• 検索結果がありません。

Memoirs on Differential Equations and Mathematical Physics Volume 73, 2018, 131–140

N/A
N/A
Protected

Academic year: 2022

シェア "Memoirs on Differential Equations and Mathematical Physics Volume 73, 2018, 131–140"

Copied!
10
0
0

読み込み中.... (全文を見る)

全文

(1)

Memoirs on Differential Equations and Mathematical Physics

Volume 73, 2018, 131–140

S. H. Saker, D. O’Regan, M. R. Kenawy, R. P. Agarwal

FRACTIONAL HARDY TYPE INEQUALITIES VIA CONFORMABLE CALCULUS

(2)

2010 Mathematics Subject Classification. 26A33, 26D10.

Key words and phrases. Hardy inequality, Conformable fractional derivative, Conformable frac- tional integral, Hölder inequality.

ÒÄÆÉÖÌÄ. ÃÀÃÂÄÍÉËÉÀ äÀÒÃÉÓ ÊËÀÓÉÊÖÒÉ ÉÍÔÄÂÒÀËÖÒ ÖÔÏËÏÁÀÈÀα-ßÉËÀÃÖÒÉ ÀÍÀËÏÂÄÁÉ.

(3)

Fractional Hardy Type Inequalities via Conformable Calculus 133

1 Introduction

In 1925, Hardy [4] used the calculus of variations to prove the inequality

0

(1 x

x 0

f(t)dt )p

dx≤( p p−1

)p 0

fp(x)dx, (1.1)

wheref 0is integrable over any finite interval(0, x)andfpis integrable and convergent over(0,) andp >1. The constant(p/(p1))p is the best possible.

In 1928, Hardy [5] generalized inequality (1.1) and proved that ifp >1 andf is non-negative for x≥0,then

0

xc (∫x

0

f(t)dt )p

dx≤( p c−1

)p 0

xpcfp(x)dx for c >1, (1.2)

and ∫

0

xc (∫

x

f(t)dt )p

dx≤( p 1−c

)p 0

xpcfp(x)dx for c <1. (1.3) The constants(p/(c1))p and(p/(1−c))p are the best possible.

In recent years, fractional inequalities were studied by using the fractional Caputo and Riemann–

Liouville derivative; for details, we refer the reader to [3] and [17]. In [1] and [7], the authors presented conformable calculus and classical inequalities with the use of conformable fractional calculus such as Opial’s inequality (see [11] and [12]), Hermite–Hadamard’s inequality (see [8] and [10]), Chebyshev’s inequality (see [2]) and Steffensen’s inequality (see [13]). In this paper, using a somewhat different approach we present new Hardy type inequalities via conformable fractional calculus. Also, one can see from our approach and presentation that the conformable fractional inequalities encountered in the literature are, in fact, special cases of weighted inequalities (for an appropriate weight function).

Our goal in this paper is, first, to show how naturally weights work in inequalities and, second, to indicate and correct some slight mistakes (usually when one integrates by parts) in the literature.

The paper is organized as follows. In Section 2, we present some concepts on conformable fractional calculus and also Hölder’s inequality forα-fractional differentiable functions which we will use to prove our main results. In Section 3, we prove some Hardy type inequalities forα-fractional differentiable functions and obtain the classical ones as special cases whenα= 1.

2 Basic concepts and lemmas

In this section, we present some basic definitions concerning conformable fractional calculus. For more details, we refer the reader to [1] and [7].

Definition 2.1. Letf : [0,) R. Then the conformable fractional derivative of orderα off is defined by

Dαf(t) =lim

ϵ0

f(t+ϵt1α)−f(t) ϵ

for allt >0and 0< α≤1, and

Dαf(0) = lim

t0+

Dαf(t).

Letα∈(0,1]andf, g beα-differentiable at a point t. Then

Dα(f g) =f Dαg+gDαf. (2.1)

Further, letα∈(0,1]andf, g be α-differentiable at a point t, withg(t)̸= 0. Then Dα

(f g

)

=gDαf−f Dαg

g2 . (2.2)

(4)

Remark 2.1. Iff is a differentiable function, then

Dαf(t) =t1αdf(t) dt .

Definition 2.2. Let f : [0,) R. Then the conformable fractional integral of order α of f is defined by

Iαf(t) =

t 0

f(x)dαx=

t 0

xα1f(x)dx (2.3)

for allt >0and0< α≤1.

