This paper introduces a generalization of the alternating harmonic series, expresses the sum in two closed forms, and examines the relationship between these sums and the harmonic numbers.
Citation: Shelby Kilmer, Songfeng Zheng. A generalized alternating harmonic series[J]. AIMS Mathematics, 2021, 6(12): 13480-13487. doi: 10.3934/math.2021781
[1] | Jiaye Lin . Evaluations of some Euler-type series via powers of the arcsin function. AIMS Mathematics, 2025, 10(4): 8116-8130. doi: 10.3934/math.2025372 |
[2] | Chunli Li, Wenchang Chu . Remarkable series concerning $ \binom{3n}{n} $ and harmonic numbers in numerators. AIMS Mathematics, 2024, 9(7): 17234-17258. doi: 10.3934/math.2024837 |
[3] | Kwang-Wu Chen, Fu-Yao Yang . Infinite series involving harmonic numbers and reciprocal of binomial coefficients. AIMS Mathematics, 2024, 9(7): 16885-16900. doi: 10.3934/math.2024820 |
[4] | Xuqiong Luo . A D-N alternating algorithm for exterior 3-D problem with ellipsoidal artificial boundary. AIMS Mathematics, 2022, 7(1): 455-466. doi: 10.3934/math.2022029 |
[5] | H. C. Vidya, B. A. Rao . Utilization of Ramanujan-type Eisenstein series in the generation of differential equations. AIMS Mathematics, 2024, 9(10): 28955-28969. doi: 10.3934/math.20241405 |
[6] | Maryam AL-Towailb . A generalization of the $ q $-Lidstone series. AIMS Mathematics, 2022, 7(5): 9339-9352. doi: 10.3934/math.2022518 |
[7] | Anh Tuan Nguyen, Le Dinh Long, Devendra Kumar, Van Thinh Nguyen . Regularization of a final value problem for a linear and nonlinear biharmonic equation with observed data in $ L^{q} $ space. AIMS Mathematics, 2022, 7(12): 20660-20683. doi: 10.3934/math.20221133 |
[8] | Ling Zhu . Asymptotic expansion of a finite sum involving harmonic numbers. AIMS Mathematics, 2021, 6(3): 2756-2763. doi: 10.3934/math.2021168 |
[9] | Fei Hou, Bin Chen . Triple correlation sums of coefficients of $ \theta $-series. AIMS Mathematics, 2023, 8(10): 25275-25287. doi: 10.3934/math.20231289 |
[10] | Kee Wah Fo, Seng Huat Ong, Choung Min Ng, You Beng Koh . An alternative hyper-Poisson integer-valued GARCH model with application to polio, internet protocol and COVID-19 data. AIMS Mathematics, 2023, 8(12): 29116-29139. doi: 10.3934/math.20231491 |
This paper introduces a generalization of the alternating harmonic series, expresses the sum in two closed forms, and examines the relationship between these sums and the harmonic numbers.
Riemann's Rearrangement Theorem says that if an infinite series of real numbers is conditionally convergent, then its terms can be rearranged in in such a way that the resulting series converges to any real sum [3]. It is well known that the alternating harmonic series converges to log2, that is,
1−12+13−14+⋯=log2. |
As the prototypical conditionally convergent series, rearrangements of the alternating harmonic series have been studied extensively [1]. However, Riemann's Rearrangement Theorem is non-constructive; there is no general method to find the sum of a re-arrangement. It was shown in [8] that assigning plus or minus signs randomly produces sums that converge almost surely.
In this article, we consider a generalized version of the alternating harmonic series, one with the assignment of plus or minus signs, as follows. For each positive integer k, we consider the series
Sk=(1+12+13+⋯+1k)−(1k+1+1k+2+⋯+12k)+(12k+1+12k+2+⋯+13k)−(13k+1+13k+2+⋯+14k)+⋯ |
We will find the infinite sum, Sk, of the infinite series in two different formats and examine the interesting relationship between this sum and the harmonic number,
Hk=1+12+13+⋯+1k=k∑n=11n, |
which appears often throughout mathematics [2].
Definition 1.1. We take H0=0 and define the terms of Sk by
an(k)=Hnk−H(n−1)k, |
for each positive integer n and k. Thus,
Sk=∞∑n=1(−1)n+1an(k). |
Clearly, for a fixed k, an(k)>0 and decreases to 0 as n increases. Hence, by the alternating series test, Sk is convergent for all positive integers k.
Our first summation formula is given in integral form, as follows.
