The main aim of this article is using the elementary method and the number of the solutions of some congruence equations modulo an odd prime p, to study the calculating problem of the sixth power mean of one kind generalized two-term exponential sums, and give a sharp asymptotic formula for it.
Citation: Jin Zhang, Xiaoxue Li. The sixth power mean of one kind generalized two-term exponential sums and their asymptotic properties[J]. Electronic Research Archive, 2023, 31(8): 4579-4591. doi: 10.3934/era.2023234
[1] | Li Wang, Yuanyuan Meng . Generalized polynomial exponential sums and their fourth power mean. Electronic Research Archive, 2023, 31(7): 4313-4323. doi: 10.3934/era.2023220 |
[2] | Jianghua Li, Zepeng Zhang, Tianyu Pan . On two-term exponential sums and their mean values. Electronic Research Archive, 2023, 31(9): 5559-5572. doi: 10.3934/era.2023282 |
[3] | Zhefeng Xu, Xiaoying Liu, Luyao Chen . Hybrid mean value involving some two-term exponential sums and fourth Gauss sums. Electronic Research Archive, 2025, 33(3): 1510-1522. doi: 10.3934/era.2025071 |
[4] | Li Rui, Nilanjan Bag . Fourth power mean values of one kind special Kloosterman's sum. Electronic Research Archive, 2023, 31(10): 6445-6453. doi: 10.3934/era.2023326 |
[5] | Hongjian Li, Kaili Yang, Pingzhi Yuan . The asymptotic behavior of the reciprocal sum of generalized Fibonacci numbers. Electronic Research Archive, 2025, 33(1): 409-432. doi: 10.3934/era.2025020 |
[6] | Li Yu, Xiaofeng Wang . Semi-local convergence of a sixth-order iterative method for solving nonlinear systems. Electronic Research Archive, 2025, 33(6): 3794-3810. doi: 10.3934/era.2025168 |
[7] | Jiafan Zhang . On the distribution of primitive roots and Lehmer numbers. Electronic Research Archive, 2023, 31(11): 6913-6927. doi: 10.3934/era.2023350 |
[8] | Meenakshi Rana, Shruti Sharma . Combinatorics of some fifth and sixth order mock theta functions. Electronic Research Archive, 2021, 29(1): 1803-1818. doi: 10.3934/era.2020092 |
[9] | Jing Huang, Qian Wang, Rui Zhang . The short interval results for power moments of the Riesz mean error term. Electronic Research Archive, 2023, 31(9): 5917-5927. doi: 10.3934/era.2023300 |
[10] | Shufen Zhao . The S-asymptotically $ \omega $-periodic solutions for stochastic fractional differential equations with piecewise constant arguments. Electronic Research Archive, 2023, 31(12): 7125-7141. doi: 10.3934/era.2023361 |
The main aim of this article is using the elementary method and the number of the solutions of some congruence equations modulo an odd prime p, to study the calculating problem of the sixth power mean of one kind generalized two-term exponential sums, and give a sharp asymptotic formula for it.
As usual, we let p be an odd prime, χ denotes a Dirichlet character modulo p. For any integers k>h≥1 and integers m and n, the generalized two-term exponential sum S(m,n,k,h,χ;p) is defined as follows:
S(m,n,k,h,χ;p)=p−1∑a=1χ(a)e(mak+nahp), |
where e(y)=e2πiy and i is the imaginary unit.
In analytic number theory, these sums play important roles, many classical number theory problems are closely related to it, such as the prime distribution and Waring's problems etc. And because of that, many number theorists and scholars had studied the various properties of S(m,n,k,h,χ;p), and obtained a series of meaningful research results. For example, R. Duan and W. P. Zhang [1] proved that for any prime p with 3∤(p−1), and any Dirichlet character λmodp, one has the identities
p−1∑m=1|p−1∑a=1λ(a)e(ma3+nap)|4={3p3−8p2 if λ=(∗p),2p3−7p2 if λ≠χ0,λ≠(∗p),2p3−3p2−3p−1 if λ=χ0, |
where (∗p) is the Legendre symbol and χ0 is the principal character modulo p.
