- Research
- Open access
- Published:
On Copson’s inequalities for \(0< p<1\)
Journal of Inequalities and Applications volume 2020, Article number: 72 (2020)
Abstract
Let \((\lambda_{n})_{n \geq1}\) be a positive sequence and let \(\varLambda_{n}=\sum^{n}_{i=1}\lambda_{i}\). We study the following Copson inequality for \(0< p<1\), \(L>p\):
We find conditions on \(\lambda_{n}\) such that the above inequality is valid with the constant being the best possible.
1 Introduction
Let \(p>0\) and \(\mathbf{x}=(x_{n})_{n \geq1}\) be a positive sequence Let \((\lambda_{n})_{n \geq1}\) be a positive sequence and let \(\varLambda_{n}=\sum^{n}_{i=1}\lambda_{i}\). The well-known Copson inequalities [4, Theorem 1.1, 2.1] state that
The above two inequalities are equivalent (see [10]) and the constants are best possible. When \(\lambda_{k}=1\), \(k \geq1\) and \(c=p\), inequality (1.1) becomes the following celebrated Hardy inequality ([12, Theorem 326]):
Note that the reversed inequality of (1.2) holds when \(c \leq0 < p <1\) ([4, Theorem 2.3]) with the constant being best possible and as pointed out in [1, p. 390], the reversed inequality of (1.2) continues to hold with constant \(p^{p}\) when \(c>0\). The particular case of \(c=p\), \(\lambda_{k}=1\), \(k \geq1\) becomes the following one given in [12, Theorem 345]:
It is noted in [12] that the constant \(p^{p}\) in (1.4) may not be best possible and the best constant for \(0< p \leq1/3\) was shown by Levin and Stečkin [13, Theorem 61] to be indeed \((p/(1-p))^{p}\). In [8], it is shown that the constant \((p/(1-p))^{p}\) stays best possible for all \(0< p \leq0.346\). It is further shown in [11] that the constant \((p/(1-p))^{p}\) is best possible when \(p=0.35\).
There exists an extensive and rich literature on extensions and generalizations of Copson’s inequalities and Hardy’s inequality (1.3) for \(p>1\). For recent developments in this direction, we refer the reader to the articles in [6–11] and the references therein. On the contrary, the case \(0< p<1\) is less known as can be seen by comparing inequalities (1.3) and (1.4). On one hand, the constant in (1.3) is shown to be best possible for all \(p>1\). On the other hand, though it is known the best constant that makes inequality (1.4) valid is \((p/(1-p))^{p}\) when \(0< p \leq 0.346\), it is shown in [8] that the constant \((p/(1-p))^{p}\) fails to be best possible when \(1/2 \leq p <1\) and the best constant in these cases remains unknown.
Our goal in this paper is to study the following variation of Copson’s inequalities for \(0< p<1\):
where \(L>p\) is a constant.
It is an open problem to determine the best possible constant to make inequality (1.5) valid in general. Our choice for presenting the constant in the form \((p/(L-p))^{p}\) in (1.5) is motivated by the study on the analogue case of inequality (1.5) when \(p>1\). We define
A result of Cartlidge [3] shows that when \(L_{\lambda} < p\) for \(p>1\), then the following inequality holds for all non-negative sequences x:
We shall see in Theorem 1.3 below that the constant given in inequality (1.5) is indeed best possible for certain sequences (\(\lambda_{n}\)) and certain ranges of p when one replaces L by \(L_{\lambda}\) in (1.5). This includes case concerning the classical inequality (1.4) (with \(p^{p}\) replaced by \((p/(1-p))^{p}\) there).
Further, let \(q<0\) be the number satisfying \(1/p+1/q=1\), we note that inequality (1.5) is equivalent to its dual version (assuming that \(x_{n}>0\) for all n):
The equivalence of the above two inequalities can be easily established following the discussions in [8, Sect. 1].
Our main result gives a condition on \(\lambda_{n}\) and L such that inequalities (1.5) and (1.8) hold. For this purpose, we define, for constants p and L,
Our main result is the following statement.
