R E S E A R C H
Open Access
A sharpened version of Hardy’s inequality for
parameter
p
= /
Yongping Deng
1, Shanhe Wu
1*and Deng He
2*Correspondence: [email protected] 1Department of Mathematics and Computer Science, Longyan University, Longyan, Fujian 364012, P.R. China
Full list of author information is available at the end of the article
Abstract
The purpose of this paper is to investigate a sharpened version of Hardy’s inequality for parameterp= 5/4. By evaluating the weight coefficientW(k, 5/4), sharpened Hardy’s inequality that contains the best coefficient
η
5/4= 0.46. . .is established. MSC: 26D15; 26D20; 26D07Keywords: Hardy’s inequality; weight coefficient; sharpened inequality; exponential parameter; best coefficient
1 Introduction
Letp> , /p+ /q= ,an≥ (n= , , . . .), <
∞
n=a p
n<∞. Then
∞
n=
n n
k= ak
p
<qp
∞
n=
apn, ()
whereqp= ( p p–)
pis the best coefficient. Inequality () is called Hardy’s inequality which is
of great use in the field of modern mathematics (see [, ]). A special case of () yields the following inequalities:
∞
n=
n n
k= ak
<
∞
n=
an, ()
∞
n=
n n
k= ak
<
∞
n=
an. ()
In , Yang and Zhu [] evaluated the weight coefficientW(k,p),
W(k,p) =k–/p
∞
n=k
np
n
j=
j/p
p–
, k= , , . . . , ()
and established an improved version of inequality () as follows:
∞
n=
n n
k= ak
<
∞
n=
– √n+
an. ()
With the same approach, that is, evaluating the weight coefficientW(k,p), Huang [–] gave some improvements on Hardy’s inequality forp= andp= /,i.e.,
∞
n=
n n
k= ak
<
∞
n=
– n/
an, ()
∞
n=
n n
k= ak
/
≤√
∞
n=
– ·
√
n+
a/n . ()
Some further extensions of Hardy’s inequality related to the range of parameterpwere given in Huang [, ].
In , Yang [] proved an inequality for the weight coefficientW(k, )
W(k, ) =√k
∞
n=k
n
n
j=
j
≤ –√
k
– W(, )
and established the following inequality:
∞
n=
n n
k= ak
<
∞
n=
–√θ
n
an, ()
whereθ= –W(, ) = . . . . is the best coefficient under the weight coefficient W(k, ).
In , Zhang and Xu made use of the monotonicity theorem [–] and obtained an improvement of inequality ():
∞
n=
n n
k= ak
p
≤
p p–
p∞
n=
– cp (n– /)–/p
apn, ()
where
cp=
(p– )[ – /p( – /p)], <p≤,
– –/p( – /p)p–, p> .
By evaluating the weight coefficientW(k,p), and with the help of an inequality-proving package called BOTTEMA [, ], He [] investigated a sharpened version of Hardy’s inequality forp∈Nand obtained the following improved version of inequality ():
∞
n=
n n
k= ak
≤
∞
n=
– θ
n/
an, ()
whereθ= – W(, ) = . . . . is the best coefficient under the weight coefficient W(k, ).
numbers):
∞
n=
n n
k= ak
p
≤
p p–
p∞
n=
– θp()
n–/p
apn, ()
whereθp() = – (p–p )pW(,p) is the best coefficient of () under the weight coefficient W(k,p).
Recently, based on the program HDISCOVER written by Deng, He and Wu [], an automated verification of inequality () is achieved forp∈Q(Qis the set of rational numbers).
For more detailed information of Hardy’s inequality, we refer the interested readers to relevant research papers [, , –].
In this paper, by evaluating the weight coefficientW(k, /), we establish an improve-ment of Hardy’s inequality for parameterp= / as follows:
∞
n=
n n
k= ak
/
≤/
∞
n=
– ·
n/+η /
a/n , ()
whereη/= /
[/–W(,/)]– = . . . . is the best coefficient under the weight coefficient W(k, /).
2 Lemmas
To prove the main results in Section , we will use the following lemmas.