L. Chen and X. Wang [2] used the elementary method to prove the identities
p−1∑m=1|p−1∑a=0e(ma4+ap)|4={2p2(p−2), if p≡7mod12,2p3, if p≡11mod12,2p(p2−10p−2α2), if p≡1mod24,2p(p2−4p−2α2), if p≡5mod24,2p(p2−6p−2α2), if p≡13mod24,2p(p2−8p−2α2), if p≡17mod24, |
where the character sum α=α(p)=p−12∑a=1(a3+ap) is an integer satisfying the identity (see [3, Theorem 4–11]):
p=α2+β2=(p−12∑a=1(a3+ap))2+(p−12∑a=1(a3+rap))2, |
and r denotes any quadratic non-residue modulo p.
Recently, W. P. Zhang and Y. Y. Meng [4] studied the sixth power mean of S(m,n,3,1,χ0;p), and proved that for any odd prime p and integer n with (n,p)=1, one has the identities
p−1∑m=1|p−1∑a=0e(ma3+nap)|6={5p3(p−1) if p≡5mod6;p2(5p2−23p−d2) if p≡1mod6, |
where 4p=d2+27⋅b2, and d is uniquely determined by d≡1mod3 and b>0.
In addition, X. Y. Liu and W. P. Zhang [5] studied the calculating problem of the sixth power mean of the generalized two-term exponential sums
∑χmodpp−1∑m=0|p−1∑a=1χ(a)e(ma3+ap)|6, | (1.1) |
and proved that for any odd prime p with 3∤(p−1), one has the identity
∑χmodpp−1∑m=0|p−1∑a=1χ(a)e(ma3+ap)|6=p(p−1)(6p3−28p2+39p+5). |
Some papers related to the exponential sums and the generalized exponential sums can also be found in [6,7,8,9,10], to save space, we will not list them all here.
Of course, the results in [5] are very neat and beautiful, the only drawback is that they do not talk about the case 3∣(p−1). Then for prime p with p≡1mod3, what is going to happen? This seems to be an open problem. In this paper, we will use the elementary and analytic methods, and the number of the solutions of some congruence equations to study this problem, and prove the following conclusion:
Theorem. For any prime p with p≡1mod3, we have the asymptotic formula
∑χmodpp−1∑m=0|p−1∑a=1χ(a)e(ma3+ap)|6=6p5+O(p4). |
Combining our theorem and the result in [5] we can deduce the following:
Corollary. For any odd prime p, we have the asymptotic formula
∑χmodpp−1∑m=0|p−1∑a=1χ(a)e(ma3+ap)|6=6p5+O(p4). |
Some notes: For prime p with p≡1mod3, we can only get an asymptotic formula for (1.1), but can not get an exact calculating formula. The reason is that we can not get an exact value in Lemma 6.
Whether there exists an exact calculating formula for (1.1) with p≡1mod3 is an open problem. It remains to be further studied.
