• Aucun résultat trouvé

Existence and regularity of weak solutions for ψ-Hilfer fractional boundary value problem

N/A
N/A
Protected

Academic year: 2021

Partager "Existence and regularity of weak solutions for ψ-Hilfer fractional boundary value problem"

Copied!
16
0
0

Texte intégral

(1)

HAL Id: hal-02562931

https://hal.archives-ouvertes.fr/hal-02562931

Preprint submitted on 5 May 2020

HAL

is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire

HAL, est

destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

Existence and regularity of weak solutions for ψ-Hilfer fractional boundary value problem

José Vanterler da Costa Sousa, M. Aurora P. Pulido, E. Capelas de Oliveira

To cite this version:

José Vanterler da Costa Sousa, M. Aurora P. Pulido, E. Capelas de Oliveira. Existence and regularity

of weak solutions for

ψ-Hilfer fractional boundary value problem. 2020. �hal-02562931�

(2)

(will be inserted by the editor)

Existence and regularity of weak solutions for ψ-Hilfer fractional boundary value problem

J. Vanterler da C. Sousa · M. Aurora P.

Pulido · E. Capelas de Oliveira

Received: date / Accepted: date

Abstract In this present paper we investigate the existence and regularity of weak solutions for ψ-Hilfer fractional boundary value problem in Cα,β;ψ2

and H (Hilbert space) spaces, using extension of the Lax-Milgram theorem.

In this sense, to finalize the paper, we discuss the integration by parts for ψ-Riemann-Liouville fractional integral andψ-Hilfer fractional derivative.

Keywords ψ-fractional differential boundary value problem, existence and regularity, weak solution, Lax-Milgram theorem.

Mathematics Subject Classification (2010) 26A33, 34A08, 35A15, 35B38.

1 Introduction and Preliminaries

In this paper we discuss the existence and regularity of weak solutions of ψ- Hilfer fractional boundary value problem

( Hα,β;ψT−

Hα,β;ψ0+ τ(t)

+τ(t) =λΦ(t, τ(t)), t∈(0, T) Iβ(β−1);ψ0+ τ(0) =Iβ(β−1);ψT τ(T) = 0,

(1)

where Hα,β;ψT− (·), Hα,β;ψ0+ (·) are the right-sided and left-sided ψ-Hilfer frac- tional derivatives of order 1/2 < α ≤ 1 and type 0 ≤ β ≤ 1, respectively, Iβ(β−1);ψT− (·) and Iβ(β−1);ψ0+ (·) are the right-sided and left-sided ψ-Riemann- Liouville fractional integrals of orderβ(β−1), respectively,λ is a parameter andΦ: [0, T]×R→Ris a continuous function [18, 19].

Department of Applied Mathematics, Imecc-State University of Campinas, 13083-859, Campinas, SP, Brazil

E-mail: vanterler@ime.unicamp.br,ra211681@ime.unicamp.br, capelas@unicamp.br

(3)

Fractional differential equations are tools of great importance in physics, biology, engineering and in various fields of science. In addition, we also high- light its importance in the discussion of qualitative properties of fractional differential equations [9, 20, 21, 22, 23, 24, 25]. Recently, the fractional calculus has been of paramount importance in the discussion of variational problems, which are formulated in the context of boundary value problems, problems involvingp-Laplacian, critical point theory, among other problems that stand out in this context [14, 27, 28]. In 2011 Jiao and Zhou [7], discussed the ex- istence of solutions to a fractional boundary value problem on a variational structure using critical point theory. Mahmudov and Unul [13], investigated the existence and uniqueness of solutions of fractional differential equation in the sense of Caputo with thep-Laplacian operator and integral conditions.

In 2016 Ledesma [10] investigated the existence of solution for the fractional p-Laplacian Dirichlet problem with mixed derivatives

DαT |Dα0+u(t)|p−2Dα0+u(t)

=f(t, u(t)), t∈[0, T] u(0) =u(T) = 0,

whereDαT(·) andDα0+(·), are Riemann-Liouville fractional derivatives the right and left of orderα with 1p < α <1,1< p < ∞ and f : [0, T]×R→R is a Carath´eodory function which satisfies some growth conditions.To obtain the result, a direct method in variational methods and mountain pass theorem was used.

Em 2017, Fattahi and Alimohammady [6] using non-smooth critical point theory and variational methods to study the existence solutions for a fractional boundary-value problem

DαT D0+α u(t)

∈λ∂F(u(t)) +µ∂F(u(t)), a.e. t∈[0, T] u(0) =u(T) = 0,

whereDTα(·) andDα0+(·), are Riemann-Liouville fractional derivatives the right and left of orderαcom 0< α≤1, and whereλ >0 andµ≥0 are two param- eters:F, G:R→Rare locally Lipschitz functions, whereF(ω) =

Z ω 0

f(s)ds, G(ω) =

Z ω 0

g(s)ds, ω ∈R and f, g : R→ R are locally essentially bounded functions∂F(u(t)) denotes the generalized Clarke gradient of functionF(u(t)) at u∈R. Fractional differential problems were studied by many authors, see for example [2, 4, 5, 11, 12, 15, 16, 26].

Afrouzi and Hadjian [1], established results of existence and uniqueness of classic infinite solutions for Caputo fractional differential equations with limit value movement, using the critical point theory. In this sense, presented two examples, in order to elucidate the investigated results. However, there are few studies on the existence and uniqueness of fractional boundary conditions, some references for a brief reading [8, 27] and references therein.