Now, we state an integration by parts formula (see [1] and [7]) which is immediate.

Lemma 2.1. Assume thatw, g: [0,)R are two functions such thatw, g are differentiable and 0< α≤1. Then for any b >0,

b 0

w(x)Dαg(x)dαx=w(x)g(x)|b0

b 0

g(x)Dαw(x)dαx. (2.4)

Next, we prove the Hölder type inequality needed in the next section (of course, it is the usual Hölder inequality for the functions under consideration (i.e., xp1)f(x) and xq1)g(x)); for com- pleteness we include its proof).

Lemma 2.2. Let f, g: [0,)Rand0< α≤1. Then for anyb >0,

b 0

|f(x)g(x)|dαx≤ (∫b

0

|f(x)|pdαx )p1(∫b

0

|g(x)|qdαx )1q

, (2.5)

where1/p+ 1/q= 1 (provided the integrals exist(and are finite)).

Proof. For nonnegative real numbersβ, γ, the classical Young inequality is β1pγq1 β

p+γ p. Suppose now, without loss of generality, that

b 0

|f(x)|pdα= 0 and

b 0

|g(x)|qdα= 0.

Applying Young’s inequality with

β = |f(x)|p

b 0

|f(x)|pdαx

, γ= |g(x)|q

b 0

|g(x)|q dαx ,

and integrating the obtained inequality from0to b, we get

b 0

|f(x)| (∫b

0

|f(s)|pdαs)1p

|g(x)| (∫b

0

|g(s)|qdαs)1q dαx

=

b 0

β1p(x)γ1q(x)dαx≤

b 0

(β p+γ

q )

dαx

(5)

Fractional Hardy Type Inequalities via Conformable Calculus 135

=

b 0

( |f(x)|p

p(∫b

0

|f(s)|pdαs)+ |g(x)|q q(∫b

0

|g(s)|qdαs) )

dαx

=

b 0

|f(x)|pdαx

p(∫b

0

|f(s)|pdαs)+

b 0

|g(x)|qdαx

q(∫b

0

|g(s)|qdαs)= 1 p+1

q = 1,

which is the desired inequality (2.5).

3 Hardy type inequalities of α-fractional order

In this section, we state and prove the main results of this paper and we begin with the fractional version of the classical Hardy type inequality. Throughout the paper, we will assume that the functions are nonnegative locallyα-integrable and the integrals throughout are assumed to exist (and are finite, i.e., convergent).

Theorem 3.1. Let f be a nonnegative function on (0,), and 0< α≤1 and p >1. Also assume xα1f(x)is continuous on[0,). Then

0

(1 x

x 0

f(s)dαs )p

dαx≤( p p−α

)p 0

(xα1f(x))pdαx. (3.1) Proof. Let

F(x) := 1 x

x 0

f(s)dαs. (3.2)

Integrating by parts, see formula (2.4) with w(x) = Fp(x)and Dαg(x) = 1 (note here g(x) = xαα), and using Remark 2.1, we obtain (heret >0)

t 0

Fp(x)dαx= Fp(x)xα α

t

0

t 0

xα

α DαFp(x)dαx= tαFp(t) α p

α

t 0

xα

α x1αFp1(x)F(x)dαx

= tαFp(t) α −p

α

t 0

xFp1(x)F(x)dαx; (3.3)

note

lim

x0+

xαpF(x) = lim

x0+

x 0

sα1f(s)ds xppα

= lim

x0+

xα1f(x)

(ppα)xαp = lim

x0+

( p p−α

)

xα1f(x)xαp = 0.

From the definition ofF, we see that

xF(x) =xα1f(x)−F(x), and substituting it into (3.3), we obtain

t 0

Fp(x)dαx= tαFp(t) α p

α

t 0

Fp1(x)(

xα1f(x)−F(x)) dαx

= tαFp(t) α p

α

t 0

xα1Fp1(x)f(x)dαx+ p α

t 0

Fp(x)dαx,

(6)

and so

( 1 p

α )∫t

0

Fp(x)dαx= tαFp(t) α −p

α

t 0

xα1Fp1(x)f(x)dαx.

Thus

t 0

Fp(x)dαx=tαFp(t) α−p + p

p−α

t 0

xα1Fp1(x)f(x)dαx.

