It is known that there is no non-constant periodic solutions on a closed bounded interval for differential equations with fractional order. Therefore, many researchers investigate the existence of asymptotically periodic solution for differential equations with fractional order. In this paper, we demonstrate the existence and uniqueness of the S-asymptotically ω-periodic mild solution to non-instantaneous impulsive semilinear differential equations of order 1<α<2, and its linear part is an infinitesimal generator of a strongly continuous cosine family of bounded linear operators. In addition, we consider the case of differential inclusion. Examples are given to illustrate the applicability of our results.
Citation: Zainab Alsheekhhussain, Ahmed Gamal Ibrahim, Rabie A. Ramadan. Existence of S-asymptotically ω-periodic solutions for non-instantaneous impulsive semilinear differential equations and inclusions of fractional order 1<α<2[J]. AIMS Mathematics, 2023, 8(1): 76-101. doi: 10.3934/math.2023004
[1] | Kui Liu . Stability analysis for $ (\omega, c) $-periodic non-instantaneous impulsive differential equations. AIMS Mathematics, 2022, 7(2): 1758-1774. doi: 10.3934/math.2022101 |
[2] | A.G. Ibrahim, A.A. Elmandouh . Existence and stability of solutions of $ \psi $-Hilfer fractional functional differential inclusions with non-instantaneous impulses. AIMS Mathematics, 2021, 6(10): 10802-10832. doi: 10.3934/math.2021628 |
[3] | Muneerah AL Nuwairan, Ahmed Gamal Ibrahim . The weighted generalized Atangana-Baleanu fractional derivative in banach spaces- definition and applications. AIMS Mathematics, 2024, 9(12): 36293-36335. doi: 10.3934/math.20241722 |
[4] | Muneerah Al Nuwairan, Ahmed Gamal Ibrahim . Nonlocal impulsive differential equations and inclusions involving Atangana-Baleanu fractional derivative in infinite dimensional spaces. AIMS Mathematics, 2023, 8(5): 11752-11780. doi: 10.3934/math.2023595 |
[5] | Velusamy Kavitha, Mani Mallika Arjunan, Dumitru Baleanu . Non-instantaneous impulsive fractional-order delay differential systems with Mittag-Leffler kernel. AIMS Mathematics, 2022, 7(5): 9353-9372. doi: 10.3934/math.2022519 |
[6] | Ping Tong, Qunjiao Zhang . Existence of solutions to Caputo fractional differential inclusions of $ 1 < \alpha < 2 $ with initial and impulsive boundary conditions. AIMS Mathematics, 2023, 8(9): 21856-21871. doi: 10.3934/math.20231114 |
[7] | Yuanlin Ding . Existence and stability analysis of solutions for periodic conformable differential systems with non-instantaneous impulses. AIMS Mathematics, 2025, 10(2): 4040-4066. doi: 10.3934/math.2025188 |
[8] | Ahmed Salem, Kholoud N. Alharbi . Fractional infinite time-delay evolution equations with non-instantaneous impulsive. AIMS Mathematics, 2023, 8(6): 12943-12963. doi: 10.3934/math.2023652 |
[9] | Ramkumar Kasinathan, Ravikumar Kasinathan, Dumitru Baleanu, Anguraj Annamalai . Hilfer fractional neutral stochastic differential equations with non-instantaneous impulses. AIMS Mathematics, 2021, 6(5): 4474-4491. doi: 10.3934/math.2021265 |
[10] | Mohamed Adel, M. Elsaid Ramadan, Hijaz Ahmad, Thongchai Botmart . Sobolev-type nonlinear Hilfer fractional stochastic differential equations with noninstantaneous impulsive. AIMS Mathematics, 2022, 7(11): 20105-20125. doi: 10.3934/math.20221100 |
It is known that there is no non-constant periodic solutions on a closed bounded interval for differential equations with fractional order. Therefore, many researchers investigate the existence of asymptotically periodic solution for differential equations with fractional order. In this paper, we demonstrate the existence and uniqueness of the S-asymptotically ω-periodic mild solution to non-instantaneous impulsive semilinear differential equations of order 1<α<2, and its linear part is an infinitesimal generator of a strongly continuous cosine family of bounded linear operators. In addition, we consider the case of differential inclusion. Examples are given to illustrate the applicability of our results.