Motivated by the work discussed above, and in order to contribute to the growth of the area, the main contributions to our study are highlighted as follows:

(4)

1. We discussed conditions for obtaining the existence and regularity of weak solutions for Eq.(1);

2. The proposed problem is more generalized, and some it in the literature are the special cases of it. Moreover, according to particular cases, our analysis can also be applied to the addressed problems by selecting the appropriate function ofψ;

3. Theψ-fractional spaces are directly linked with their respective fractional operator. Therefore, each fractional problem discussed, as well as its re- spective particular cases, are well defined;

Let (X,k·kX) be a real Banach space with dual spaceX. Denote byBr(x0) the ballBr(x0) ={x∈X,kx−x0k ≤r}and consider the set

∆= max

Φ(t, s),(t, s)∈[0, T]×

−(ψ(T)−ψ(0))α−1/2

Γ(α)(2α−1)1/2 ,(ψ(T)−ψ(0))α−1/2 Γ(α)(2α−1)1/2

.

Let 0< α≤1 and 0≤β ≤1. The left-sidedψ-fractional derivative space Cα,β;ψ2 :=Cα,β;ψ2 ([0, T],R) is defined by the closure ofC0([0, T],R) [8, 17],

Cα,β;ψ2 =

(y∈L2([0, T],R) ;Hα,β;ψ0+ y∈L2([0, T],R), Iβ(β−1)0+ u(0) =Iβ(β−1)T u(T) = 0

)

=C0([0, T],R).(2)

On the other hand, let 0 < α≤1, 0≤β ≤1 and 1 < p <∞. The Left- sidedψ-fractional derivative spaceC2α,β;ψp :=C2α,β;ψp ([0, T],R) is defined by the closure ofC0([0, T],R),is given by

C2α,β;ψp =









y∈Lp([0, T],R) ; Hα,β;ψ0+ y∈Lp([0, T],R) and Hα,β;ψT

Hα,β;ψ0+ y

∈Lp([0, T],R), Iβ(β−1)0+ y(0) =Iβ(β−1)0+ y(T) = 0, Hα,β;ψT

Hα,β;ψ0+ y(0)

= Hα,β;ψT

Hα,β;ψ0+ y(T)

= 0









with the following norm kykC2α,β;ψp =

kykpLp+ Hα,β;ψ0+ y

p Lp+

Hα,β;ψT−

Hα,β;ψ0+ y

p Lp

1/p

∀y∈C2α,β;ψp andHα,β;ψ0+ (·) is theψ-Hilfer fractional derivative with 0< α≤1 and 0≤β ≤1.

Let (0, T) be a finite or infinite interval of the real lineRandα >0. Also letψ(t) be an increasing and positive monotone function on (0, T], having a continuous derivativeψ0(t) on (0, T). The left-sided and right-sided fractional integrals of a functionfwith respect to another functionψon [0, T] are defined by [18, 19]

Iα;ψ0+f(x) = 1 Γ(α)

Z x a

ψ0(t) (ψ(x)−ψ(t))α−1f(t)dt (3)

(5)

and

Iα;ψTf(x) = 1 Γ(α)

Z T x

ψ0(t) (ψ(t)−ψ(x))α−1f(t)dt. (4) Let n−1 < α < n with n ∈ N, I = [0, T] is the interval and f, ψ ∈ Cn([0, T],R) two functions such that ψ is increasing and ψ0(t) 6= 0, for all t ∈ I. The ψ-Hilfer fractional derivative left-sided and right-sidedHα,β;ψ0+ (·) andHα,β;ψT (·) of function of orderαand type 0≤β ≤1, are defined by [18, 19]

Hα,β;ψ0+ f(x) =Iβ(n−α);ψ0+

1 ψ0(x)

d dx

n

I(1−β)(n−α);ψ

0+ f(x) (5)

and

Hα,β;ψT f(x) =Iβ(n−α);ψT

− 1 ψ0(x)

d dx

n

I(1−β)(n−α);ψ

T f(x). (6)

Theψ−Hilfer fractional derivatives defined as above can be written in the following form

Hα,β;ψ0+ f(x) =Iγ−α;ψ0+ Dγ;ψ0+f(x) (7) and

Hα,β;ψT− f(x) =Iγ−α;ψT− Dγ;ψT−f(x) (8) with γ = α+β(n−α) and Iγ−α;ψ0+ (·), Iγ−α;ψT (·), are given by Eq.(3) and Eq.(4), respectively. The ψ-Riemann-Liouville fractional derivatives DTγ;ψ(·) andD0+γ;ψ(·), can be obtained in the following references [18, 19].

Lemma 1 [18, 19] Let α >0 andδ > 0. Then, we have the following semi- group property given by

Iα;ψ0+ Iδ;ψ0+f(x) =Iα+δ;ψ0+ f(x) (9) and

Iα;ψT− Iδ;ψT−f(x) =Iα+δ;ψT− f(x). (10) Theorem 1 [18, 19]If f ∈Cn([0, T],R),n−1< α < nand0≤β ≤1,then

Iα;ψ0+ Hα,β;ψ0+ f(x) =f(x)−

n

X

k=1

(ψ(x)−ψ(a))γ−k

Γ(γ−k+ 1) fψ[n−k]I(1−β)(n−α);ψ

a+ f(0)

(11) and

Iα;ψT Hα,β;ψT f(x) =f(x)−

n

X

k=1

(ψ(b)−ψ(x))γ−k

Γ(γ−k+ 1) fψ[n−k]I(1−β)(n−α);ψ

T− f(T),

(12) wherefψ[n−k](x) :=

1 ψ0(x)

d dx

n−k

f(x).