Applying Hölder’s inequality with indicespandp/(p−1), and using the fact thattαFp(x)/(α−p)is negative, we get

t 0

Fp(x)dαx≤ p p−α

(∫t

0

Fp(x)dαx

)pp1(∫t

0

(xα1f(x))p

dαx )1p

,

and so,

(∫t

0

Fp(x)dαx )p1

p p−α

(∫t

0

(xα1f(x))p

dαx )p1

.

Hence,

t 0

Fp(x)dαx≤( p p−α

)p

t 0

(xα1f(x))p

dαx.

Lett→ ∞, and then

0

Fp(x)dαx≤( p p−α

)p 0

(xα1f(x))p

dαx,

which is the desired inequality (3.1).

Remark 3.1. From the proof of Theorem 3.1 we see that if the condition “xα1f(x) is continuous on[0,)” is replaced either by

(i) xα1f(x)is continuous on(0,)and lim

x0+xα1+αpf(x) = 0, or

(ii) lim

x0+xαFp(x) = 0, then (3.1) is again true.

Corollary 3.1. In Theorem3.1, if α= 1,then we obtain the classical Hardy inequality (1.1).

Theorem 3.2. Let f be a nonnegative function on (0,)and0< α≤1. Letc >1andp >1. Also assume thatxα1f(x)is continuous on[0,)andp > c−α. Then

0

xc (∫x

0

f(t)dαt )p

dαx≤( p c−α

)p 0

(xαpcf(x))pdαx. (3.4)

Proof. Let

F(x) :=

x 0

f(s)dαs.

(7)

Fractional Hardy Type Inequalities via Conformable Calculus 137

Integrating by parts

t 0

xcFp(x)dαx(heret >0) with w(x) =xcFp(x), Dαg(x) = 1

(

g(x) =xα α

) , Dαw(x) =x1α(

−cxc1Fp(x) +pxcFp1(x)F(x)) , we obtain

t 0

xcFp(x)dαx=xcFp(x)xα α

t

0

t 0

xα αx1α(

−cxc1Fp(x) +pxcFp1(x)F(x)) dαx

=tαcFp(t)

α + c

α

t 0

xcFp(x)dαx− p α

t 0

x1cFp1(x)F(x)dαx;

note

lim

x0+

xα−cp F(x) = lim

x0+

x 0

sα1f(s)ds xcpα

= lim

x0+

xα1f(x) (cpα)xc−αp 1

= lim

x0+

( p c−α

)

xα1f(x)x1+α−cp = 0.

Thus ∫t

0

xcFp(x)dαx=tαcFp(t) α−c + p

c−α

t 0

x1cFp1(x)F(x)dαx.

SinceF(x) =xα1f(x), we obtain

t 0

xcFp(x)dαx= tαcFp(t) α−c + p

c−α

t 0

x1cFp1(x)xα1f(x)dαx

p c−α

t 0

xαcFp1(x)f(x)dαx≤ p c−α

t 0

xαcFp1(x)f(x)dαx

p c−α

t 0

xαc Fp1(x) (xc)p−1p (xc)p−1p

f(x)dαx≤ p c−α

t 0

xαc (xc)p−1p

((xcFp(x)))p−1p

f(x)dαx

p c−α

t 0

xαpc(

xcFp(x))p−1p

f(x)dαx.

Applying Hölder’s inequality with indicespandp/(p−1), we obtain (noteα−c <0)

t 0

xcFp(x)dαx≤ p c−α

(∫t

0

(

(xcFp(x))p−1p )p−1p

dαx

)p−1p (∫t

0

(xαcpf(x))pdαx )1p

p c−α

(∫t

0

xcFp(x)dαx

)p−1p (∫t

0

(xαpcf(x))pdαx )p1

.

Thus (∫t

0

xcFp(x)dαx )1p

( p c−α

)(∫t

0

(xαcpf(x))pdαx )1p

,

(8)

and so

t 0

xcFp(x)dαx≤( p c−α

)p

t 0

(xαpcf(x))pdαx.

Lett→ ∞, and then

0

xcFp(x)dαx≤( p c−α

)p 0

(xαpcf(x))pdαx, which is the desired inequality (3.4).

Remark 3.2. From the proof of Theorem 3.2 we see that if the condition “xα1f(x) is continuous on[0,)” is replaced either by

(i) xα1f(x)is continuous on(0,)and lim

x0+

xα+α−cp f(x) = 0, or

(ii) lim

x0+

xαcFp(x) = 0, then (3.4) is again true.