It is known that the action of instantaneous impulses seems not describe some certain dynamics of evolution processes in Pharmacotherapy. For example, in the case of a decompensation, (high or low levels of glucose) one can prescribe some intravenous drugs (insulin). The introduction of the drugs in the bloodstream and the consequent absorption for the body are gradual and continuous processes. Thus, we do not expect to use the instantaneous impulses to describe such a process. In fact, the above situation is fallen in a new case of impulsive action, which starts at any arbitrary fixed point and stays active on a finite time interval. To this end, Hernándaz and O'Regan [1] introduced the non-instantaneous impulsive differential equations. For recent contributions on non-instantaneous impulsive differential equations and inclusions, we refer the reader to [2,3,4,5,6,7].
There are some papers where the nonexistence of non-constant periodic solutions on closed bounded interval for differential equations with fractional order are considered such as [8,9,10,11,12]. Many authors investigated the existence of S-asymptotically ω-periodic solutions for many types of differential equations of fractional order. For example, Maghsoodi et al. [13] considered an evolution equation of order α∈(0,1) generated by an evolution system U(θ,s). Ren et al.[12] studied semilinear differential equation of order α∈(0,1) and generated by exponentially stable C0-semigroup. Ren et al.[14] considered semilinear differential equations of order α∈(1,2) generated by a sectorial operator. Mu et al.[15] investigated an evolution equation with the Weyl-Liouville fractional derivative of order α∈(0,1) and generated by C0-semigroup. Zhao at al.[16] demonstrated the existence of an asymptotically almost automorphic mild solution to a semilinear fractional differential equation, and Wang et al. [17] studied delay fractional differential equations with an almost sectorial operator of order α∈(0,1). Moreover, Muslim et al. [18] investigated the existence, uniqueness and stability of solutions to second order nonlinear differential equations with non- instantaneous impulses. Very recently, Alsheekhhussain et al. [19] proved the existence of S-asymptotically w-periodic solutions for non-instantaneous impulsive differential equations and inclusions generated by sectorial operators. For more information regarding this subject, we refer the reader to [20,21,22,23,24,25].
It is worth noting that the problems considered in all the cited works above, except [19], do not contain impulseses effects and the right-hand side is a single-valued function. Moreover, to the best of the authors' knowledge, the literature concerning S-asymptotically w -periodic solutions for differential inclusions subject to non-instantaneous impulses and generated by an infinitesimal generator of a cosine family {C(θ):θ≥0} is very new, and this fact is the main aim in the present paper.
When the considered problem contains non-instantaneous impulses, there are two approaches in the literature to prove the existence of the solution. The first one is by keeping the lower limit of the fractional derivative at zero. The second one is by switching it at the impulsive points, which will be considered in the present paper.
Let α∈(1,2), E be a Banach space, N be the set of natural numbers, m ∈ N, ω>0,J=[0,∞),
0=s0< θ1<s1<⋯<θm<sm=ω<θm+1=ω+θ1<sm+1=s1+ω<..., |
with limi→∞θi=∞,sm+i=si+ω;i∈{0}∪N, θm+i=θi+ω; i∈N, and A is the infinitesimal generator of cosine family {C(θ):θ≥0}. Moreover, let Π:J×E→E, gi:[θi,si]×E⟶Ei∈N, x0∈D(A) (the domain of A ), and x1∈E a fixed point.
Motivated by the above cited works, we demonstrate the existence and uniqueness of an S-asymptotically ω-periodic solution to the following non-instantaneous impulsive semilinear differential equation:
{cDα0,θx(θ)=Ax(θ)+Π(θ,x(θ)), a.e. θ∈(si, θi+1],i∈N∪{0},x(θ)=gi(θ,x(θ−i)),θ∈(θi si],i∈N,x(0)=x0,x´(0)=x1, | (1.1) |
where, cDα0,θx(θ) is the Caputo derivative of the function x at the point θ with lower limit at 0 [26].