(6)

Definition 1 [8, 17] Let 0< α≤1 and 0≤β≤1. Theψ-fractional derivative spaceCα,β;ψ2 is defined by the closure ofC0([0, T]) with respect to the norm,

kτkCα,β;ψ2

=

kτk2L2(0,T)+ Hα,β;ψ0+

2 L2(0,T)

1/2 .

Theorem 2 [8, 17] Let0< α≤1 and0≤β ≤1. Theψ-fractional derivative spaceCα,β;ψ2 is reflexive and separable Banach space.

Theorem 3 [8, 17] Assume that α > 1/2, 0 ≤ β ≤ 1 and the sequence τk converges weakly to τ inCα,β;ψ2 , thenτk→τ inC([0, T]).

Remark 1 We have

kτk≤(ψ(T)−ψ(0))α−1/2 Γ(α)(2α−1)1/2 kτk

Cα,β;ψ2 , ∀x∈Cα,β;ψ2 .

Definition 2 [8, 17] A functionτ ∈Cα,β;ψ2 is a weak solution of Eq.(1), pro- vided that for anyξ∈Cα,β;ψ2 , yields

Z T 0

Hα,β;ψ0+ τ(t)Hα,β;ψ0+ ξ(t) +τ(t)ξ(t) dt=λ

Z T 0

Φ(t, τ(t), ξ(t))dt. (13) Theorem 4 [14] Suppose thatHis a Hilbert space,B(τ, ξ)is continuous co- ercive bilinear form onHandΞ:H → H satisfying the following conditions:

(C1)There exists a constantN >0such that||Ξ(u)||H≤N, ∀u∈B1(0), whereB1(0) ={u∈ H:||u||H≤1};

(C2) If {uk} is a sequence in Hsuch that uk * uweakly in H, then the sequence{Ξ(uk)}has a subsequence{Ξ(uk)}such thatΞ(ukn)*Ξ(u)weakly inH.

The rest of the article is divided as follows. In section 2, we discussed the existence and regularity of weak solutions for ψ-Hilfer fractional boundary value problem in Cα,β;ψp ([0, T],R) using the Lax-Milgram theorem and some results that will be discussed during the section. In addition, we discuss the integration by parts for ψ-Riemann-Liouville fractional integral andψ-Hilfer fractional derivative and some particular cases are investigated. Concluding remarks closed the paper.

2 Existence and regularity of weak solutions

In this section, we investigate the existence and regularity of weak solutions of ψ-Hilfer fractional boundary problem, that is, Theorem 5 and Theorem 10. In addition, during the section, some particular cases are discussed, in order to elucidate the broad class of particular cases that the operator as well as their respective space holds.

The first main result of this section, is given by the following theorem:

(7)

Theorem 5 Suppose that 1/2 < α ≤ 1, 0 ≤ β ≤ 1 and Φ ∈ C([0, T]× R,R), then for any|λ| ≤ 1

∆(ψ(T)−ψ(0))1/2, theEq.(1)has at least one weak solution.

Proof Consider B(τ, ξ) =

Z T 0

Hα,β;ψ0+ τ(t)Hα,β;ψ0+ ξ(t) +τ(t)ξ(t) dt.

and by means of H¨older’s inequality, yields

|B(τ, ξ)| ≤ Z T

0

Hα,β;ψ0+ τ(t)

Hα,β;ψ0+ ξ(t) dt+

Z T 0

|τ(t)| |ξ(t)|dt

Hα,β;ψ0+ τ 2

Hα,β;ψ0+ ξ

2+kτk2kξk2≤2kτk2

Cα,β;ψ2

kξk2

Cα,β;ψ2

, therefore,|B(τ, ξ)| ≥ kξk2

Cα,β;ψ2

.

It follows thatBis a continuous coercive bilinear form onCα,β;ψ2 . We define Ξ:Cα,β;ψ2

Cα,β;ψ2

and

hΞ(τ), τi= Z T

0

Φ(t, τ(t))ξ(t)dt.

Assumeτ∈Cα,β;ψ2 withkτk

Cα,β;ψ2 ≤1. Then, by means of Remark 1, yields kτk≤ (ψ(T)−ψ(0))α−1/2

Γ(α)(2α−1)1/2 kτk

Cα,β;ψ2 ≤ (ψ(T)−ψ(0))α−1/2

Γ(α)(2α−1)1/2 , (14)

and we obtain kτk ≤ (ψ(T)−ψ(0))α−1/2

Γ(α)(2α−1)1/2 , for any t ∈ [0, T]. So, we can conclude that|hΞ(τ), ξi| ≤∆, for any t∈[0, T].

For anyv∈Cα,β;ψ2 withkξk

Cα,β;ψ2

= 1, by the H¨older’s inequality, yields

|hΞ(τ), ξi|=

Z T 0

Φ(t, τ(t))ξ(t)dt

≤ Z T

0

Φ(t, τ(t))dt

!1/2

kξk2L2(0,T)≤∆T1/2.