Theorem 1.1
Let\(0< p<1\)be fixed. Let\(\lambda=(\lambda_{n})_{n \geq1}\)be a positive sequence and let\(\varLambda_{n}=\sum^{n}_{i=1}\lambda_{i}\). If there exists a positive constant\(L>p\)such that, for any integer\(n \geq1\),
then inequality (1.5) holds for all non-negative sequencesx. In particular, let\(L_{\lambda}\)be defined by (1.6) and let\(a_{i}(L,p)\), \(i=1,2\)be defined by (1.9), then inequality (1.5) holds withLreplaced by\(L_{\lambda}\)there for all non-negative sequencesxwhen\(L_{\lambda} \geq 1\), \(0< p \leq1/3\)and\(a_{1}(L_{\lambda},p)\geq0\)or when\(0< L_{\lambda} < 1\), \(0< p \leq L_{\lambda}/4\)and\(a_{2}(L_{\lambda},p) \geq0\).
We note that when \(0< p<1/2\), we have \(\lim_{L \rightarrow\infty }a_{1}(L,p)<0\), \(\lim_{p \rightarrow0^{+}}a_{2}(4p,p)/p<0\). This implies that the values of \(a_{i}(L,p)\), \(i=1,2\) do give restrictions on the validity of inequality (1.5). We note that when \(\lambda_{n}=1\), then \(L_{\lambda}=1\), \(a_{1}(1,p)=3(\frac{1}{p}-2)^{2}-1\) and Theorem 1.1 implies the above-mentioned result of Levin and Stečkin.
In [2, 5–8], two special cases of inequality (1.7) corresponding to \(\lambda_{n}=n^{\alpha}-(n-1)^{\alpha}\) and \(\lambda_{n}=n^{\alpha-1}\) for \(p>1\), \(\alpha p >1\) were studied. It follows from this work that inequality (1.7) holds in either case with best possible constant \((\alpha p/(\alpha p-1))^{p}\) except for the case when \(\lambda_{n}=n^{\alpha-1}\), \(1< p \leq4/3\), \(1\leq\alpha <1+1/p\) or \(4/3< p<2\), \(1 \leq\alpha<2\).
It is now interesting to study the following \(0< p<1\), \(\alpha<1/p\) analogues of the above inequalities:
Note that when \(\alpha=1\), the above two inequalities become inequality (1.4) (with the constant \(p^{p}\) there being replaced by the best possible constant \((p/(1-p))^{p}\)). We note that it is shown in [8, (4.14)] that inequality (1.12) holds when \(0< \alpha<1\), \(0< p<1/(\alpha +2)\). It is also shown in [9, Theorem 1.1] that inequality (1.11) holds when \(\alpha>0\), \(0< p<1/(\alpha+2)\) when one replaces \(k^{\alpha}-(k-1)^{\alpha}\) by \((k+1)^{\alpha}-k^{\alpha }\). As \((k+1)^{\alpha}-k^{\alpha} \leq k^{\alpha}-(k-1)^{\alpha}\) when \(0<\alpha\leq1\), this implies that inequality (1.11) holds when \(0<\alpha\leq1\), \(0< p<1/(\alpha+2)\).
Note that the values of p are not given explicitly in (1.10), nor by the conditions \(a_{i}(L_{\lambda},p)\geq0\), \(i=1,2\). Thus, Theorem 1.1 is not readily applied in practice. For this reason, and with future applications in mind, we develop the following result.
Theorem 1.2
Let\(0< p<1\)be fixed. Let\(\lambda=(\lambda_{n})_{n \geq1}\)be a positive sequence and let\(\varLambda_{n}=\sum^{n}_{i=1}\lambda_{i}\). Let\(L_{\lambda}\)be defined by (1.6) such that\(0< L_{\lambda} <1\)and that
then inequality (1.5) holds withLreplaced by\(L_{\lambda}\)there for all non-negative sequencesx.
Suppose that there exist positive constants\(1/2< L<1\), \(0< M<1\), \(L+2M<1\)such that, for any integer\(n \geq1\),
Then inequality (1.5) holds for all non-negative sequencesxwhen
We remark here that it is easy to see that the minimum on the right-hand side of (1.15) can take either values. We now apply Theorem 1.2 to study inequalities (1.11)–(1.12). As the case \(\alpha=1\) yields the classical inequality (1.4), we concentrate on the case \(\alpha>1\) and we deduce readily from Theorem 1.2 the following result.