In this section, we decompose the proof of the theorem into six simple lemmas. Of course, the proofs of these lemmas need some knowledge of elementary or analytic number theory, all these can be found in [3,11,12], we will not repeat them here. First we have the following:
Lemma 1. Let p be an odd prime with p≡1mod3, then we have the identity
p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1a+b+c≡d+e+1modpabc≡demodp1=p3−3p2+5p−5. |
Proof. For any integer a, from the properties of the character sums
∑χmodpχ(a)={p−1 if a≡1modp;0 otherwise, |
and the properties of the classical Gauss sums modulo p
p−1∑a=1χ(a)e(map)=¯χ(m)p−1∑a=1χ(a)e(ap)=¯χ(m)τ(χ), |
we have
∑χmodpp−1∑m=0(p−1∑a=1χ(a)e(map))3(p−1∑a=1¯χ(a)e(−map))3=p(p−1)p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1p−1∑f=1a+b+c≡d+e+fmodpabc≡defmodp1=p(p−1)p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1p−1∑f=1af+bf+cf≡df+ef+fmodpabcf3≡def3modp1=p(p−1)2p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1a+b+c≡d+e+1modpabc≡demodp1. | (2.1) |
On the other hand, we also have
∑χmodpp−1∑m=0(p−1∑a=1χ(a)e(map))3(p−1∑a=1¯χ(a)e(−map))3=(p−1)6+∑χmodpp−1∑m=1τ3(χ)¯τ(χ)3=(p−1)6+(p−1)+(p−1)∑χ≠χ0τ3(χ)¯τ(χ)3=(p−1)6+(p−1)[1+(p−2)p3]=p(p−1)2(p3−3p2+5p−5), | (2.2) |
where τ(χ)=p−1∑a=1χ(a)e(ap) denotes the classical Gauss sums, χ0 denotes the principal character modulo p, ¯χ(a) and ¯τ(χ) denote the complex conjugate of χ(a) and τ(χ).
Combining (2.1) and (2.2) we have the identity
p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1a+b+c≡d+e+1modpabc≡demodp1=p3−3p2+5p−5. |
This proves Lemma 1.
Lemma 2. Let p be an odd prime with p≡1mod3. Then for any third-order character λ modulo p, we have the identity
p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1a+b+c≡d+e+1modpabc≡demodpλ(a)¯λ(d)=p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1a+b+c≡d+e+1modpabc≡demodp¯λ(a)λ(d)=p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1a+b+c≡d+e+1modpabc≡demodpλ(a)=p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1a+b+c≡d+e+1modpabc≡demodp¯λ(a)=p2−2p−1. |
Proof. First from the properties of the reduced residue system modulo p we have
p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1a+b+c≡d+e+1modpabc≡demodpλ(a)=p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1a−1+d(b−1)+e(c−1)≡0modpabc≡1modpλ(a). | (2.3) |
For any integer b with (b,p)=1, let ¯b denotes the multiplicative inverse of b modulo p. If a=b=c=1, then the congruence equations a−1+d(b−1)+e(c−1)≡0modp and abc≡1modp have (p−1)2 solutions and λ(1)=1.
If a=1, b≠1 and c=¯b, then the congruence equations a−1+d(b−1)+e(c−1)≡0modp and abc≡1modp have (p−1)(p−2) solutions.
Similarly, if b=1, a≠1 and c=¯a, then we have
p−1∑a=2p−1∑d=1p−1∑e=1a−1+e(¯a−1)≡0modpλ(a)=(p−1)p−1∑a=2λ(a)=−(p−1). | (2.4) |
If c=1, a≠1 and b=¯a, then we also have
p−1∑a=2p−1∑d=1p−1∑e=1a−1+d(¯a−1)≡0modpλ(a)=(p−1)p−1∑a=2λ(a)=−(p−1). | (2.5) |
If a≠1, b≠1, c≠1 and abc≡1modp, then we have
p−1∑a=2p−1∑b=2p−1∑c=2p−1∑d=1p−1∑e=1a−1+d(b−1)+e(c−1)≡0modpabc≡1modpλ(a)=p−1∑a=2p−1∑b=2p−1∑c=2p−1∑d=1p−1∑e=11+d+e≡0modpabc≡1modpλ(a)=(p−2)[(p−1)p−1∑a=1λ(a)−(p−2)−2p−1∑a=2λ(a)−1]=(p−2)(3−p). | (2.6) |
Combining (2.3)–(2.6) we have the identity
p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1a+b+c≡d+e+1modpabc≡demodpλ(a)=p2−2p−1. | (2.7) |
From the properties of the reduced residue system modulo p and (2.7) we also have
p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1a+b+c≡d+e+1modpabc≡demodpλ(a)¯λ(d)=p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1ad+bd+cd≡d+ed+1modpabcd3≡d2emodpλ(ad)¯λ(d)=p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1a+b+c≡1+e+¯dmodpabc≡¯demodpλ(a)=p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1a+b+c≡d+e+1modpabc≡demodpλ(a)=p2−2p−1. | (2.8) |
Now Lemma 2 follows from (2.7) and (2.8).