Lemma (see[]) If p> ,then for all integers n≥,it holds that p
p– n
–/p– p p– +
+
n/p
≤
n
j=
j/p ≤ p p– n
–/p– p p– +
+
n/p +
p–
pn+/p. ()
Lemma (see[]) If p> ,then for all integers n≥k≥,it holds that
(p– )kp–+
kp <
∞
n=k
np <
(p– )kp–+
kp +
p
kp+.
Lemma Let p> , /p+ /q= ,and let gr,glbe the functions defined by
gr(x) = –
p+ q– pqx/q +
qx–
pqx, gl(x) = – p+ q
pqx/q +
qx, x∈[, +∞). Then– <gr(x) < , – <gl(x) < .
Further, we have
gr(x) = p+ q– pqx+/q –
qx+
pqx
≥p+ q–
pqx –
qx +
pqx
=(px+x+ p)(p– ) px > ,
and consequently,gris strictly increasing on [, +∞).
Now, fromgr() = –/p> – andlimx→+∞gr(x) = , it follows thatgr(x)≥gr() = –/p>
– andgr(x) < .
Similarly, from
gl(x) = p+ q pqx+/q –
qx ≥
p+ q
pqx –
qx =
p– px > , gl() = –/p> – and lim
x→+∞gl(x) = ,
we deduce that – <gl(x) < .
Lemma is proved.
Lemma Let– <g(x) < .Ifα∈(, ],then
+g(x) + (α– )g(x) +(α– )(α– ) g
(x)
≤ +g(x)α≤ +αg(x) +α(α– ) g
(x).
Ifα∈[, ],then
+g(x)α≥ +αg(x) +α(α– ) g
(x).
Proof Whenα∈(, ]. By using the Maclaurin formula
+g(x)α= +αg(x) +α(α– ) g
(x)
+α(α– )(α– )( +θg(x))
α–
g
(x), θ∈(, ),
and noticing – <g(x) < , we find
+θg(x) > +g(x) > ,
α(α– )(α– )( +θg(x))α–
g
(x)≤,
(α– )(α– )(α– )( +θg(x))α–
g
Thus
+g(x)α≤ +αg(x) +α(α– ) g
(x),
+g(x)α= +g(x) +g(x)α–
≥ +g(x) + (α– )g(x) +(α– )(α– ) g
(x)
.
Whenα∈[, ]. We have
α(α– )(α– )( +θg(x))α–
g
(x)≥.
Thus
+g(x)α≥ +αg(x) +α(α– ) g
(x).
The proof of Lemma is complete.
Lemma Let p> , /p+ /q= ,n≥k≥,and let[x]denote the greatest integer less than or equal to the real number x.Then we have
W(k,p)≤qp–k/q
∞
n=k
n+/q
+gr(n)
[p]–
×
+p– [p]gr(n) +
(p– [p])(p– [p] – ) g
r(n)
.
Proof By Lemma and the identitypq=p+q,q(p– ) =p, it follows that
W(k,p) =k–/p
∞
n=k
np
n
j=
j/p
p–
≤k–/p
∞
n=k
np
p p– n
–/p– p p– +
+
n/p+
p–
pn+/p
p–
=k/q
∞
n=k
np
qn/q–p+ q– p +
n/p –
pn+/p
p–
=k/q
∞
n=k
npq
p–n(p–)/q
–p+ q– pqn/q +
qn –
pqn
p–
=qp–k/q
∞
n=k
n+/q
+gr(n)
p–
.
Combining Lemmas and , we obtain
W(k,p)
≤qp–k/q
∞
n=k
n+/q
+gr(n)
[p]–
+gr(n)
≤qp–k/q
∞
n=k
n+/q
+gr(n)
[p]–
×
+p– [p]gr(n) +
(p– [p])(p– [p] – ) g
r(n)
.
This completes the proof of Lemma .
Lemma Let/p+ /q= ,n≥k≥.If p∈(, ),then
W(k,p)≥qp–k/q
∞
n=k
n+/q
+gl(n)
[p]
×
+p– [p] – gl(n) +
(p– [p] – )(p– [p] – )
g
l(n)
.