After that, we prove the existence of S-asymptotically ω-periodic solutions for the following non-instantaneous impulsive semilinear differential inclusion:
{cDα0,θx(θ)∈Ax(θ)+F(θ,x(θ)), a.e. θ∈(si, θi+1],i∈N∪{0},x(θ)=gi(θ,x(θ−i)),θ∈(θi si],i∈N,x(0)=x0,x´(0)=x1, | (1.2) |
where F:J×E→2E−{ϕ} is a multi-valued function.
Unlike the differential equations of integer order, the existence of non-constant periodic solutions for fractional differential equations is not guaranteed. For this reason, the concept of an asymptotically periodic solution is introduced for fractional differential equations. Many researchers uses this approach to investigate the existence of the solution for fractional differential equations. However, up to now, there are no work studying the problem mentioned above. In this paper, we construct sufficient conditions that assure the existence of asymptotically periodic mild solutions for Problems (1.1) and (1.2). Moreover, our results generalize the obtained ones in [12], and our method can be used to study the existence of asymptotically periodic mild solutions for the problems considered in [13,15,16,17,20,21,22,23,24,25], when these problems contain impulseses effects and the right hand side is a multi-valued function.
Since a multivalued function is a function values are sets, so, our technique to find an asymptotically periodic solution for Problem (2) can be used to extend many recent publications on the same subject in which the right hand side is a single-function see, for example, [27,28,29].
In Section 3, we prove the existence and uniqueness of S-asymptotically ω-periodic solution for Problem (1.1). Section 4 is devoted to prove the existence of S-asymptotically ω-periodic solutions to Problem (1.2). Finally, examples are given to show that the obtained results are applicable.
Let J0=[0,θ1], Ji=(θi,θi+1], and i∈N. Because Problem (1.1) contains non-instantaneous impulses effect, we consider the two Banach spaces:
PC(J,E):={x:J→E, x|Ji∈C(Ji,E),x(θ+i) and x(θ−i) exist, i∈N }, |
and
PCb(J,E):={x ∈PC(J,E):x is bounded, x|Ji∈C(Ji,E)}, |
where
||x||PC(J,E):=maxθ∈J||x(θ)||E, |
||x||PCb(J,E)):=maxθ∈J||x(θ)||E, |
and x(θ+i) and x(θ−i) are the right and left limits of x at θi.
Definition 2.1. Let ω be a positive real number. A function x∈PCb(J,E) is said to be S-asymptotically ω-periodic if it satisfies the relation:
limθ→∞||x(θ+ω)−x(θ)||=0. |
Definition 2.2. [19] By SAPωPCb(J,E), we mean the Banach space of all S-asymptotically ω-periodic functions x∈PCb(J,E), where the norm is given by
||x||PCb(J,E)):=maxθ∈J||x(θ)||E. |
Definition 2.3. [30] A family {C(θ):θ∈R}, where C(θ):D(C(θ))=E→E is a bounded linear operator, is called a strongly cosine family if:
(ⅰ) C(0)=I,
(ⅱ) C(θ+τ)+C(τ−θ)=2C(τ)C(θ) for all τ,θ∈R,
(ⅲ) the map θ⟼C(θ)x is continuous for each x∈E.
If {C(θ):θ∈R} is a strongly cosine family, then the strongly continuous sine family associated with it is defined by:
S(θ)x=∫θ0C(s)xds;θ∈R,x∈E. |
Definition 2.4. The infinitesimal generator of a cosine family {C(θ):θ∈R} is an operator A:D(A)⟼E defined by
Ax=d2dθ2C(θ)x|θ=0, |
where D(A)={x∈E:C(t)x is twice continuously differentiable of t}.
Lemma 2.1. ([30], Propositions 2.2 and 2.3]) Let {C(t):t∈R} be a strongly cosine family in E with infinitesimal generator A and
Z={z∈E:C(θ)x is once continuously differentiable of θ }. |
Then, the following statements hold:
1- D(A) is dense in E, and A is a closed operator.
2- If z∈E, then S(θ)z∈Z.
3- If z∈Z, then
(ⅰ) S(θ)z∈D(A) and d2dθ2S(θ)z=AS(θ)z,
(ⅱ) S(θ)z∈D(A) and ddθC(θ)z=AS(θ)z.
4- If z∈D(A), then
(ⅰ) C(θ)z∈D(A) and d2dθ2C(θ)z=AC(θ)x=C(θ)Az;
(ⅱ) S(θ)z∈D(A) and AS(θ)z=S(θ)Az.