TakingN =∆T1/2, the condition (C1) of Theorem 4 holds. Suppose{τk} is a sequence in Cα,β;ψ2 such that τk → τ weakly in Cα,β;ψ2 . Then, by the Theorem 3, we get that for anyt∈[0, T],τk(t)→τ(t), ∀t∈[0, T].

Using this and the continuity of the functionΦ, yields

Φ(t, τk(t))→Φ(t, τ(t)), as k→ ∞, ∀t∈[0, T]. (15)

(8)

The sequence, {τk} is a bounded subset of Cα,β;ψ2 . In other words, there exists a constantM>0 such thatkτkk

Cα,β;ψ2 ≤MwithM∈N. Thus, of the Eq.(14), yields

kk≤M(ψ(T)−ψ(0))α−1/2

Γ(α)(2α−1)1/2 , k∈N. Therefore, there exists ∆0>0 (constant), such that

|Φ(t, τk(t))|< ∆0, ∀t∈[0, T], k= 1,2, . . . (16) In this sense, of Eq.(15), Eq.(16) and of Lebesgue dominated theorem, we conclude that

Z T 0

|Φ(t, τk(t))−Φ(t, τ(t))|2dt→0.

For anyξ∈Cα,β;ψ2 withkvk

Cα,β;ψ2 = 1, yields

|hΞ(τk)−Ξ(u), vi|=

Z T 0

(Φ(t, τk(t))−Φ(t, τ(t)))v(t)dt

≤ kΦ(t, τk(t))−Φ(t, τ(t))k2→0,

which yields that Ξ satisfies (C2). Then by Theorem 4, we concluded the proof.

Theorem 6 Taking ψ(t) = t and the limit β → 1 in Eq.(1). Suppose that 1/2< α≤1 andΦ∈C([0, T]×R,R), then for any|λ| ≤ 1

∆(t)1/2, the Eq.(1) has at least one weak solution in the Caputo fractional derivative sense.

Proof The proof is straight from Theorem 5, taking ψ(t) = t and the limit β→1.

Theorem 7 Takingψ(t) =tρ (ρ >0) and the limitβ→0inEq.(1). Suppose that 1/2< α≤1 andΦ∈C([0, T]×R,R), then for any |λ| ≤ 1

∆(tρ)1/2, the Eq.(1)has at least one weak solution in the Katugampola fractional derivative sense.

Proof The proof is straight from Theorem 5, taking ψ(t) = tρ and the limit β→0.

Theorem 8 Takingψ(t) =tρ (ρ >0) and the limitβ→1inEq.(1). Suppose that 1/2< α≤1 andΦ∈C([0, T]×R,R), then for any |λ| ≤ 1

∆(tρ)1/2, the Eq.(1) has at least one weak solution in the Caputo-Katugampola fractional derivative sense.

Proof The proof is straight from Theorem 5, taking ψ(t) = tρ and the limit β→1.

(9)

Note that we can discuss numerous other particular cases, however we cannot chooseψ(t) = lnt, since it is not set tot∈[0, T].

We are going to present the definition of fractional derivative in the weak sense, and some results are discussed.

Definition 3 Letτ, w, ξ∈L2(0, T) with Z T

0

τ(t)Hα,β;ψT ϕ(t)dt= Z T

0

ξ(t)ϕ(t)dt, ∀ϕ∈C0(0, T), and

Z T 0

τ(t)Hα,β;ψ0+ ϕ(t)dt= Z T

0

w(t)ϕ(t)dt, ∀ϕ∈C0(0, T).

Note that,H

α,β;ψ

T− τ(t) =ξ(t) andH

α,β;ψ

0+ τ(t) =w(t).

The functions ξ and w given above will be called the weak left and the weak right fractional derivative of orderα∈(0,1] and typeβ(0≤β≤1) ofτ respectively. Here, we denote them byH

α,β;ψ

0+ τ(t) andH

α,β;ψ

T− τ(t) respectively.

From the choice of the functionψ(·), we can obtain a wide class of fractional derivatives in the weak sense, as particular cases.

In view of Definition 1, τ ∈Cα,β;ψ2 means thatτ is the limit of a Cauchy sequence {τn} ⊂C0(0, T), i.e., τn →τ in L2(0, T) and ∃w ∈L2(0, T) such thatHα,β;ψ0+ τn→winL2(0, T).

Then, for any ϕ∈C0(0, T), yields Z T

0

w(t)ϕ(t)dt= lim

n→∞

Z T 0

Hα,β;ψ0+ τn(t)ϕ(t)dt= lim

n→∞

Z T 0

ψ0(t)τn(t)Hα,β;ψ0+

ϕ(t) ψ0(t)

dt

= Z T

0

ψ0(t)τ(t)Hα,β;ψ0+

ϕ(t) ψ0(t)

dt.

Therefore,w=Hα,β;ψ0+ however it is not clear whetherHα,β;ψ0+ τ(t) exists in the usual sense, for anyt∈[0, T].

Next, we will prove the integration by parts for the ψ-Riemann-Liouville fractional integral and theψ-Hilfer fractional derivative.

The relation Z b

a

Iα;ψ0+τ(t)

ξ(t) dt = Z b

a

τ(t)ψ0(t)Iα;ψT ξ(t)

ψ0(t)

dt (17)

is valid.