Theorem 2.1. If k is a positive integer, then
Sk=k−1∑m=0∫10xm1+xkdx. | (2.1) |
Proof. Since
∫1011+xdx=log2, |
Equation (2.1) holds for k=1. For k≥2, we first note that the harmonic number has an integral expression [7] as
Hn=1+12+⋯+1n=∫10dx+∫10xdx+⋯+∫10xn−1dx=∫10(1+x+⋯+xn−1)dx=∫101−xn1−xdx. |
Hence when k≥2,
an(k)=Hnk−H(n−1)k=∫101−xnk1−xdx−∫101−x(n−1)k1−xdx=∫10x(n−1)k−xnk1−xdx=∫10x(n−1)k1−xk1−xdx. | (2.2) |
It follows that for k≥2,
Sk=∞∑n=1(−1)n+1an(k)=∞∑n=1(−1)n−1∫10x(n−1)k1−xk1−xdx=∫101−xk1−x∞∑n=1(−xk)n−1dx=∫101−xk1−x11+xkdx=∫101+x+⋯+xk−11+xkdx=k−1∑m=0∫10xm1+xkdx, |
proving the theorem.
As an application of Theorem 2.1, we have
Example 2.2.
S2=∫101+x1+x2dx=arctanx|10+12log(1+x2)|10=π4+12log2. |
Our second summation formula is given in terms of trigonometric functions, as follows.
Corollary 2.3. If If k is a positive integer, then
Sk=π2kk−1∑m=1cscmπk+1klog2. |
Proof. By Theorem 2.1,
Sk=k−1∑m=0∫10xm1+xkdx=k−2∑m=0∫10xm1+xkdx+∫10xk−11+xkdx=k−1∑m=1∫10xm−11+xkdx+1klog2=12(k−1∑m=1∫10xm−11+xkdx+k−1∑m=1∫10xm−11+xkdx)+1klog2=12(k−1∑m=1∫10xm−11+xkdx+k−1∑m′=1∫10xk−m′−11+xkdx)+1klog2 | (2.3) |
=12k−1∑m=1∫10xm−1+xk−m−11+xkdx+1klog2, | (2.4) |
where in Eq (2.3), for the second summation in the parenthesis, we changed the index from m to m′=k−m. We simplify Eq (2.4) using the following well known result, see page 323 of [5].
For k>m>0,∫10xm−1+xk−m−11+xkdx=πkcscmπk. |
This yields the desired result,
Sk=π2kk−1∑m=1cscmπk+1klog2, |
and this proves the Theorem.
Example 2.4.
S3=π62∑m=1cscmπ3+13log2=2π√39+13log2. |
We now set about finding the relationship between Sk and Hk. Since on [0,1],1≤1+xk≤2 for all integer k≥1, we have that,
k−1∑m=0∫10xm2dx≤k−1∑m=0∫10xm1+xkdx≤k−1∑m=0∫10xmdx. |
Hence by Theorem 2.1, we have
12Hk≤Sk≤Hk, for all k≥1. |
We will now calculate and simplify the difference between Hk and Sk. Notice that
Hk−Sk=a1(k)−∞∑n=1(−1)n+1an(k)=−∞∑n=2(−1)n+1an(k)=∞∑n=1(−1)n+1an+1(k). |
We make the following definition.
Definition 3.1. Let
Dk=Hk−Sk=∞∑n=1(−1)n+1an+1(k). |
Note that for each integer k≥1, Dk is a convergent alternating series, which easily follows from the properties of an(k). More interestingly, Dk itself forms a convergent sequence, as given in the following.
Theorem 3.2.
limk→∞Dk=limk→∞(Hk−Sk)=logπ2. |
Proof. Let ε be an arbitrarily fixed positive real number. By the Wallis product formula [6], we have
∞∏n=1(2n)2(2n−1)(2n+1)=π2. |
Since the natural logarithm is a continuous function, there exist a positive integer N1 such that when l≥N1,
|logl∏n=1(2n)2(2n−1)(2n+1)−logπ2|<ε3. | (3.1) |
Let m be any positive integer and let k≥1. By (2.2), we have
|∞∑n=m(−1)n+1an+1(k)|=|∞∑n=m(−1)n+1∫10xnk1−xk1−xdx|=|−∞∑n=m∫10(−xk)n1−xk1−xdx|≤∫101−xk1−x|∞∑n=m(−xk)n|dx≤∫101−xk1−xxmk1+xkdx=∫10(1+x+⋯+xk−1)xmk1+xkdx≤∫10kxmkdx=kmk+1<1m. |
Hence, there exists a positive integer N2 such than when m≥N2,
|∞∑n=m(−1)n+1an+1(k)|<ε3, | (3.2) |
for each and every k≥1.