Theorem 1.3
Let\(\alpha\geq1\)and\(p_{1/\alpha}\)be defined as in (1.13). Then inequality (1.11) holds for all non-negative sequencesxwhen\(\alpha> 1\), \(0< p \leq p_{1/\alpha}\)and inequality (1.12) holds for all non-negative sequencesxwhen\(\alpha\geq2\), \(0< p \leq p_{1/\alpha}\). The constants are best possible in both cases.
In fact, note that [5, Lemma 2.1] implies that (1.6) is satisfied for \(\lambda_{n}=n^{\alpha}-(n-1)^{\alpha}\) with \(L_{\lambda }=1/\alpha\) when \(\alpha\geq1\) and (1.6) is satisfied for \(\lambda_{n}=n^{\alpha-1}\) with \(L_{\lambda}=1/\alpha\) when \(\alpha \geq2\). That the constant is best possible can be seen by setting \(x_{n}=n^{-1/p-\epsilon}\) and letting \(\epsilon\rightarrow0^{+}\).
2 Proof of Theorem 1.1
For the first assertion of Theorem 1.1, our goal is to find conditions on the \(\lambda_{n}\) such that the following inequality holds for \(0< p<1\), \(L>p\):
It suffices to prove the above inequality by replacing the infinite sums by finite sums from \(n=1\) to N (and \(k=n\) to N) for any integer \(N \geq1\). Note that, as in [6, Sect. 3], we have, for any positive sequence \(\mathbf{w}=(w_{n})\),
By Hölder’s inequality, we have
It follows that
Suppose we can find a sequence \(\mathbf{w}=(w_{n})\) of positive terms, such that, for any integer \(n\geq1\),
Then the desired inequality follows. Now we make a change of variables: \(w_{i} \mapsto\lambda_{i}w_{i}\) to recast the above inequality as
We now define our sequence \(\mathbf{w}=(w_{n})\) to satisfy \(w_{1}=1\) and we inductively see that, for \(n \geq2\),
This allows us to deduce that, for \(n \geq1\),
Applying (2.2), (2.3) in (2.1), we see that inequality (2.1) becomes (1.10). This completes the proof for the first assertion of Theorem 1.1.
We now prove the second assertion of Theorem 1.1. We set \(x=\lambda_{n}/\varLambda_{n}\), \(y=\lambda_{n+1}/\varLambda_{n+1}\) to recast inequality (1.10) as
To facilitate also the proof of Theorem 1.2 below, we proceed by taking the condition (1.14) into consideration to assume that L is a constant such that \(1/y \leq1/x+L+Mx\), where \(M \geq0\) is another constant. As the function \(t \mapsto (t-1)^{(1+p)/(1-p)}t^{-p/(1-p)}\) is an increasing function of \(t \geq 1\), we deduce that it suffices to prove the above inequality for \(1/y=1/x+L+Mx\), which is equivalent to showing that \(f_{L,M,p}(x) \geq0\) for \(0< x \leq1\), where
Suppose that (1.6) is valid and \(L_{\lambda} \geq1\). In this case we set \(L=L_{\lambda}\) and \(M=0\) so that it suffices to show that \(f_{L,0,p}(x) \geq0\). Calculation shows that
where
Suppose that \(0< p \leq1/3\). We want to show that \(g_{L,p}(x) \geq0\) for \(0< x \leq1\). We first note that we have
We may now assume that
Otherwise, we have trivially \(g_{L,p}(x) \geq0\). We then estimate \((1+(L-1)x)^{\frac{1-3p}{1-p}}\) trivially by \((1+(L-1)x)^{\frac {1-3p}{1-p}} \geq1\) and we apply a Taylor expansion to see that
It then follows that \(g_{L,p}(x) \geq u_{L,p}(x)\), where
Suppose first that
We then deduce that
We regard the last expression above as a linear function of x to see that its derivative with respect to x is
As we have \(0< p \leq1/3\) and \(L \geq1\), we deduce that
It follows from this that the last expression in (2.5) is minimized at \(x=0\) with corresponding value being \(u_{L,p}(0)\). On the other hand, if the inequality in (2.4) does not hold, then one checks that \(u_{L,p}(x)\) is a quadratic function of x with negative leading coefficient, hence is minimized at \(x=0\) or \(x=1\). Thus, in either case, we conclude that \(u_{L,p}(x) \geq\min\{ u_{L,p}(0), u_{L,p}(1) \}\) for \(0 \leq x \leq1\). One checks that
When we regard the above expression as a function of L, it is readily seen that \(u_{L,p}(0)\) is convex in L such that
We thus deduce that \(u_{L,p}(0) \geq0\) when \(0< p \leq1/3\). On the other hand, we have \(u_{L,p}(1)=a_{1}(L,p) \geq0\) by our assumption, where \(a_{1}(L,p)\) is defined in (1.9). It follows that \(g_{L,p}(x) \geq u_{L,p}(x) \geq0\), hence \(f''_{L,0,p}(x) \geq0\) for \(0< x \leq1\). As \(f_{L,0,p}(0)=f'_{L,0,p}(0)=0\), we then deduce that \(f_{L,0,p}(x) \geq0\) and this completes the proof for the case \(L_{\lambda} \geq1\) of the second assertion of Theorem 1.1.