One can prove Eq.(17) directly by interchanging the order of integration by the Dirichlet formula in the particular case Fubini theorem, i.e.,

Z b a

Iα;ψ0+τ(t)

ξ(t) dt = Z b

a

1 Γ(α)

Z t a

ψ0(s) (ψ(t)−ψ(s))α−1τ(s) dsξ(t) dt

= Z b

a

1 Γ(α)

Z b t

ψ0(s) (ψ(t)−ψ(s))α−1ξ(t) dtτ(s) ds

= Z b

a

τ(t)ψ0(t)Iα;ψT ξ(t)

ψ0(t)

dt.

(10)

Theorem 9 Let ψ(·) be an increasing and positive monotone function on [0, T], having a continuous derivative ψ0(·)6= 0 on (0, T). If 0 < α≤ 1 and 0≤β ≤1, then

Z b a

Hα,β;ψ0+ τ(t)

ξ(t) dt = Z b

a

τ(t)ψ0(t) Hα,β;ψT

ξ(t) ψ0(t)

dt (18) for any τ ∈AC1 and ξ∈C1 satisfying the boundary conditions τ(0) = 0 = τ(T).

Proof In fact, using the Eq.(8), Eq.(10) and Theorem 1 (Eq.(11)), yields Z b

a

τ(t)ψ0(t) Hα,β;ψT

ξ(t) ψ0(t)

dt

= Z b

a

τ(t)ψ0(t)I1−α;ψT D1;ψT ξ(t)

ψ0(t)

dt

= Z b

a

ψ0(t)

"

Iα;ψ0+ Hα,β;ψ0+ τ(t) +(ψ(t)−ψ(a))γ−1 Γ(γ) dj

#

I1−α;ψT D1;ψT−

ξ(t) ψ0(t)

dt

wheredj = 1

ψ0(t) d dt

I(1−β)(1−α);ψ

0+ τ(0)

= Z b

a

ψ0(t)Iα;ψ0+ Hα,β;ψ0+ τ(t)I1−α;ψT− DT1;ψ ξ(t)

ψ0(t)

dt

+ dj Γ(γ)

Z b a

ψ0(t) (ψ(t)−ψ(a))γ−1I1−γ;ψT DT1;ψ ξ(t)

ψ0(t)

dt

= Z b

a

Iα;ψ0+ Hα,β;ψ0+ τ(t)I−α;ψT ξ(t)

ψ0(t)

dt

= Z b

a

Hα,β;ψ0+ τ(t) ξ(t) dt.

Remark 2 From Theorem 9, forξ, τ ∈Cα,β;ψ2 , we have Z T

0

Hα,β;ψ0+ τ(t)

ξ(t)dt= Z T

0

ψ0(t)τ(t)Hα,β;ψT

ξ(t) ψ0(t)

dt,

sinceC0(0, T)⊂Cα,β;ψ2 , then Hα,β;ψ0+ τ(t) =H

α,β;ψ

0+ τ(t) a.e. ont∈[0, T].

Lemma 2 Letτ ∈Cα,β;ψ2 , thenH

α,β;ψ

0+ τ(t)is almost every where equal to the weak derivative ofI1−α;ψ0+ τ(t) in theH1(0, T)sense, i.e.,

H

α,β;ψ

0+ τ(t) =Iβ(1−α);ψ0+ DψI1−γ;ψ0+ τ(t),

whereDψ= 1 ψ0(t)

d

dt, a.e on [0, T]andγ=α+β(1−α).

(11)

Proof For anyϕ∈C0(0, T), yields Z T

0

v(t)ϕ(t)dt= Z T

0

ψ0(t)u(t)H

α,β;ψ T

ϕ(t) ψ0(t)

dt

= Z T

0

ψ0(t)u(t)Iγ−α;ψT RLDγ;ψT ϕ(t)

ψ0(t)

dt

= Z T

0

ψ0(t)u(t)Iγ−α;ψT CDγ;ψT ϕ(t)

ψ0(t)

dt

=− Z T

0

ψ0(t)u(t)Iγ−α;ψT I1−γ;ψT

Dψ

ϕ(t) ψ0(t)

dt

=− Z T

0

ψ0(t)u(t)I1−α;ψT 1

ψ0(t) d dt

ϕ(t) ψ0(t)

dt

=− Z T

0

ψ0(t)u(t)I1−α;ψ

Dψ(t) ψ0(t)

dt

=− Z T

0

Iβ(1−α);ψ0+ I(1−β)(1−α);ψ

0+ u(t)Dψ(t)dt

=− Z T

0

ψ0(t)I(1−β)(1−α);ψ

0+ u(t)Iβ(1−α);ψT

Dψ(t) ψ0(t)

dt

=− Z T

0

ψ0(t)I(1−β)(1−α);ψ

0+ u(t)Iβ(1−α);ψT Dψ

ϕ(t) ψ0(t)

dt

= Z T

0

ψ0(t)I(1−β)(1−α);ψ

0+ u(t)CD1−β(1−α);ψT ϕ(t)

ψ0(t)

dt

= Z T

0

CD1−β(1−α);ψ0+ I(1−β)(1−α);ψ

0+ u(t)ϕ(t)dt

= Z T

0

Iβ(1−α);ψ0+ Dψ(t)I(1−β)(1−α);ψ

0+ u(t)ϕ(t)dt.

