Citation: He Yanqin, Zhu Chaoxi, Chen Zhuoyu. A sum analogous to Kloosterman sum and its fourth power mean[J]. AIMS Mathematics, 2020, 5(3): 2569-2576. doi: 10.3934/math.2020168
[1] | Junfeng Cui, Li Wang . The generalized Kloosterman's sums and its fourth power mean. AIMS Mathematics, 2023, 8(11): 26590-26599. doi: 10.3934/math.20231359 |
[2] | Xiaoxue Li, Wenpeng Zhang . A note on the hybrid power mean involving the cubic Gauss sums and Kloosterman sums. AIMS Mathematics, 2022, 7(9): 16102-16111. doi: 10.3934/math.2022881 |
[3] | Jianghua Li, Xi Zhang . On the character sums analogous to high dimensional Kloosterman sums. AIMS Mathematics, 2022, 7(1): 294-305. doi: 10.3934/math.2022020 |
[4] | Jinmin Yu, Renjie Yuan, Tingting Wang . The fourth power mean value of one kind two-term exponential sums. AIMS Mathematics, 2022, 7(9): 17045-17060. doi: 10.3934/math.2022937 |
[5] | Wenjia Guo, Yuankui Ma, Tianping Zhang . New identities involving Hardy sums $S_3(h, k)$ and general Kloosterman sums. AIMS Mathematics, 2021, 6(2): 1596-1606. doi: 10.3934/math.2021095 |
[6] | Jin Zhang, Wenpeng Zhang . A certain two-term exponential sum and its fourth power means. AIMS Mathematics, 2020, 5(6): 7500-7509. doi: 10.3934/math.2020480 |
[7] | Yan Zhao, Wenpeng Zhang, Xingxing Lv . A certain new Gauss sum and its fourth power mean. AIMS Mathematics, 2020, 5(5): 5004-5011. doi: 10.3934/math.2020321 |
[8] | Wenpeng Zhang, Jiafan Zhang . The hybrid power mean of some special character sums of polynomials and two-term exponential sums modulo $ p $. AIMS Mathematics, 2021, 6(10): 10989-11004. doi: 10.3934/math.2021638 |
[9] | Xuan Wang, Li Wang, Guohui Chen . The fourth power mean of the generalized quadratic Gauss sums associated with some Dirichlet characters. AIMS Mathematics, 2024, 9(7): 17774-17783. doi: 10.3934/math.2024864 |
[10] | Wenpeng Zhang, Yuanyuan Meng . On the fourth power mean of one special two-term exponential sums. AIMS Mathematics, 2023, 8(4): 8650-8660. doi: 10.3934/math.2023434 |
Let q≥3 be an integer, χ denotes any Dirichlet character mod q. Then for any integers m and n, the generalized Kloosterman sum K(m,n,χ;q) is defined as
K(m,n,χ;q)=q∑a=1χ(a)e(ma+n¯aq), |
where, e(y)=e2πiy, and a⋅¯a≡1modq.
As is known to all, the research on the properties of K(m,n,χ;q) is a very important topic in analytic number theory, and many famous number theory problems are closely related to it. Because of this reason, many scholars have studied the properties of K(m,n,χ;q), and obtained a series of meaningful results, interested readers may refer to references [3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21]. For example, Zhang Wenpeng [3] proved the following conclusion:
Let q>2 be an odd square-full number, n be any integer with (n,q)=1. Then for any primitive character χmodq, we have the identity
q∑′m=1|q∑a=1χ(a) e(ma+n¯aq)|4=q2ϕ(q)∏pα‖q(α+1−5p−1), |
where pα‖q denotes that pα|q and pα+1∤q.
If q=p is a prime, then Zhang Wenpeng [4], Li Jianghua and Liu Yanni [5] also proved the identity
p−1∑m=1|K(m,n,χ;p)|4=p−1∑m=1|p−1∑a=1χ(a)e(ma+n¯ap)|4={2p3−3p2−3p−1if χ is the principal character mod p,3p3−8p2if χ is the Legendre symbol mod p,p2(2p−7)if χ is a non-real character mod p, |
where (n,p)=1.
In addition, Kloosterman sum has been widely used in the study of analytic number theory problem. In fact, Zhang Yitang's very important works [6] on the gaps between primes is obtained based on the sieve methods and some special mean value estimate for Kloosterman sums.