Let N=max{N1,N2}. Recall the Euler-Mascheroni constant [4] is defined as
γ=limn→∞(Hn−logn). |
Since {Hn−logn} converges, it is a Cauchy sequence. We therefore may choose a positive integer K such that when j,m≥K,
|(Hj−Hm)−logjm|=|(Hj−logj)−(Hm−logm)|<ε6N. | (3.3) |
For convenience, we denote
D2N(k)=2N∑n=1(−1)n+1an+1(k) and T2N(k)=∞∑n=2N+1(−1)n+1an+1(k). |
Clearly, Dk=D2N(k)+T2N(k) and by (3.2), for any k≥1, there is
|T2N(k)|<ε3. | (3.4) |
Also notice that, for any integer k≥1, we have
D2N(k)=2N∑n=1(−1)n+1an+1(k)=a2(k)−a3(k)+a4(k)−a5(k)+⋯+a2N(k)−a2N+1(k)=N∑n=1[a2n(k)−a2n+1(k)]=N∑n=1[(H2nk−H(2n−1)k)−(H(2n+1)k−H2nk)]. | (3.5) |
Let
WN=logN∏n=1(2n)2(2n−1)(2n+1). |
Henceforth, let k≥K, where K is defined by Eq (3.3). Using the representation of D2N(k) in Eq (3.5), we have
|D2N(k)−WN|=|N∑n=1[(H2nk−H(2n−1)k)−(H(2n+1)k−H2nk)]−logN∏n=1(2n)2(2n+1)(2n−1)|=|N∑n=1{[(H2nk−H(2n−1)k)−(H(2n+1)k−H2nk)]−log(2n)2(2n+1)(2n−1)}|≤N∑n=1|[(H2nk−H(2n−1)k)−(H(2n+1)k−H2nk)]−log(2n)2(2n−1)(2n+1)|=N∑n=1|[(H2nk−H(2n−1)k)−log2n2n−1]−[(H(2n+1)k−H2nk)−log2n+12n]|≤N∑n=1{|(H2nk−H(2n−1)k)−log2nk(2n−1)k|+|(H(2n+1)k−H2nk)−log(2n+1)k2nk|}≤N(ε6N+ε6N)=ε3, | (3.6) |
where "≤" in (3.6) follows from (3.3) because k≥K, which makes all of 2nk, (2n−1)k, and (2n+1)k greater than K.
Finally, for k≥K, we have
|Hk−Sk−logπ2|=|Dk−logπ2|=|D2N(k)+T2N(k)−WN+WN−logπ2|≤|D2N(k)−WN|+|T2N(k)|+|WN−logπ2|<ε, |
where the last step follows from (3.1), (3.4), and (3.6). Therefore, by the arbitrariness of ε, we have
limk→∞Dk=limk→∞(Hk−Sk)=logπ2, |
and this completes the proof.
The authors would like to extend their sincere gratitude to the anonymous reviewers for their constructive suggestions and comments, which have greatly helped improve the quality of this paper. We are grateful to Prof. Les Reid for stimulating discussions during the preparation of the manuscript. S. Zheng was supported by a Summer Faculty Fellowship from Missouri State University.
The authors declare that there is no conflicts of interest.
[1] |
C. Cowen, K. Davidson, R. Kaufman, Rearranging the Alternating Harmonic Series, Amer. Math. Monthly, 87 (1980), 817–819. doi: 10.1080/00029890.1980.11995162
![]() |
[2] |
D. W. DeTemple, The Non-Integer Property of Sums of Reciprocals of Consecutive Integers, The Mathematical Gazette, 75 (1991), 193. doi: 10.2307/3620253
![]() |
[3] | Y. Dybskiy, K. Slutsky, Riemann Rearrangement Theorem for some types of convergence, J. Math. Anal. Appl., 373 (2010), 605–613. |
[4] | S. R. Finch, Mathematical Constants, Cambridge University Press, 2003. |
[5] | I. S. Gradshteyn, I. M. Ryzhik, Table of Integrals, Series, and Products, 7th Edition, Academic Press, 2014. |
[6] | H. Jeffreys, B. S. Jeffreys, B. Swirles, Methods of Mathematical Physics, 3rd ed, Cambridge University Press, 1999. |
[7] | C. E. Sandifer, How Euler did it, MAA Spectrum, Mathematical Association of America, Washington, D.C., 2007. |
[8] | B. Schmuland, Random Harmonic Series, Amer. Math. Monthly, 110 (2003), 407–416. |
1. | Chunli Li, Wenchang Chu, Parametric Integrals for Binomial Series with Harmonic Polynomials, 2024, 13, 2075-1680, 885, 10.3390/axioms13120885 |