Lemma 3. Let p be an odd prime with p≡1mod3. Then for any third-order character λ modulo p, we have the identity
p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1a+b+c≡d+e+1modpabc≡demodpλ(d)=p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1a+b+c≡d+e+1modpabc≡demodp¯λ(d)=p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1a+b+c≡d+e+1modpabc≡demodpλ(a)¯λ(b)=p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1a+b+c≡d+e+1modpabc≡demodpλ(b)¯λ(a)=−(p−1). |
Proof. First from the properties of the reduced residue system we have
p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1a+b+c≡d+e+1modpabc≡demodpλ(d)=p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1d(a−1)+e(b−1)+c−1≡0modpabc≡1modpλ(d). | (2.9) |
If a=b=c=1, then from (2.9) we have
p−1∑d=1p−1∑e=1λ(d)=0. | (2.10) |
If a=1, b≠1 and c=¯b, then from (2.9) we have
p−1∑b=2p−1∑c=2p−1∑d=1p−1∑e=1e(b−1)+c−1≡0modpbc≡1modpλ(d)=p−1∑b=2p−1∑c=2p−1∑e=1e(b−1)+c−1≡0modpbc≡1modpp−1∑d=1λ(d)=0. | (2.11) |
Similarly, if b=1, a≠1 and c=¯a, then from (2.9) we have
p−1∑a=2p−1∑d=1p−1∑e=1d(a−1)+¯a−1≡0modpλ(d)=(p−1)p−1∑a=2λ(d)=−(p−1). | (2.12) |
If c=1, a≠1 and b=¯a, then we also have
p−1∑a=2p−1∑d=1p−1∑e=1d(a−1)+e(¯a−1)≡0modpλ(d)=p−1∑a=2p−1∑e=1λ(¯a)λ(e)=0. | (2.13) |
If a≠1, b≠1, c≠1 and abc≡1modp, then note that λ(¯a)=¯λ(a), from (2.9) we have the identity
p−1∑a=2p−1∑b=2p−1∑c=2p−1∑d=1p−1∑e=1d(a−1)+e(b−1)+(c−1)≡0modpabc≡1modpλ(d)=p−1∑a=2p−1∑b=2p−1∑c=2p−1∑d=1p−1∑e=1d+e+c−1≡0modpabc≡1modp¯λ(a−1)λ(d)=p−1∑a=2p−1∑b=2p−1∑c=2p−1∑e=1abc≡1modp¯λ(a−1)λ(1−c−e)=−p−1∑a=1p−1∑b=2p−1∑c=1abc≡1modp¯λ(a−1)λ(c−1)=−p−1∑a=1p−1∑c=1¯λ(a−1)λ(c−1)+p−1∑a=1¯λ(a−1)λ(¯a−1)=−1−p−1∑a=2λ(¯a)=0. | (2.14) |
Combining (2.9)–(2.14) we have the identity
p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1a+b+c≡d+e+1modpabc≡demodpλ(d)=−(p−1). | (2.15) |
From the properties of the reduced residue system modulo p and (2.15) we have
p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1a+b+c≡d+e+1modpabc≡demodpλ(a)¯λ(b)=p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1ab+b+cb≡db+eb+1modpab3c≡dbebmodpλ(ab)¯λ(b)=p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1a+1+c≡d+e+¯bmodpac≡de¯bmodpλ(a)=p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1a+1+c≡d+e+bmodpac≡debmodpλ(a)=−(p−1). | (2.16) |
Now Lemma 3 follows from (2.15) and (2.16).