On the other hand, Lv Xingxing and Zhang Wenpeng [7] introduced a new sum, which can be regarded as a companion of the Kloosterman sum. It is defined by
H(m,n,k,χ;p)=p−1∑a=1χ(ma+n¯a)e(kaq). | (1.1) |
Then they used the analytic methods and the properties of character sums to study the hybrid power mean problem involving H(m,n,k,χ;p) and the quadratic Gauss sums, and obtained an exact computational formula for
∑χmodp|p−1∑a=1χ(a)e(ma2p)|2⋅|p−1∑b=1χ(b+¯b)e(nbp)|2. |
Shane Chern [8] also studied the mean value properties of H(m,n,r,χ;q), and obtained the identity
∑χmodp|p−1∑m=1|p−1∑a=1χ(ma+n¯a)e(kap)|2|2=(p−1)(p4−7p3+17p2−5p−25), |
where p is an odd prime, n and k are integers with (nk,p)=1.
But Shane Chern said in [8], his ultimate goal is to find the fourth power mean of H(m,n,k,χ;p). That is,
∑χmodpp−1∑m=1|p−1∑a=1χ(ma+n¯a)e(kap)|4. | (1.2) |
Unfortunately, he did not succeed, and ultimately he only proved a different form of the following:
∑χmodpp−1∑m=1|p−1∑a=1χ(ma+n¯a)e(kap)|4=(p−1)(p2−10p+37+pT(p)−TL(p)), |
where T(p) and TL(p) are defined by
T(p)=p−1∑u=1p−1∑a=1p−1∑b=1(a−1)2(u¯a−1)(ub−1)≡(b−1)2(u¯b−1)(ua−1)modpχ0((a−1)2(u¯a−1)(ub−1)) |
and
TL(p)=p−1∑u=1(p−1∑a=1((ua−1)(u¯a−1)p))2, |
χ0 denotes the principal character, and (∗p) denotes the Legendre's symbol mod p.
According to our experience, the main term in this formula is focused on T(p), which is the reason why one can not see any asymptotic properties of (1.2).
Inspired by references [7,8], we are interested in the computing problem of the following 2k-th power mean
p−1∑m=0∑χmodp|p−1∑a=1χ(a+¯a)e(map)|2k, k≥2. | (1.3) |
Although this problem looks very similar to the mean value (1.2), it is substantially different.
Of course, it seems no one to study this 2k-th power mean and we have not seen any related results so far.
In this paper, we will use the number of the solutions of some congruence equations mod p and the properties of the Legendre's symbol to study the calculating problem of (1.3) with k=2, and give a sharp asymptotic formula for it. That is, we will prove the following result:
Theorem. Let p be an odd prime with p≡3mod4, then we have the asymptotic formula
∑χmodpp−1∑m=0|p−1∑a=0χ(a+¯a)e(map)|4=3p4+O(p72⋅lnp). |
Some notes: For prime p with p≡1mod4, because the congruence equation a+¯a≡0modp has two solutions, so it is a little bit complicated, and we do not have a corresponding result. For p≡3mod4 and k=2, if one can get an exact value for the summation
p−1∑a=1p−1∑b=1(a2b2(a+b)2−4ab(a+b)2−4abp), | (1.4) |
then we can obtain an exact calculating formula for (1.3).
Whether there exists an exact calculating formula for (1.4) is an open problem.
In this section, we first give three simple lemmas, which are necessary in the proof of our main result. Of course, in order to prove these lemmas, we need some basic knowledge of elementary and analytic number theory, these content can be found in reference [1], here we do not need to repeat them. First we prove the following:
Lemma 1. For any odd prime p, we have the identity
p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1ab≡cdmodpa+b≡c+dmodp1=2p2−5p+3. |
Proof. Note that if a, b, c passed through a reduced residue system mod p, then da, db and cd also pass through a reduced residue system mod p, providing (d,p)=1. So from these properties we have
p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1ab≡cdmodpa+b≡c+dmodp1=p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1abd2≡cd2modpad+bd≡cd+dmodp1=(p−1)p−1∑a=1p−1∑b=1a+b≡ab+1modp1=(p−1)p−1∑a=1p−1∑b=1(a−1)(b−1)≡0modp1=(p−1)(2p−1∑a=11−1)=(p−1)(2p−3)=2p2−5p+3. |
This proves Lemma 1.