Therefore H

α,β;ψ

0+ τ(t) =Iβ(1−α);ψ0+ Dψ(t)I(1−β)(1−α);ψ

0+ τ(t),

withDψ(t) = 1 ψ0(t)

d dt.

Lemma 3 Let 0< α≤1 and0≤β≤1.

1. If1/2< α≤1 andτ∈L2(0, T), thenIα+β(β−1);ψ0+ τ(0) = 0;

2. Ifτ∈C([0, T]), thenIα,β;ψ0+ τ∈C([0, T]);

3. Ifτ∈C1([0, T])andIβ(β−1);ψ0+ τ(0) = 0, thenIα;ψ0+τ∈C1([0, T]).

Proof Let ˜α=α+β(β−1), then we have

Iα;ψ0+˜ τ(t)

≤(ψ(t)−ψ(0))α−1/2˜

Γ( ˜α)(2 ˜α−1)1/2 kτkL2[0,T],

(12)

which completes the proof of 1.

Lett0∈[0, T] and{tn}be a sequence in [0, T] such thattn→t0. Consider M= max0≤t≤T|τ(t)|.

As τ(tn−s)→τ(t0−s) almost every on [0, T],

Ψα0(t−s)|τ(tn−s)−τ(t0−s)| ≤2MΨα0(t−s), and by means of Lebesgue dominated theorem, we conclude

Z tn 0

Ψα0(t−s)τ(tn−s)ds− Z t0

0

Ψα0(t−s)τ(t0−s)ds

≤ Z ξ

0

Ψα0(t−s)|τ(tn−s)−τ(t0−s)|ds+ Z tn

t0

Ψα0(t−s)|τ(t0−s)|ds→0.

whereξ= max{t0, t1, t2, . . .}andΨα0(t−s) :=ψ0(t−s)(ψ(t)−ψ(t−s))α−1. This concludes the proof of 2.

Now, suppose that K = max0≤t≤T0(t)| then

τn(tn−s)−τ(t0−s) tn−t0

≤ K. Therefore, using the Lebesgue dominated theorem, yields

Γ(α) Iα;ψtn τ(tn)−Iα;ψt0 τ(t0) tn−t0

!

= Z tn

0

Ψα0(t−s)(τ(tn−s)−τ(t0−s)) tn−t0

ds

+ Z tn

t0

Ψα0(t−s)(τ(tn−s)−τ(t0−s)) tn−t0

ds

+ Z tn

t0

Ψα0(t−s)(τ(t0−s)−τ(0)) t0−s

t0−s tn−t0

ds

→ Z t0

0

Ψα0(t−s)τ0(t0−s)ds.

Thus,h

Iα;ψ0+τ(t)i0

(t0) =Iα;ψ0+ τ0(t0). From this and part 2, we conclude 3.

Lemma 4 Suppose that for someτ ∈L2(0, T),H

α,β;ψ

0+ τ(t) exists and is a.e.

equal to a function inC([0, T]). Then

1. τ is almost every equal to a functionτ˜∈C([0, T]);

2. Hα,β;ψ0+ τ(t)exist for anyt∈[0, T]andHα,β;ψ0+ τ(t)∈C([0, T]).

Proof First of all, note thatH

α,β;ψ

0+ Φ(x) =RLD1−β(1−α);ψ0+ I1−γ;ψ0+ Φ(x), yields H

α,β;ψ

0+ Φ(x) =Dψ

I1−α;ψ0+ Φ(x) ,

and by the Lemma 2, we obtainDψ

I1−α;ψ0+ f(x)

is almost everywhere equal.

(13)

Therefore,I1−α;ψ0+ τ(t) is almost everywhere equal to a function inC1([0, T]).

Thus by Lemma 3, yields Z t

0

τ(s)ds=Iα;ψ0+I1−α;ψ0+

τ(t) ψ0(t)

∈C1([0, T]).

Take ˜τ=D Z t

0

τ(s)ds

[3], this complete the proof of 1.

On the other hand, by means of Lemma 3 implies thatI1−α;ψ0+ τ˜∈C([0, T]).

Since I1−α;ψ0+ τ(t) = I1−α;ψ0+ τ(t),˜ ∀t ∈ [0, T], then I1−α;ψ0+ τ(t) ∈ C([0, T]). By using it and the fact thatI1−α;ψ0+ τ(t) is almost everywhere equal to a function inC1([0, T]), we can concludeI1−α;ψ0+ τ(t)∈C1([0, T]).

So, yields

Hα,β;ψ0+ τ(t) =RLDα;ψ0+τ(t) =Dψ

I1−α;ψ0+ τ(t) ,

in the weak sense.

Below we present two particular cases from Lemma 4, in the sense of Riemann-Liouville and Katugampola fractional derivatives.

Lemma 5 (Riemann-Liouville). Suppose that for someτ∈L2(0, T),H

α,0;t 0+ τ(t) exists and is a.e. equal to a function inC([0, T]). Then

1. τ is almost every equal to a functionτ˜∈C([0, T]);

2. Hα,0;t0+ τ(t) exist for anyt∈[0, T]andHα,0;t0+ τ(t)∈C([0, T]).

Lemma 6 (Katugampola). Suppose that for some τ ∈ L2(0, T), H

α,0;tρ 0+ τ(t) (ρ >0) exists and is a.e. equal to a function in C([0, T]). Then

1. τ is almost every equal to a functionτ˜∈C([0, T]);

2. Hα,0;t

ρ

0+ τ(t)exist for anyt∈[0, T]andHα,0;t

ρ

0+ τ(t)∈C([0, T]).