Corollary 3.2. In Theorem3.2, if α= 1, then we have the weighted Hardy inequality (1.2).

Corollary 3.3. In Theorem3.2, ifc=pandα= 1, then we have the classical Hardy inequality(1.1).

Theorem 3.3. Letf be a nonnegative function on(0,)and0< c < α≤1. Letp >1. In addition, assume thatxα1f(x)is continuous on(0,)and lim

t→∞tα+α−cp f(t) = 0. Then

0

xc (∫

x

f(t)dαt )p

dαx≤( p α−c

)p 0

(xαpcf(x))pdαx. (3.5)

Proof. LetF(x) :=

x

f(s)dαs=∫

x

sα1f(s)dsand integrate by parts the term

t ϵ

xcFp(x)dαx(here t >0 and0< ϵ < t small) with

w(x) =xcFp(x), Dαg(x) = 1 (

g(x) =xα α

) , Dαw(x) =x1α(

−cxc1Fp(x) +pxcFp1(x)F(x)) . Then we obtain

t ϵ

xcFp(x)dαx= xcFp(x)xα α

t

ϵ

t ϵ

xα α x1α(

−cxc1Fp(x) +pxcFp1(x)F(x)) dαx

= tαcFp(t)

α −ϵαcFp(ϵ)

α + c

α

t ϵ

xcFp(x)dαx− p α

t ϵ

x1cFp1(x)F(x)dαx

tαcFp(t)

α + c

α

t ϵ

xcFp(x)dαx−p α

t ϵ

x1cFp1(x)F(x)dαx,

and therefore (lettingϵ→0+), sinceα−c >0, we have

t 0

xcFp(x)dαx≤tαcFp(t) α−c p

α−c

t 0

x1cFp1(x)F(x)dαx.

(9)

Fractional Hardy Type Inequalities via Conformable Calculus 139

SinceF(x) =−xα1f(x),we obtain

t 0

xcFp(x)dαx≤ tαcFp(t) α−c + p

α−c

t 0

x1cFp1(x)xα1f(x)dαx

= tαcFp(t) α−c + p

α−c

t 0

xαcFp1(x)f(x)dαx= tαcFp(t) α−c +

t 0

xαc Fp1(x) (xc)p−1p .(xc)p−1p

f(x)dαx

=tαcFp(t) α−c +

t 0

xαc (xc)p−1p

((xcF(x))p)p−1p

f(x)dαx

= tαcFp(t) α−c +

t 0

xαcp(xcFp(x))p−1p f(x)dαx.

Applying Hölder’s inequality with indicespandp/(p−1), we obtain

t 0

xcFp(x)dαx≤ tαcFp(t) α−c + p

α−c (∫t

0

((xcFp(x))p−1p )p−1p dαx

)p−1p (∫t

0

(xαcpf(x))p

dαx )1p

tαcFp(t) α−c + p

α−c (∫t

0

xcFp(x)dαx

)p−1p (∫t

0

(xαpcf(x))p

dαx )1p

,

so, (∫t

0

xcFp(x)dαx )1p

≤tαcFp(t) α−c +

( p α−c

)(∫t

0

(xαcpf(x))pdαx )1p

.

Thus ∫t

0

xcFp(x)dαx≤ tαcFp(t) α−c +

( p α−c

)p

t 0

(xαpcf(x))pdαx.

Lett→ ∞and note

tlim→∞tα−cp F(t) = lim

t→∞

t

sα1f(s) tc−αp

ds= lim

t→∞ tα1f(t) (cpα)tc−αp 1

= lim

t→∞

( p c−α

)

f(t)tα+α−cp = 0,

so, ∫

0

xcFp(x)dαx≤( p α−c

)p 0

(xαpcf(x))pdαx, which is the desired inequality (3.5).

Corollary 3.4. In Theorem3.3, if α= 1, then we have the weighted Hardy inequality (1.3).

References

[1] T. Abdeljawad, On conformable fractional calculus.J. Comput. Appl. Math.279(2015), 57–66.

[2] A. Akkurt, M. E. Yildirim and H. Yildirim, On some integral inequalities for conformable frac- tional integrals.RGMIA Research Report Collection19(2016), Article 107, 8 pp.

[3] K. Bogdan and B. Dyda, The best constant in a fractional Hardy inequality. Math. Nachr.284 (2011), no. 5-6, 629–638.