Lemma 2. Let p be an odd prime with p≡3mod4, then we have the asymptotic formula
p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1(a+b)2+1≡abcdmodpa+b≡c+dmodp1=p2+O(p32⋅lnp). |
Proof. Since p≡3mod4, so for any integers a and b, we have p∤((a+b)2+1). From the properties of the reduced residue system, quadratic residue and Legendre's symbol mod p we have
p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1(a+b)2+1≡abcdmodpa+b≡c+dmodp1=p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=0(a+b)2+1≡abcdmodpa+b≡c+dmodp1−p−1∑a=1p−1∑b=1p−1∑c=1(a+b)2+1≡0modpa+b≡cmodp1=p−1∑a=1p−1∑b=1p−1∑c=1(a+b)2+1≡abc(a+b−c)modp1=p−1∑a=1p−1∑b=1p−1∑c=1abc2−ab(a+b)c+(a+b)2+1≡0modp1=p−1∑a=1p−1∑b=1p−1∑c=1(ab)2c2−(ab)2(a+b)c+ab((a+b)2+1)≡0modp1=p−1∑a=1p−1∑b=1p−1∑c=1c2−ab(a+b)c+ab((a+b)2+1)≡0modp1=p−1∑a=1p−1∑b=1p−1∑c=0(2c−ab(a+b))2+4ab((a+b)2+1)≡a2b2(a+b)2modp1=p−1∑a=1p−1∑b=1p−1∑c=0c2+4ab((a+b)2+1)≡a2b2(a+b)2modp1=p−1∑a=1p−1∑b=1(1+(a2b2(a+b)2−4ab(a+b)2−4abp))=(p−1)2+p−1∑a=1p−1∑b=1(a2b6(a+1)2−4ab4(a+1)2−4ab2p)=(p−1)2+p−1∑a=1p−1∑b=1(a2(a+1)2−4a¯b2(a+1)2−4a¯b4p)=(p−1)2+p−1∑a=1p−1∑b=1(a2(a+1)2−4ab2(a+1)2−4ab4p)=(p−1)2+p−1∑a=1p−1∑b=1(−ap)((2b2+(a+1)2)2−a(a+1)2−(a+1)4p). | (2.1) |
It is clear that there exist at most three integers 1≤a≤p−1 such that p∣(a+1)2((a+1)2+a). If p∤(a+1)2((a+1)2+a), that is, a≠−1 and p∤(a+1)2+a, then (2x2+(a+1)2)2−a(a+1)2−(a+1)4 is not a complete square of an integral coefficient polynomial f(x). So from the Weil's important work [2] we have the estimate
p−1∑b=1((2b2+(a+1)2)2−a(a+1)2−(a+1)4p)≪√p⋅lnp. | (2.2) |
Then combining (2.1) and (2.2) we have the estimate
p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1(a+b)2+1≡abcdmodpa+b≡c+dmodp1=p2+O(3p)+O(p32⋅lnp)=p2+O(p32⋅lnp). |
This proves Lemma 2.
Lemma 3. Let p be a prime, then we have the identity
p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1(a+b)2+1≡abcdmodpab≡cdmodpa+b≡c+dmodp1≤4(p−1). |
Proof. It is clear that the conditions ab≡cdmodp and (a+b)2+1≡abcdmodp are equivalent to ab≡cdmodp and (a+b)2+1≡(ab)2modp, and the condition (a+b)2+1≡(ab)2modp is equivalent to (a+b+ab)(a+b−ab)≡−1modp.
Then we have
p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1(a+b)2+1≡abcdmodpab≡cdmodpa+b≡c+dmodp1=p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1(a+b+ab)(a+b−ab)≡−1modpab≡cdmodpa+b≡c+dmodp1=p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑t=1(a+b+ab)≡−tmodpa+b−ab≡1tmodpab≡cdmodpa+b≡c+dmodp1=p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑t=1(c+d)≡(a+b)≡(1t−t)/2modpcd≡ab≡(−t−1t)/2modp1 |
Notice that for any given s,t∈F∗q, the equation
{x+y=sxy=t |
has at most 2 zeros in Fq, hence we have
p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1(a+b)2+1≡abcdmodpab≡cdmodpa+b≡c+dmodp1≤4(p−1). |
This proves Lemma 3.
In this section, we use the three simple lemmas in section two to prove our main result. In fact, from Lemma 1, Lemma 2, Lemma 3, the trigonometric identity
p−1∑a=0e(map)={p, if p∣m;0, if p∤m |
and the orthogonality of characters modp
∑χmodpχ(a)={p−1, if a≡1modp;0, otherwise |
. we have
1p(p−1)∑χmodpp−1∑m=0|p−1∑a=1χ(a+¯a)e(map)|4=1pp−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1(a+¯a)(b+¯b)≡(c+¯c)(d+¯d)modpp−1∑m=0e(m(a+b−c−d)p)=p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1¯ab(a2+1)(b2+1)≡¯cd(c2+1)(d2+1)modpa+b≡c+dmodp1=p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1¯ab((ab−1)2+(a+b)2)≡¯cd((cd−1)2+(c+d)2)modpa+b≡c+dmodp1=p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1(ab−cd)((a+b)2+1−abcd)≡0modpa+b≡c+dmodp1=p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1ab≡cdmodpa+b≡c+dmodp1+p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1(a+b)2+1≡abcdmodpa+b≡c+dmodp1−p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1(a+b)2+1≡abcdmodpab≡cda+b≡c+dmodp1=2p2−5p+3+p2+O(p32⋅lnp)+O(p)=3p2+O(p32⋅lnp). |
This completes the proof of our Theorem.