Lemma 4. Let p be an odd prime with p≡1mod3. Then for any third-order character λ modulo p, we have the identities
|p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1a+b+c≡d+e+1modpabc≡demodpλ(ab)|=|p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1a+b+c≡d+e+1modpabc≡demodp¯λ(c)λ(de)|≤p2; |
|p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1a+b+c≡d+e+1modpabc≡demodpλ(abc)|≤p2 and |p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1a+b+c≡d+e+1modpabc≡demodpλ(dec)|≤p2. |
In fact, if abc≡demodp and λi(a)λj(b)λk(c)λh(d)λs(e)≠1, 0≤i, j, k, h, s≤2, then we can also prove the estimate
|p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1a+b+c≡d+e+1modpabc≡demodpλi(a)λj(b)λk(c)λh(d)λs(e)|≤p2. |
Proof. From the properties of the orthogonality of the characters we have
∑χmodpp−1∑m=0(p−1∑a=1λ(a)χ(a)e(map))2(p−1∑a=1χ(a)e(map))×(p−1∑b=1¯χ(b)e(−mbp))2e(−mp)=p(p−1)p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1a+b+c≡d+e+1modpabc≡demodpλ(ab). | (2.17) |
On the other hand, from the properties of the classical Gauss sums we also have
∑χmodpp−1∑m=0(p−1∑a=1λ(a)χ(a)e(map))2(p−1∑a=1χ(a)e(map))×(p−1∑b=1¯χ(b)e(−mbp))2e(−mp)=∑χmodpτ2(λχ)τ(χ)¯τ(χ)2p−1∑m=1¯λ(m)¯χ(m)e(−mp)=∑χmodp|τ(λχ)|2|τ(χ)|2τ(λχ)¯τ(χ) |
or
|∑χmodpp−1∑m=0(p−1∑a=1λ(a)χ(a)e(map))2(p−1∑a=1χ(a)e(map))|×|(p−1∑b=1¯χ(b)e(−mbp))2e(−mp)|=|∑χmodp|τ(λχ)|2|τ(χ)|2τ(λχ)¯τ(χ)|≤p3(p−1). | (2.18) |
Combining (2.17) and (2.18) we may immediately deduce the estimate
|p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1a+b+c≡d+e+1modpabc≡demodpλ(ab)|≤p2. |
Similarly, we can also deduce the other estimations in Lemma 4.
Lemma 5. Let p be an odd prime with p≡1mod3. Then we have the identity
p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1a3+b3+c3≡d3+e3+1modpa+b+c≡d+e+1modpabc≡demodp1=7p2−29p+44. |
Proof. From the properties of the congruence equation modulo p we have
p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1a3+b3+c3≡d3+e3+1modpa+b+c≡d+e+1modpabc≡demodp1=p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1(a+b+c)3−a3−b3−c3≡(d+e+1)3−d3−e3−1modpa+b+c≡d+e+1modpabc≡demodp1=p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1(a+b+c)(a2+b2+c2−d2−e2−1)≡0modpa+b+c≡d+e+1modpabc≡demodp1=p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1a+b+c≡0modpa+b+c≡d+e+1modpabc≡demodp1+p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1a2+b2+c2≡d2+e2+1modpa+b+c≡d+e+1modpabc≡demodp1−p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1a2+b2+c2≡d2+e2+1modpa+b+c≡d+e+1≡0modpabc≡demodp1. | (2.19) |
Now we calculate the three sums in (2.19) separately. First we have