To prove the case \(0< L_{\lambda} < 1\) of the second assertion of Theorem 1.1, we first note that
One checks that
It follows from this and the arithmetic–geometric mean inequality that
Suppose that (1.6) is valid and \(0< L_{\lambda}<1\). In this case we can also set \(L=L_{\lambda}\) and \(M=0\) to see that inequality (2.7) becomes
where
Calculation shows that
where
One checks that \(v_{L,p}(x)\) is a quadratic polynomial of x with a negative leading coefficient when \(L \geq2p\). It follows that \(v_{L,p}(x) \geq\min\{ v_{L,p}(0), v_{L,p}(1) \}\) for \(0< x \leq1\) and one checks that \(v_{L,p}(0)=p u_{L,p}(0)\), where \(u_{L,p}(0)\) is defined in (2.6). Similar to our discussions for the case \(L>1\), one checks that \(u_{L,p}(0)\) is convex in L such that
with \(u_{L,p}(0) \geq0\) when \(L \geq4p\). On the other hand, we have \(v_{L,p}(1)=a_{2}(L,p) \geq0\) by our assumption, where \(a_{2} (L,p)\) is defined in (1.9). It follows that \(h'_{L,p}(x) \geq0\), hence \(h_{L,p}(x) \geq h_{L,p}(0) \geq1\). It follows from this and (2.8) that \(f'_{L,0,p}(x) \geq1\) for \(0< x \leq1\). Now, since \(f_{L,0,p}(x)= 0\), this implies that \(f_{L,0,p}(x) \geq0\) for \(0< x \leq1\), which then completes the proof for the case \(0< L_{\lambda}<1\) of the second assertion of Theorem 1.1.
3 Proof of Theorem 1.2
First, we assume that (1.6) is valid and we set \(L=L_{\lambda}\) in this case. It suffices to find a value of p such that \(a_{2}(L,p) \geq0\) by Theorem 1.1. Note that \(\lim_{p \rightarrow0^{+}}a_{2}(a p,p)/p<0\) when \(a >1\), it is therefore not possible to show \(a_{2}(L,p) \geq0\) by assuming that \(p \leq L/a\) for any \(a>1\). We therefore seek to show \(a_{2}(L,p) \geq0\) for \(p \leq L^{2}/4\). We first note that
where the last inequality follows from the observation that the function \(p \mapsto4-5p-\frac{2}{1-p}\) is non-negative for \(0< p \leq1/4\). This completes the proof for the first assertion of Theorem 1.2.