If p∈[, +∞),then
W(k,p)≥qp–k/q
∞
n=k
n+/q
+gl(n)
[p]–
×
+p– [p] + gl(n) +
(p– [p] + )(p– [p]) g
l(n)
.
Proof Sincepq=p+q,q(p– ) =p, using Lemma gives
W(k,p) =k–/p
∞
n=k
np
n
j=
j/p
p–
≥k–/p
∞
n=k
np
p p– n
–/p– p p– +
+
n/p
p–
=k/q
∞
n=k
np
qn/q–p+ q p +
n/p
p–
=k/q
∞
n=k
npq
p–n(p–)/q
– p+ q pqn/q+
qn
p–
=qp–k/q
∞
n=k
n+/q
+gl(n)
p–
.
Whenp∈(, ). From Lemmas and , we have
W(k,p)≥qp–k/q
∞
n=k
n+/q
+gl(n)
[p]–
+gl(n)
p–[p]
≥qp–k/q
∞
n=k
n+/q
+gl(n)
[p]
×
+p– [p] – gl(n) +
(p– [p] – )(p– [p] – )
g
l(n)
Whenp∈[, +∞). Using Lemmas and , we obtain
W(k,p)≥qp–k/q
∞
n=k
n+/q
+gl(n)
[p]–
+gl(n)
p–[p]+
≥qp–k/q
∞
n=k
n+/q
+gl(n)
[p]–
×
+p– [p] + gl(n) +
(p– [p] + )(p– [p]) g
l(n)
.
Lemma is proved.
Lemma (see[]) Let p> ,an≥ (n= , , . . .), <
∞
n=a p
n<∞.Then
∞ n= n n k= ak p ≤ ∞ k=
k–/p
∞
n=k
np n j=
j/p
p– apk
=
∞
k=
W(k,p)apk.
3 Main results
Theorem For an arbitrary natural number k,the following inequality holds true:
W(k, /) <R/(k),
where
R/(k) = /
– k/–
, ,k/ +
k–
k/ –
, ,k/ +
,k
+ ,k/ +
,k +
,k/+
,k+
,k
.
Proof Using Lemma gives
W(k, /)≤/k/
∞
n=k
n/
+ gr(n) –
g
r(n)
= /k/
∞
n=k r/(n),
where
r/(n) =
n/ –
n/ –
, ,n/+
n/ +
,n/–
,n/
– ,n/ +
,n/ –
,n/.
Hence
W(k, /)≤/k/
∞
n=k
n/ –
n/–
, ,n/+
n/ +
,n/
– ,n/ –
,n/+
,n/–
,n/
Using Lemma and takingp= /, /, /, /, /, /, /, /, / in the right-hand side of inequality (), respectively, we get
∞
n=k
n/ <
k/ +
k/ +
k/,
–
∞
n=k
n/ < –
k/ –
,k/,
. . . ,
∞
n=k
,n/ <
,k/ +
,k/ +
,k/,
–
∞
n=k
,n/< –
,k/ –
,k/.
Adding up the above inequalities, we obtain
W(k, /) <R/(k).
Theorem is proved.
Theorem For an arbitrary natural number k,the following inequality holds true:
W(k, /) >L/(k), where
L/(k) = /
– k/–
k/ –
, ,k/+
k–
k/ +
, ,k/
– , ,k/+
,k –
,k/+
,k/–
, ,k/
– ,k–
,k/+
,k–
, ,k/
.
Proof Utilizing Lemma gives
W(k, /)≥/k/
∞
n=k
n/
+gl(n)
– gl(n) +
g
l(n)
= /k/
∞
n=k l/(n),
where
l/(n) =
n/ –
n/–
,n/ –
, ,n/ +
n/ +
,n/
+ , ,n/ –
,n/ –
,n/ +
Hence
W(k, /)≥/k/
∞
n=k
n/ –
n/–
,n/ –
, ,n/ +
n/
+ ,n/+
, ,n/ –
,n/
– ,n/ +
,n/
.