Definition 2.5. ([31]) By a mild solution for Problem (1.1), we mean a function x∈PC(J,E) such that
x(θ)={Cq(θ)x0+Kq(θ)x1+∫θ0(θ−τ)q−1Pq(θ−τ)Π(τ,x(τ))dτ,θ∈[0,θ1],gi(θ,x(θ−i)),θ∈(θi,si],i∈N,Cq(θ−si)gi(si,x(θ−i))+Kq(θ−si)g′i(si,x(θ−i))−∫si0(si−τ)q−1Pq(si−τ)Π(τ,x(τ))dτ+∫θ0(θ−τ)q−1Pq(θ−τ)Π(τ,x(τ))dτ,θ∈[si,θi+1],i∈N, | (2.1) |
where q=α2, and, for ϑ≥0,
Cq(ϑ)=∫∞0ξq(θ)C(ϑqθ)dθ,Kq(ϑ)=∫ϑ0Cq(τ)dτ, |
Pq(ϑ)=q∫∞0θξq(θ)S(ϑqθ)dθ, |
ξq(θ)=1qθ−1−1qwq(θ−1q), θ∈(0,∞), |
and
wq(θ)=1π∞∑n=1(−1)n−1θ−qn−1Γ(nq+1)n!sin(nπq), θ∈(0,∞). |
Remark 2.1. The solution function given by (2.1) satisfies the following properties:
1- x(0)=Cq(0)x0=x0.
2 x´(0)=x1.
3- x is continuous on Ji; i∈{0}∪N.
We will need the following lemma which gives some properties for the operators Cq(θ),Kq(θ) and Pq(θ).
Lemma 2.2. ([31], Lemma 8). Assume that
(HA) A:D(A)→E is the infinitesimal generator of strongly continuous cosine family of linear operators {C(θ):θ≥0} which is uniformly bounded by M >0. Then,
(ⅰ) For any fixed θ≥0,Cq(θ),Kq(θ) and Pq(θ) are linear bounded operators.
(ⅱ) For γ∈[0,1],∫∞0θγξα(θ)dθ=Γ(1+γ)Γ(1+αγ).
(ⅲ) If ||Cq(θ)||≤M,θ≥0, then for any x∈E, ||Cq(θ)x|| ≤M||x||, ||Kq(θ)x|| ≤θM||x|| and ||Pq(θ)x|| ≤MΓ(2q)||x||θq.
(ⅳ) {Cq(θ),θ≥0},{Kq(θ),θ≥0} and {θq−1Pq(θ),θ≥0} are strongly continuous.
We make the following assumptions:
(HA)∗A:D(A)→E satisfies (HA), and the family {C(θ): θ≥0} is exponentially stable. That is, there exist positive numbers a, M such that ||C(θ)||≤e−aθM, θ≥0.
(HΠ) Π:J×E→E is a strongly measurable function, and there are h1,h2∈C(J,R+) such that h1is bounded,
||Π(θ,x)−Π(θ,y)||E≤h1(θ)||x−y||E,∀θ∈J, x,y∈E, | (3.1) |
and
||Π(θ+ω,x)−Π(θ,x)||E≤h2(θ)(||x||E+1),∀θ∈J, x∈E. | (3.2) |
(Hg) For any i∈N, gi:[θi,si]×E⟶E (i∈N) such that, for any x∈E, the function θ↦gi(θ,x) is differentiable at si , and that:
(ⅰ)
limθ→∞i→∞||gi+m(θ+ω,z)−gi(θ,z)||E=0,∀z∈E, | (3.3) |
and
limi→∞||g′i+m(si+ω,z)−g′i(si,z)||E=0,∀z∈E. | (3.4) |
(ⅱ) There are N>0 such that
||gi(θ,z1)−gi(θ,z2)||E≤N||z1−z2||E, ∀θ∈[θi,si], ∀z1,z2∈E. | (3.5) |
(ⅲ) There is N>0 such that
||g′i(si,z1)−g′i(si,z2)||E≤N||z1−z2||E, ∀ z1,z2∈E. | (3.6) |
(ⅳ) There is κ1>0 such that
supi∈Nsupθ∈J||gi(θ,z)||E≤κ1(||z||E+1),∀z∈E. | (3.7) |
(ⅴ) There is κ2>0 with
supi∈N||g′i(si,z)||E≤κ2(||z||E+1),∀z∈E. | (3.8) |
The following lemma provides additional properties for the operators Cq(θ) and Pq(θ) when {C(θ):θ≥0} is exponentially stable.