p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1a+b+c≡0modpa+b+c≡d+e+1modpabc≡demodp1=p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1da+eb+c≡0modpd+e+1≡0modpabc≡1modp1=p−1∑a=1p−1∑b=1p−1∑d=1p−1∑e=1da+eb+¯ab≡0modpd+e+1≡0modp1=p−1∑a=1p−1∑b=1p−1∑d=1da−(d+1)b+¯ab≡0modp1−p−1∑a=1p−1∑b=1−a+¯ab≡0modp1=p−1∑a=1p−1∑b=1p−1∑d=1d(a−b)≡b−¯abmodp1−(p−1)=3(p−1)+3(p−2)+(p−3)(p−4)−(p−1)=p2−2p+4. | (2.20) |
p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1a2+b2+c2≡d2+e2+1modpa+b+c≡d+e+1modpabc≡demodp1=p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1ab+ac+bc≡de+d+emodpa+b+c≡d+e+1modpabc≡demodp1=p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1ab+(a+b)c≡abde+ad+ebmodpa+b+c≡ad+be+1modpc≡demodp1=p−1∑a=1p−1∑b=1p−1∑d=1p−1∑e=1(a−1)(b−1)(de−1)≡0modpa+b+de≡ad+be+1modp1=1+p−1∑d=1p−1∑e=1(d−1)(e−1)≡0modp(de−1,p)=11+2p−1∑b=2p−1∑d=1(d−1)(b−¯d)≡0modp1+2p−1∑b=2p−1∑d=1p−1∑e=1(e−1)(b−d)≡0modp(de−1,p)=11+p−1∑a=2p−1∑b=2p−1∑d=1p−1∑e=1a+b≡ad+bemodpde≡1modp1=1+2(p−2)+4(p−2)+2(p−2)(2p−5)+p−1∑a=2p−1∑b=2p−1∑e=1(a+b)e≡a+be2modp1=1+4(p−2)(p−1)+p−1∑a=2p−1∑b=2p−1∑e=1(e−1)(eb−a)≡0modp1=1+4(p−2)(p−1)+2(p−2)2−(p−2)=6p2−21p+19. | (2.21) |
p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1a2+b2+c2≡d2+e2+1modpa+b+c≡d+e+1≡0modpabc≡demodp1=p−1∑a=1p−1∑b=1p−1∑d=1p−1∑e=1(a−1)(b−1)(de−1)≡0modpa+b+de≡ad+be+1≡0modp1=p−1∑d=1p−1∑e=12+de≡d+e+1≡0modp1+2p−1∑b=2p−1∑d=12+b≡d+b¯d+1≡0modp1+2p−1∑b=2p−1∑d=1p−1∑e=11+b+de≡d+be+1≡0modp(de−1,p)=11+p−1∑a=2p−1∑b=2p−1∑d=1a+b+1≡ad+b¯d+1≡0modp1=2+4+2(p−4+p−5)+p−4+p−5=6p−21. | (2.22) |
Combining (2.19)–(2.22), we have the desired result
p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1a3+b3+c3≡d3+e3+1modpa+b+c≡d+e+1modpabc≡demodp1=7p2−29p+44. |
Lemma 6. Let p be an odd prime with p≡1mod3. Then we have the asymptotic formula
p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1a3+b3+c3≡d3+e3+1modpabc≡demodp1=p3+O(p2). |
Proof. Let λ be a third-order character modulo p. Note that if abc≡demodp, then λ(abc)¯λ(de)=¯λ(abc)λ(de)=1, so from the properties of the third-order characters modulo p and Lemmas 1–4 we have the asymptotic formula
p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1a3+b3+c3≡d3+e3+1modpabc≡demodp1=13p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1a3+b3+c3≡d3+e3+1modpa3b3c3≡d3e3modp1=13p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1a+b+c≡d+e+1modpabc≡demodp(1+λ(a)+¯λ(a))(1+λ(b)+¯λ(b))×(1+λ(c)+¯λ(c))(1+λ(d)+¯λ(d))(1+λ(e)+¯λ(e))=p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑e=1a+b+c≡d+e+1modpabc≡demodp1+O(p2)=p3+O(p2). |
This proves Lemma 6.