Using Lemma and takingp= /, /, /, /, /, /, /, /, /, / in the left-hand side of inequality (), respectively, we get
∞
n=k
n/ >
k/ +
k/,
–
∞
n=k
n/ > –
k/–
k/–
k/,
. . . , –
∞
n=k
,n/> –
,k/–
,k/–
, ,k/,
∞
n=k
,n/ >
,k/ +
,k/.
Adding up the above inequalities, we obtain
W(k, /) >L/(k).
Theorem is proved.
Theorem Let an≥ (n= , , . . .), <
∞
n=a/n <∞.Then
∞
n=
n n
k= ak
/
≤/
∞
n=
– ·
n/+η /
a/
n , ()
whereη/= /
[/–W(,/)] – = . . . .is the best possible under the weight coefficient W(k, /).
Proof By Lemma , we have
∞
n=
n n
k= ak
/
≤ ∞
k=
W(k, /)a/k .
Therefore, to prove inequality (), it suffices to show that
W(k, /)≤/
– ·
k/+η /
Obviously, inequality () becomes an equality fork= . In what follows, we will assume thatk≥.
By Theorem W(k, /) <R/(k), we need only to prove that
R/(k)≤/
– ·
k/+η /
.
Note that
η/=
/
[/–W(, /)]– = . . . . >
, it suffices to show
R/(k)≤/
– ·
k/+ /
. ()
Substitutingk=xin (), inequality () becomes
R/
x≤/
– ·
x+ /
, wherex≥√,
which is equivalent to the following inequality:
/
– x–
, ,x +
x –
x–
, ,x +
,x+
,x
+ ,x +
,x+
,x+
,x
≤/
– ·
x+ /
()
⇔ –
x–
, ,x +
x –
x–
, ,x +
,x+
,x
+ ,x +
,x+
,x+
,x ≤ –
·
x+ /
⇔–
x–
, ,x +
x –
x–
, ,x +
,x+
,x
+ ,x +
,x+
,x+
,x+
·
x+ /≤
⇔– f(x)
,,x(x+ )≤,
where
f(x) = ,,x+ ,,,x+ ,,x– ,,,x
– ,,x+ ,,x+ ,,x– ,,x
– ,,x– ,,x– ,,x– ,,x
From the hypothesisx≥√ > ., we have
,,x+ ,,,x+ ,,x– ,,,x
– ,,x+ ,,x+ ,,x
>,,×.+ ,,,×.+ ,,×. – ,,,x– ,,x+ ,,x+ ,,x
= ,,,x– ,,x+ ,,x+ ,,x
=(,,,x– ,,)x+ ,,x+ ,,x
>(,,,×. – ,,)×. + ,,×. + ,,x
= ,,,x.
Further, we have
f(x) > ,,,x– ,,x– ,,x
– ,,x– ,,x– ,,x
– ,,x– ,x– ,x– ,x– , > ,,,x– ,,x– ,,x
– ,,x– ,,x
– ,,x– ,,x– ,x– ,x
– ,x– ,x
= ,,x> .
Consequently, inequality () holds true, and inequality () is proved. Let us now show thatη/=
/
[/–W(,/)] – = . . . . is the best possible under the
weight coefficientW(k, /).
Consider inequality () in a general form as
W(k, /)≤/
– ·
k/+η /
. ()
Puttingk= in () yields
η/≥
/
[/–W(, /)].
Thus the best possible value forη/in () should beηmin= / [/–W(,/)].
Remark From the definition ofW(k,p) and in the same way as in [], we can establish the following accurate estimates ofW(, /):
. <W(, /) < .. () Further, the approximation ofη/can be derived as follows:
η/=
/
[/–W(, /)]– = . . . . .
Remark Forp= /, inequality () can be written as
∞
n=
n n
k= ak
/
≤/
∞
n=
– – (
)
/W(, /) n/
a/n . ()
It is easy to observe that
·
n/+η/>
·
n/+ ,/,>
n/
and
–
/
W(, /) < –
/
×. = . . . . < , hence
– ()/W(, /) n/ <
·
n/+η /
.
This implies that inequality () is stronger than inequality ().
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
YD finished the proof and the writing work. SW gave YD some advice on the proof and writing. DH gave YD lots of help in revising the paper. All authors read and approved the final manuscript.