The following Lemma 7 is similar to Lemma 3 and Lemma 4.

Lemma 7 Suppose thatτ ∈L2([0, T]),Hα,β;ψT− τ(t)exists and is almost every- where equal to a function inC([0, T]). Then

1. w is almost everywhere equal to a function ˜τ∈C([0, T]).

2. Hα,β;ψT τ(t)exists for anyt∈[0, T]andHα,β;ψT τ(t)∈C([0, T]).

Theorem 10 Consider the conditions ofTheorem 5. Then, every weak solu- tion of Eq.(1)is a classical solution.

Proof Suppose that τ is the weak solution of Eq.(1) and consider g(t) = λΦ(t, τ(t))−τ(t). Using the definition of weak solution, yields

Z T 0

Hα,β;ψ0+ τ(t)Hα,β;ψ0+ ξ(t)dt= Z T

0

g(t)ξ(t)dt,

(14)

for allξ∈Hα,β;ψp .

Using the Definition 3, we have g(t) = H

α,β;ψ

T− Hα,β;ψ0+ τ(t). Using the Re- mark 2 and Theorem 3, we concludeg(t) =H

α,β;ψ

T− H

α,β;ψ

0+ τ(t)∈C([0, T]).

Then Lemma 7 implies that Hα,β;ψT H

α,β;ψ

0+ τ(t) exists for any t ∈ [0, T], Hα,β;ψT− H

α,β;ψ

0+ τ(t) ∈ C([0, T]) and H

α,β;ψ

0+ τ(t) is almost everywhere equal to an element of C([0, T]). So by Lemma 4 (part 2), there exist Hα,β;ψ0+ τ(t) for any t ∈ [0, T]. By Remark 2, we have H

α,β;ψ

0+ τ(t) = Hα,β;ψ0+ τ(t) a.e on [0, T].

Then, we concluded thatHα,β;ψT Hα,β;ψ0+ τ(t) exist for anyt∈[0, T].

Since g and Hα,β;ψT− Hα,β;ψ0+ τ(t) are almost everywhere equal and they con- tinuous, we have

Hα,β;ψT Hα,β;ψ0+ τ(t) =g(t), ∀t∈[0, T], which concludes the proof.

The results above are also valid for their respective particular cases, as discussed for Theorem 5, from the choice of the functionψ(t).

3 Concluding remarks

We have obtained some results of existence and regularity of weak solutions for ψ-Hilfer fractional boundary value problem inCα,β;ψ2 andH, using the exten- sion of the Lax-Milgram Theorem and some results discussed in section 2. The results discussed here are general and global that hold a wide class of particu- lar cases, both the problem investigated and its respectiveψ-fractional space.

Other issues in this regard have been discussed in other variational problems, for example, what conditions should be imposed in order to obtain multiplic- ity of weak solutions in the spaceC2α,β;ψ2 ? Finally, a variational structure has been discussed, in order to ensure that the variational problem discussed is well established.

Acknowledgment

JVDCS acknowledges PNPD-CAPES (no88882.32909093/2019-01) for finan- cial support scholarship of the Postgraduate Program in Applied Mathematics of IMECC-Unicamp.

References

1. Afrouzi, G. A., Hadjian, A. ”A variational approach for boundary value problems for impulsive fractional differential equations.” Frac. Cal. Appl. Anal. 21.6, 1565-1584 (2018).

2. Ali, K. B., Ghanmi, A., Kefi, K.. ”Existence of solutions for fractional differential equa- tions with Dirichlet boundary conditions.” Elect. J. Diff. Equ. 2016.116, 1-11 (2016).

(15)

3. Brezis, H. Functional analysis, Sobolev spaces and partial differential equations. Springer Science & Business Media, (2010).

4. Chai, G. ”Infinitely many solutions for nonlinear fractional boundary value problems via variational methods.” Adv. Diff. Equ. 2016.1, 213 (2016).

5. Cruz, G. A. M., Ledesma, C. T.. ”Multiplicity of solutions for fractional Hamiltonian systems with Liouville-Weyl fractional derivatives.” Frac. Cal. Appl. Anal. 18.4, 875-890 (2015).

6. Fattahi, F., Alimohammady, M.. ”Existence of infinitely many solutions for a fractional differential inclusion with non-smooth potential.” Electron. J. Differ. Equ. 66, 1–13 (2017).

7. Jiao, F., Zhou, Y. ”Existence of solutions for a class of fractional boundary value problems via critical point theory.” Comput. Math. Appl. 62.3, 1181-1199 (2011).

8. Jiao, F., Zhou, Y. ”Existence results for fractional boundary value problem via critical point theory.” Inter. J. Bifur. Chaos 22.04, 1250086 (2012).

9. Kilbas, A. A., Srivastava, H. M., Trujillo,J. J.. Theory and applications of fractional differential equations. Elsevier Science Limited, Amsterdam, (2006).

10. Ledesma, C. T. ”Boundary value problem with fractional p-Laplacian operator.” Adv.

Nonlinear Anal. 5.2, 133-146 (2016).

11. Ledesma, C. Torres. ”Existence and concentration of solution for a class of fractional Hamiltonian systems with subquadratic potential.” Proc. Math. Sci. 128.4, 50 (2018).