Now we apply the lemmas in Section 2 to complete the proof of our theorem. Note that the trigonometrical identities
p−1∑a=0e(nap)={p if p∣n;0 if p∤n. |
From the orthogonality of the characters sums, Lemma 5 and 6 we have
∑χmodpp−1∑m=0|p−1∑a=1χ(a)e(ma3+ap)|6=p(p−1)p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑h=1p−1∑f=1a3+b3+c3≡d3+h3+f3modpabc≡dhfmodpe(a+b+c−d−h−fp)=p(p−1)p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑h=1a3+b3+c3≡d3+h3+1modpabc≡dhmodpp−1∑f=1e(f(a+b+c−d−h−1)p)=p2(p−1)p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑h=1a3+b3+c3≡d3+h3+1modpa+b+c≡d+h+1modpabc≡dhmodp1−p(p−1)p−1∑a=1p−1∑b=1p−1∑c=1p−1∑d=1p−1∑h=1a3+b3+c3≡d3+h3+1modpabc≡dhmodp1=p2(p−1)(7p2−29p+44)−p(p−1)(p3+O(p2))=6p5+O(p4). |
This completes the proof of our theorem.
The main result of this paper is to give an asymptotic formula for the sixth power mean of one kind special two-term exponential sums. That is, for any prime p with p≡1mod3, we have the asymptotic formula
∑χmodpp−1∑m=0|p−1∑a=1χ(a)e(ma3+ap)|6=6p5+O(p4). |
Whether there exists an exact formula for the above power mean is still an open problem. We will keep working on it.
The authors declare they have not used Artificial Intelligence (AI) tools in the creation of this article.
The authors would like to thank the referees for their very helpful and detailed comments, which have significantly improved the presentation of this paper. All authors have equally contributed to this work. All authors have read and approved the final manuscript. This work is supported by the N. S. F. (12126357) of P. R. China, and Scientific Research Program Funded by Shaanxi Provincial Education Department (Program No.22JK0424).
The authors declare that there are no conflicts of interest regarding the publication of this paper.
[1] | R. Duan, W. P. Zhang, On the fourth power mean of the generalized two-term exponential sums, Math. Rep., 72 (2020), 205–212. |
[2] |
L. Chen, X. Wang, A new fourth power mean of two-term exponential sums, Open Math., 17 (2019), 407–414. https://doi.org/10.1515/math-2019-0034 doi: 10.1515/math-2019-0034
![]() |
[3] | W. P. Zhang, H. L. Li, Elementary Number Theory, Shaanxi Normal University Press, Xi'an, 2013. |
[4] |
W. P. Zhang, Y. Y. Meng, On the sixth power mean of the two-term exponential sums, Acta Math. Sin., Engl. Ser., 38 (2022), 510–518. https://doi.org/10.1007/s10114-022-0541-8 doi: 10.1007/s10114-022-0541-8
![]() |
[5] |
X. Y. Liu, W. P. Zhang, On the high-power mean of the generalized Gauss sums and Kloosterman sums, Mathematics, 7 (2019), 907. https://doi.org/ 10.3390/math7100907 doi: 10.3390/math7100907
![]() |
[6] | H. Zhang, W. P. Zhang, The fourth power mean of two-term exponential sums and its application, Math. Rep., 69 (2017), 75–81. |
[7] |
W. P. Zhang, D. Han, On the sixth power mean of the two-term exponential sums, J. Number Theory, 136 (2014), 403–413. http://dx.doi.org/10.1016/j.jnt.2013.10.022 doi: 10.1016/j.jnt.2013.10.022
![]() |
[8] | H. N. Liu, W. M. Li, On the fourth power mean of the three-term exponential sums, Adv. Math., 46 (2017), 529–547. |
[9] | X. C. Du, X. X. Li, On the fourth power mean of generalized three-term exponential sums, J. Math. Res. Appl., 35 (2015), 92–96. |
[10] |
X. Y. Wang, X. X. Li, One kind sixth power mean of the three-term exponential sums, Open Math., 15 (2017), 705–710. http://dx.doi.org/10.1515/math-2017-0060 doi: 10.1515/math-2017-0060
![]() |
[11] | T. M. Apostol, SIntroduction to Analytic Number Theory, Springer-Verlag, New York, 1976. https://doi.org/10.1007/978-1-4757-5579-4 |
[12] | K. Ireland, M. Rosen, A Classical Introduction to Modern Number Theory, Springer-Verlag, New York, 1982. https://doi.org/10.1007/978-1-4757-1779-2 |