Lemma 3.1. ([32], Proposition 2.1). If (HA)∗ is verified, then there is L>0 such that
||Cq(θ)||≤L(1+θ)q,||Pq(θ)||≤L(1+θ)2q,∀θ∈J. | (3.9) |
Lemma 3.2. ([33], Lemma 2.11]) Let γ∈[0,1], 0<a<b. Then, |bγ−aγ|≤(b−a)γ.
Remark 3.1. In what follows, we mean by || || the norm in the Banach space E.
Theorem 3.1. Under conditions (HA)∗,(HΠ),(Hgi) and (H), Problem (1.1) has a unique S-asymptotically ω-periodic mild solution providedthat the following assumptions are verified:
ς=supθ∈J∫θ0(θ−τ)q−1(1+θ−τ)2qh1(τ)dτ<∞, | (3.10) |
MN+MωN+2Lς<1, | (3.11) |
ξ=supτ∈[0,ω]||Π(τ,0)||E<∞, | (3.12) |
and
limθ→∞∫θ0(θ−τ)q−1(1+θ−τ)2qh2(τ)dτ=0, | (3.13) |
where h1 and h2 are specified in (HΠ).
Proof. First, we clarify that if x∈SAPωPCb(J,E), then the function Φ(x) defined by
Φ(x)(θ)={Cq(θ)x0+Kq(θ)x1+∫θ0(θ−τ)q−1Pq(θ−τ)Π(τ,x(τ))dτ,θ∈[0,θ1],gi(θ,x(θ−i)),θ∈(θi,si],i∈N,Cq(θ−si)gi(si,x(θ−i))+Kq(θ−si)g′i(si,x(θ−i))−∫si0(si−τ)q−1Pq(si−τ)Π(τ,x(τ))dτ+∫θ0(θ−τ)q−1Pq(θ−τ)Π(τ,x(τ))dτ,θ∈[si,θi+1],i∈N, | (3.14) |
belongs to SAPωPCb(J,E). The proof will be given in the following steps.
Step 1. we will show that limθ→∞||Φ(x)(θ+ω)−Φ(x)(θ)||=0.
Let ϵ>0. Because x∈SAPωPC(J,E), limθ→∞||x(θ+ω)−x(θ)||E=0, and hence there is θϵ>θ1 such that
supθ>θϵ||x(θ+ω)−x(θ)||E<ϵLς. | (3.15) |
Let θ>θϵ. If θ∈(θi,si], i∈N, then θ+ω∈(θi+ω,si+ω]=(θi+m,si+m]. So, relations (3.3), (3.5) and (3.14) imply that
limθ→∞||Φ(x)(θ+ω)−Φ(x)(θ)||E=limθ→∞||gi+m(θ+ω,x(θ−i+m))−gi(θ,x(θ−i))||≤limθ→∞i→∞||gi+m(θ+ω,x(θ−i+ω))−gi+m(θ,x(θ−i+ω))||+limθ→∞i→∞||gi+m(θ,x(θ−i+ω))−gi(θ,x(θ−i))||≤N limθi→∞||x(θ−i+ω)−x(θ−i)||E=0. | (3.16) |
Let θ∈[si,θi+1],i∈N. Then, θ+ω∈[si+ω,θi+1+ω]=[si+m,θi+m+1]. By arguing as in (3.16), one obtains
limθ→∞ ||Cq(θ+ω−(si+ω))gi+m(si+ω,x(θ−i+ω))−Cq(θ−si)gi(si,x(θ−i))||=M limθ→∞||gi+m(si+ω,x(θ−i+ω))−gi(si,x(θ−i))||=0. | (3.17) |
