6] Hua L. K., Additive prime number theory, Science Press, Beijing, 1957.Search in Google Scholar
[7] Schmidt W. M., Equations over finite fields: An elementary approach, Lecture Notes in Mathematics, Springer-Verlag, New York, 1976, 536, pp. 45.10.1007/BFb0080437Search in Google Scholar
[8] Wang T., Zhang W., On the hybrid mean value of Dedekind sums and two-term exponential sums, Front. Math. China, 2011, 6, 557-563.10.1007/s11464-011-0132-ySearch in Google Scholar
[9] Weil A., On some exponential sums, Proc. Nat. Acad. Sci. U.S.A., 1948, 34, 204-207.10.1007/978-1-4757-1705-1_48Search in Google Scholar
[10] Zhang W., On the fourth power mean of the general Kloosterman sums, Indian J. Pure and Applied Mathematics, 2004, 35, 237-242.10.1016/j.jnt.2016.05.018Search in Google Scholar
[11] Zhang W., Han D., On the sixth power mean of the two-term exponential sums, J. Number Theory, 2014, 136, 403-413.10.1016/j.jnt.2013.10.022Search in Google Scholar
[12] Yu Y., Zhang W., On the sixth power mean value of the generalized three-term exponential sums, Abstr. Appl. Anal., 2014, Article ID 474726, 1-4.10.1155/2014/474726Search in Google Scholar
[13] Apostol T. M., Introduction to Analytic Number Theory, Springer-Verlag, New York, 1976.10.1007/978-1-4757-5579-4Search in Google Scholar
© 2017 Wang and Li
This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 3.0 License.