Author details
1Department of Mathematics and Computer Science, Longyan University, Longyan, Fujian 364012, P.R. China.2Gangtou Middle School, Fuqing, Fujian 350317, P.R. China.
Acknowledgements
The present investigation was supported, in part, by the Natural Science Foundation of Fujian province of China under Grant 2012J01014 and, in part, by the Foundation of Scientific Research Project of Fujian Province Education Department of China under Grant JK2012049.
Received: 10 December 2012 Accepted: 1 February 2013 Published: 20 February 2013
References
1. Hardy, GH, Littlewood, JE, Pólya, G: Inequalities, 2nd edn. Cambridge University Press, Cambridge (1952) 2. Mitrinovi´c, DS, Peˇcari´c, JE, Fink, AM: Inequalities Involving Functions and Thier Integrals and Derivatives. Kluwer
Academic, Dordrecht (1991)
5. Huang, QL: A sharpness and improvement of Hardy’s inequality forp= 3/2. J. Guangxi Norm. Univ. (Nat. Sci. Ed.)18(1), 38-41 (2000)
6. Huang, QL: An improvement of Hardy’s inequality in an interval. Acta Sci. Natur. Univ. Sunyatseni39(3), 20-24 (2000) 7. Huang, QL: An improvement of Hardy’s series inequality in the interval [2, 5]. J. South China Univ. Technol. (Natl. Sci.
Ed.)28(2), 64-68 (2000)
8. Huang, QL: An extension of a strengthened inequality. J. Guangdong Educ. Inst.21(2), 17-20 (2001) 9. Yang, BC: On a strengthened version of Hardys inequality. J. Guangdong Educ. Inst.25(5), 5-8 (2005) 10. Kuang, JC: Applied Inequalities, 4th edn., p. 576. Shangdong Science and Technology Press, Jinan (2010) 11. Zhang, XM: A new proof method of analytic inequality. RGMIA Research Report Collection12(1), 18-29 (2009) 12. Zhang, XM, Chu, YM: New Discussion to Analytic Inequalities, pp. 260-275. Harbin Institute of Technology press,
Harbin (2009)
13. Zhang, XM, Chu, YM: A new method to study analytic inequalities. J. Inequal. Appl. (2010). http://www.hindawi.com/journals/jia/2010/698012
14. Yang, L, Xia, BC: Automated Discovering and Proving for Mathematical Inequalities, pp. 33-42, 117-142. Science Press, Beijing (2008)
15. Yang, L, Zhang, JZ, Hou, XR: Nonlinear Algebraic Equation Systems and Automated Theorem Proving, pp. 137-166. Shanghai Science’s and Technology’s Education Publishing House, Shanghai (1996)
16. He, D: The automatic verification of Hardy inequality’s improvements and its improving type ofP∈N. J. Guangdong Educ. Inst.32(3), 28-35 (2012)
17. Deng, YP, He, D, Wu, SH: A sharpened version of Hardy’s inequality and the automated verification program. J. Math. Pract. Theory (2012, submitted)
18. Huang, YZ, He, D: An improvement of Hardy’s inequality forp= 3/2. J. Shantou Univ. (Natl. Sci. Ed.)27(3), 15-22 (2012) 19. Yang, BC: On a strengthened Hardy’s inequality. J. Xinyang Teach. Coll. (Natl. Sci. Ed.)15(1), 37-39 (2002)
20. Huang, QL: A strengthened Hardy’s inequality in the interval [1, 2]. J. Guangdong Educ. Inst.22(2), 20-23 (2002) 21. Huang, QL, Yang, BC: A strengthened dual form of Hardy’s inequality. J. Math. (PRC)25(3), 307-311 (2005) 22. Zhu, YH, Yang, BC: Improvement on Euler’s summation formula and some inequalities on sums of powers. Acta Sci.
Natur. Univ. Sunyatseni36(4), 21-26 (1997)
23. Qian, X, Meixiu, Z, Zhang, X: On a strengthened version of Hardy’s inequality. J. Inequal. Appl.2012, 300 (2012) doi:10.1186/1029-242X-2013-63