The authors express their sincere thanks to the reviewer for pointing out the errors in Lemma 3, which have significantly improved the presentation of this paper.
This work is supported by the N. S. F. (11771351) and (11826205) of P. R. China.
The authors declare no conflict of interest.
[1] | T. M. Apostol, Introduction to Analytic Number Theory, Springer-Verlag, New York, 1976. |
[2] |
A. Weil, On some exponential sums, Proc. Nat. Acad. Sci. U. S. A., 34 (1948), 203-210. doi: 10.1073/pnas.34.5.203
![]() |
[3] |
W. P. Zhang, On the fourth power mean of the general Kloosterman sums, J. Number Theory, 169 (2016), 315-326. doi: 10.1016/j.jnt.2016.05.018
![]() |
[4] | W. P. Zhang, On the fourth power mean of the general Kloosterman sums, Indian J. Pure Ap. Mat., 35 (2004), 237-242. |
[5] | J. H. Li, Y. N. Liu, Some new identities involving Gauss sums and general Kloosterman sums, Acta Math. Sinica (Chinese Series), 56 (2013), 413-416. |
[6] |
Y. T. Zhang, Bounded gaps between primes, Ann. Math., 179 (2014), 1121-1174. doi: 10.4007/annals.2014.179.3.7
![]() |
[7] |
X. X. Lv, W. P. Zhang, A new hybrid power mean involving the generalized quadratic Gauss sums and sums analogous to Kloosterman sums, Lith. Math. J., 57 (2017), 359-366. doi: 10.1007/s10986-017-9366-z
![]() |
[8] | S. Chern, On the power mean of a sum analogous to the Kloosterman sum, Bull. Math. Soc. Sci. Math. Roumanie, 62 (2019), 77-92. |
[9] |
W. P. Zhang, On the fourth and sixth power mean of the classical Kloosterman sums, J. Number Theory, 131 (2011), 228-238. doi: 10.1016/j.jnt.2010.08.008
![]() |
[10] |
W. P. Zhang, S. M. Shen, A note on the fourth power mean of the generalized Kloosterman sums, J. Number Theory, 174 (2017), 419-426. doi: 10.1016/j.jnt.2016.11.020
![]() |
[11] | T. Estermann, On Kloostermann's sums, Mathematica, 8 (1961), 83-86. |
[12] |
H. D. Kloosterman, On the representation of numbers in the form ax2 +by2 +cz2 +dt2, Acta Math., 49 (1926), 407-464. doi: 10.1007/BF02564120
![]() |
[13] | H. Iwaniec, Topics in Classical Automorphic Forms, Grad. Stud. Math., 17 (1997), 61-63. |
[14] | H. Salié, Uber die Kloostermanschen Summen S (u, v; q), Math. Z., 34 (1931), 91-109. |
[15] | A. V. Malyshev, A generalization of Kloosterman sums and their estimates, (in Russian), Vestnik Leningrad University, 15 (1960), 59-75. |
[16] |
W. P. Zhang, On the general Kloosterman sums and its fourth power mean, J. Number Theory, 104 (2004), 156-161. doi: 10.1016/S0022-314X(03)00154-9
![]() |
[17] |
W. P. Zhang, X. X. Li, The fourth power mean of the general 2-dimensional Kloostermann sums mod p, Acta Math. Sinica, English Series, 33 (2017), 861-867. doi: 10.1007/s10114-016-6347-9
![]() |
[18] |
W. P. Zhang, X. X. Lv, The fourth power mean of the general 3-dimensional Kloostermann sums mod p, Acta Math. Sinica, English Series, 35 (2019), 369-377. doi: 10.1007/s10114-018-7455-5
![]() |
[19] |
Y. Ye, Identities of incomplete Kloostermann sums, Proc. Amer. Math. Soc., 127 (1999), 2591-2600. doi: 10.1090/S0002-9939-99-05037-6
![]() |
[20] |
R. A. Smith, On n-dimensional Kloostermann sums, J. Number Theory, 11 (1979), 324-343. doi: 10.1016/0022-314X(79)90006-4
![]() |
[21] |
W. Luo, Bounds for incomplete Kloostermann sums, J. Number Theory, 75 (1999), 41-46. doi: 10.1006/jnth.1998.2340
![]() |