12. Li, D., Chen, F., An, Y.. ”The existence of solutions for an impulsive fractional coupled system of (p, q)-Laplacian type without the Ambrosetti-Rabinowitz condition.” Math.

Meth. Appl. Sci. 42.5, 1449-1464 (2019).

13. Mahmudov, N. I., Unul, S.. ”Existence of solutions of fractional boundary value problems with p-Laplacian operator.” Bound. Value Prob. 2015.1 (2015): 99.

14. Nyamoradi, N., Hamidi,M. R. ”An extension of the Lax-Milgram theorem and its ap- plication to fractional differential equations.” Elec. J. Diff. Equ. 2015.95 ,1-9 (2015).

15. Nyamoradi, N., Tersian, S.. ”Existence of solutions for nonlinear fractional order p- Laplacian differential equations via critical point theory.” Frac. Cal. Appl. Anal. 22.4, 945-967 (2019).

16. Nyamoradi, N., Zhou, Y., Ahmad, B., Alsaedi, A. ”Variational approach to homoclinic solutions for fractional Hamiltonian systems.” J. Opt. Theory Appl. 174.1, 223-237 (2017).

17. Sousa, J. Vanterler da C., Tavares, L. S., Ledesma, C. T. Variational structure :ψ-Hilfer fractional operator, (Submitted), (2020).

18. Sousa, J. Vanterler da C., Oliveira, E. Capelas de. ”On theψ-Hilfer fractional deriva- tive.” Commun. Nonlinear Sci. Numer. Simul. 60, 72-91 (2018).

19. Sousa, J. Vanterler da C., Oliveira, E. Capelas de. ”Leibniz type rule:ψ-Hilfer fractional operator.” Commun. Nonlinear Sci. Numer. Simul. 77, 305-311 (2019).

20. Sousa, J. Vanterler da C., Oliveira, E. Capelas de. ”Ulam–Hyers stability of a nonlinear fractional Volterra integro-differential equation.” Appl. Math. Lett. 81, 50-56 (2018).

21. Sousa, J. Vanterler da C., Kucche, Kishor D., Oliveira, E. Capelas de. ”Stability of ψ-Hilfer impulsive fractional differential equations.” Appl. Math. Lett. 88, 73-80 (2019).

22. Sousa, J. Vanterler da C., Oliveira, D. S., Oliveira, E. Capelas de. ”On the existence and stability for noninstantaneous impulsive fractional integrodifferential equation.” Math.

Meth. Appl. Sci. 42.4, 1249-1261 (2019).

23. Sousa, J. Vanterler da C., Santos, Magun N. N., Magna, L. A., Oliveira, E. Capelas de. ”Validation of a fractional model for erythrocyte sedimentation rate.” Comput. Appl.

Math. 37.5, 6903-6919 (2018).

24. Sousa, J. Vanterler da C., Rodrigues, Fabio G., Oliveira, E. Capelas de. ”Stability of the fractional Volterra integro-differential equation by means ofψ-Hilfer operator.” Math.

Meth. Appl. Sci. 42.9, 3033-3043 (2019).

25. Sousa, J. Vanterler da C., and E. Capelas de Oliveira. ”On the stability of a hyperbolic fractional partial differential equation.” Diff. Equ. Dyn. Sys. (2019).

https://doi.org/10.1007/s12591-019-00499-3.

26. Ziheng, Z., Yuan, R.. ”Variational approach to solutions for a class of fractional Hamil- tonian systems.” Math. Meth. Appl. Sci. 37.13, 1873-1883 (2014).

27. Zhang, W., Liu, W. ”Variational approach to fractional Dirichlet problem with instantaneous and non-instantaneous impulses.” Appl. Math. Lett. (2020) https://doi.org/10.1016/j.aml.2019.07.024.

(16)

28. Zhou, J., Deng, Y., Wang, Y. ”Variational approach top-Laplacian fractional differential equations with instantaneous and non-instantaneous impulses.” Appl. Math. Lett. 104 (2020) https://doi.org/10.1016/j.aml.2020.106251.

Références

Documents relatifs

Existence and Non-existence of Global Solutions to Initial Boundary Value Problems for Nonlinear Evolution Equations with the Strong Dissipation.. Akisato Kubo a , Jean-Pierre Loh´

In 2008 Banas and O’Regan [6] investigated the existence and local attractivity of solutions of a quadratic Volterra integral equation of fractional order in Ba- nach spaces and in

Tableau 17 : Fréquence des étiologies selon le terme de survenue de l’hémorragie Tableau 18 : fréquence des étiologies selon l’abondance de l’hémorragie Tableau 19 :

Dans le cadre d’un marché du travail frictionnel comme celui analysé, une taxe progressive peut avoir des effets positifs sur le fonctionnement du marché, en particulier elle peut

Besides, thanks to the high quality of the results gained from this experimental loop together with the fact that the VERDON laboratory is unique in the world for studies

Simulated I d (V ds ) characteristics in the off-state, showing the off- state current and the breakdown voltage of the normally-on HEMT and the normally-off HEMT with

Hence, our results show that heterogeneous data such as preterm newborns, infants, adults, in vivo or post mortem data, and even synthetic data like the ICBM brain template, can

Changes in the geographical distribution of economic activities, which are in the nature of the growth process, may go hand-to-hand with internal distortions in the structure