(10)

[4] G. H. Hardy, Notes on some points in the integral calculus. LX. An inequality between integrals.

Messenger of Math.54(1925), 150–156.

[5] G. H. Hardy, Notes on some points in the integral calculus. LXIV. Further inequalities between integrals.Messenger of Math.57(1928), 12–16.

[6] S. Iqbal, K. K. Himmelreich and J. Pečarić, Weighted Hardy-type inequalities for monotone convex functions with some applications.Fract. Differ. Calc.3(2013), no. 1, 31–53.

[7] R. Khalil, M. Al Horani, A. Yousef and M. Sababheh, A new definition of fractional derivative.

J. Comput. Appl. Math.264(2014), 65–70.

[8] M. A. Khan, T. Ali, S. S. Dragomir and M. Z. Sarikaya, Hermite-Hadamard type inequalities for conformable fractional integrals,RGMIA Research Report Collection 19(2016), Article 138, 16 pp.

[9] F. J. Martin-Reyes and E. Sawyer, Weighted inequalities for Riemann-Liouville fractional inte- grals of order one and greater.Proc. Amer. Math. Soc.106(1989), no. 3, 727–733.

[10] M. Z. Sarikaya, A. Akkurt, H. Budak, M. E. Yildirim and H. Yildirim, Hermite–Hadamard’s inequalities for conformable fractional integrals.RGMIA Research Report Collection 19(2016), Article 83, 15 pp.

[11] M. Z. Sarikaya and H. Budak, New inequalities of Opial type for conformable fractional integrals, RGMIA Research Report Collection19(2016) Article 104, 10 pp.

[12] M. Z. Sarikaya and H. Budak, Opial type inequalities for conformable fractional integrals.RGMIA Research Report Collection19(2016), Article 93, 11 pp.

[13] M. Z. Sarikaya, H. Yaldiz and H. Budak, Steffensen’s integral inequality for conformable fractional integrals.RGMIA Research Report Collection19 (2016), Article 94, 9 pp.

[14] A. Tiryaki, Recent developments of Lyapunov-type inequalities.Adv. Dyn. Syst. Appl. 5(2010), no. 2, 231–248.

[15] X. Yang, Y. Kim and K. Lo, Lyapunov-type inequality for a class of odd-order differential equa- tions.J. Comput. Appl. Math.234(2010), no. 10, 2962–2968.

[16] X. Yang, Y. Kim and K. Lo, Lyapunov-type inequality for a class of linear differential systems.

Appl. Math. Comput.219(2012), no. 4, 1805–1812.

[17] C. Yildiz, M. E. Ozdemir and H. K. Onelan, Fractional integral inequalities for different functions.

New Trends Math. Sci.3(2015), no. 2, 110–117.

(Received 06.06.2017) Authors’ addresses:

S. H. Saker

Department of Mathematics, Faculty of Science, Mansoura University, Mansoura-Egypt.

E-mail: [email protected] D. O’Regan

School of Mathematics, Statistics and Applied Mathematics, National University of Ireland, Gal- way, Ireland.

M. R. Kenawy

Department of Mathematics, Faculty of Science, Fayoum University, Fayoum-Egypt.

E-mail: [email protected] R. P. Agarwal

Department of Mathematics, Texas A & M University, Kingsvilie, Texas, 78363, USA.

参照

関連したドキュメント

Functional differential equations, positive solutions, asymptotic behavior, regularly varying

Below we will describe a general method of solution of spatial axi- symmetric problems of the jet and filtration theories with partially unknown boundaries.. The liquid motion is

In the present work we suggest a general method of solution of spatial axisymmetric problems of steady liquid motion in a porous medium with partially unknown boundaries.. The

OPTIMAL PROBLEMS WITH DISCONTINUOUS INITIAL CONDITION.. systems governed by quasi-linear neutral differential equations with dis- continuous initial condition is considered.

Moreover, assuming only the Fundamental Factor- ization Theorem, we provide a complete proof of an important result from Shargorodsky [16], on the factorization of an

ary value problems are explicitly solved for the Bitsadze equation in the unit disc of the complex plane.. The results are obtained from iterations of related results for

Higher-order Sobolev space, linear extension operator, boundary trace operator, complex interpolation, weighted Sobolev space, Besov space, boundary value problem, Poisson problem

Problems of a contact between finite or infinite, isotropic or anisotropic plates and an elastic inclusion are reduced to the integral differential equa- tions with Prandtl