To prove the second assertion of Theorem 1.2, we see from the proof of Theorem 1.1 that it suffices to show the right-hand side expression of (2.7) is ≥1 for \(0< x \leq1\). We simplify it to see that it is equivalent to showing that
We assume that
This implies that the function
is a concave function of x, hence is minimized at \(x=0\) or 1. When \(x=0\), the above function takes the value 1. We further assume that the above function takes a value ≥1 when \(x=1\). That is,
We then deduce that
We apply the above estimation and the estimation that \(1+2Mx/L \geq1\) in (3.1) to see that it suffices to show that, for \(0< x \leq1\),
We now assume that
Then calculation shows that
where
It is easy to see that \(v_{L,M,p}(x)\) is a quadratic polynomial of x with negative leading coefficient when
The above inequality is certainly valid when \(L/p=2\). We may thus assume that \(L/p>2\) to see that the above inequality is a consequence of the following inequality:
Assuming the above inequality, we see that \(v_{L,M,p}(x) \geq\min\{ v_{L,M,p}(0), v_{L,M,p}(1) \}\) for \(0< x \leq1\) and we have
We note that, when \(M < 1\),
We then deduce that \(v_{L,M,p}(0) \geq0\) when
and that \(v_{L,M,p}(1) \geq0\) when
Thus, one just needs to find values of p to satisfy inequalities (3.2)–(3.7). Indeed, if (3.2)–(3.7) hold, then \(v_{L,M,p}(x) \geq0\) for \(0 < x \leq1\), which implies that \(h'_{L,M,p}(x) \geq0\) for \(0 < x \leq1\). Now, since \(h_{L,M,p}(0) = 1\), we get \(h_{L,M,p}(x) \geq1\) for \(0 < x \leq1\) as desired.
We note that
We apply the above estimates to see that inequality (3.5) is a consequence of the following inequality:
As inequality (3.7) implies inequalities (3.6) and (3.8), we first find values of p so that inequality (3.7) holds. To do so, we first simply inequality (3.7) by noting that
Using this, we see that it suffices to find values of p that satisfy
We recast the above as
Note as \(1-L-M>0\), we have \((1-L-M)p^{2} \geq0\), so that the above inequality follows from
which is equivalent to
One checks that the above inequality implies inequalities (3.2) and (3.4). One further notes that inequality (3.3) is equivalent to
Combining inequalities (3.9) and (3.10), one readily deduces the second assertion of Theorem 1.2 and this completes the proof of Theorem 1.2.
References
Bennett, G.: Some elementary inequalities. II. Q. J. Math. Oxf. Ser. (2) 39, 385–400 (1988)
Bennett, G.: Sums of powers and the meaning of \(l^{p}\). Houst. J. Math. 32, 801–831 (2006)
Cartlidge, J.M.: Weighted mean matrices as operators on \(l^{p}\). Ph.D. thesis, Indiana University (1978)
Copson, E.T.: Note on series of positive terms. J. Lond. Math. Soc. 3, 49–51 (1928)
Gao, P.: A note on Hardy-type inequalities. Proc. Am. Math. Soc. 133, 1977–1984 (2005)
Gao, P.: Hardy-type inequalities via auxiliary sequences. J. Math. Anal. Appl. 343, 48–57 (2008)
Gao, P.: On \(l^{p}\) norms of weighted mean matrices. Math. Z. 264, 829–848 (2010)
Gao, P.: On a result of Levin and Stečkin. Int. J. Math. Math. Sci. 2011, Article ID 534391 (2011)
Gao, P.: On weighted remainder form of Hardy-type inequalities. Houst. J. Math. 38, 177–199 (2012)
Gao, P.: Extensions of Copson’s inequalities. J. Math. Anal. Appl. 401, 430–435 (2013)
Gao, P.: On \(l^{p}\) norms of factorable matrices. arXiv:1301.3262
Hardy, G.H., Littlewood, J.E., Pólya, G.: Inequalities. Cambridge University Press, Cambridge (1952)
Levin, V.I., Stečkin, S.B.: Inequalities. Transl. Am. Math. Soc. (2) 14, 1–29 (1960)
Acknowledgements
The authors are very grateful to the referees for their very careful reading of the paper and many valuable comments and suggestions.
Availability of data and materials
Data sharing is not applicable to this article as no datasets were generated or analysed during the current study.
Funding
Not applicable.
Author information
Authors and Affiliations
Contributions
The authors completed the paper, read and approved the final manuscript.
Corresponding author
Ethics declarations
Competing interests
The authors declare that there are no competing interests.
Additional information
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Gao, P., Zhao, H. On Copson’s inequalities for \(0< p<1\). J Inequal Appl 2020, 72 (2020). https://doi.org/10.1186/s13660-020-02339-3
Received:
Accepted:
Published:
DOI: https://doi.org/10.1186/s13660-020-02339-3