Similarly, by (3.4) and (3.6), we get
limθ→∞||Kq(θ+ω−(si+ω))g′i+m(si+ω,x(θ−i+ω))−Kq(θ−si)g′i(si,x(θ−i))||=limθ→∞||Kq(θ−si)|| ||g´i+m(si+ω,x(θ−i+ω))−g′i(si,x(θ−i))||≤limθ→∞M (θ−si)||g´i+m(si+ω,x(θ−i+ω))−g′i(si,x(θ−i))||≤M (θi+1−si)[limθ→∞i→∞||g´i+m(si,x(θ−i+ω))−g´(si,x(θ−i))||+Nlimθ→∞||x(θ−i+ω)−x(θ−i)||=0. | (3.18) |
Next, notice that
∫θ+ω0(θ+ω−τ)q−1Pq(θ+ω−τ)Π(τ,x(τ))dτ=∫θ−ω(θ−τ)q−1Pq(θ−τ)Π(τ+ω,x(τ+ω))dτ. |
Then,
||∫θ+ω0(θ+ω−τ)q−1Pq(θ+ω−τ)Π(τ,x(τ))dτ−∫θ0(θ−τ)q−1Pq(θ−τ)Π(τ,x(τ))dτ||=||∫θ−ω(θ−τ)q−1Pq(θ−τ)Π(τ+ω,x(τ+ω))dτ−∫θ0(θ−τ)q−1Pq(θ−τ)Π(τ,x(τ))dτ||≤||∫0−ω(θ−τ)q−1Pq(θ−τ)Π(τ+ω,x(τ+ω))dτ||+||∫θ0(θ−τ)q−1Pq(θ−τ)(Π(τ+ω,x(τ+ω))−Π(τ,x(τ+ω)))dτ||+||∫θ0(θ−τ)q−1Pq(θ−τ)(Π(τ,x(τ+ω))−Π(τ,x(τ)))dτ||.=Q1+Q2+Q3. | (3.19) |
Note that, from Lemma 3.1, (θ+ω)q−θq≤ωq. Hence, by taking into account τ∈[−ω,0]⟹τ+ω∈[0,ω], it yields from (3.9)
Q1=||∫0−ω(θ−τ)q−1Pq(θ−τ)Π(τ+ω,x(τ+ω))dτ||≤Lsups∈[0,ω],||v||≤||x||SAPωPC(J,E)||Π(s,v)||∫0−ω(θ−τ)q−1(1+θ−τ)2qdτ≤Lϰ.(1+θ)2q∫0−ω(θ−τ)q−1dτ=Lσx,q(1+θ)2q((θ+ω)q−θq)≤Lϰ.ωqq(1+θ)2q, | (3.20) |
where, ϰ=sups∈[0,ω],||v||≤||x||SAPωPCb(J,E)||Π(s,v)||.
Next, by (3.1), (3.2), (3.9), (3.10) and (3.15), we get
Q2=||∫θ0(θ−τ)q−1Pq(θ−τ)(Π(τ+ω,x(τ+ω))−Π(τ,x(τ+ω)))dτ||≤L∫θ0(θ−τ)q−1(1+θ−τ)2q(1+||x(τ+ω))||h2(τ)dτ≤L(1+||x||SAPωPCb(J,E))∫θ0(θ−τ)q−1(1+θ−τ)2qh2(τ)dτ, | (3.21) |
and
Q3=∫θ0(θ−τ)q−1||Pq(θ−τ)|| ||Π(τ,x(τ+ω))−Π(τ,x(τ))||dτ≤L∫θ0(θ−τ)q−1(1+θ−τ)2q||x(τ+ω))−x(τ)||h1(τ)dτ≤L∫θϵ0(θ−τ)q−1(1+θ−τ)2q||x(τ+ω))−x(τ)||h1(τ)dτ+L∫θθϵ(θ−τ)q−1(1+θ−τ)2q||x(τ+ω))−x(τ)||h1(τ)dτ<c1c2L∫θϵ0(θ−τ)q−1(1+θ−τ)2qdτ+ϵ<c1c2L∫θϵ0(θ−τ)−q−1dτ+ϵ<c1c2Lθ−q−(θ−θϵ)−qq+ϵ, | (3.22) |
where c1=supτ∈[0,θϵ]||x(τ+ω))−x(τ)|| and c2=supτ∈[0,θϵ]h1(τ). Combining (3.19–3.22), one obtains,
limθ→∞||∫θ+ω0(θ+ω−τ)q−1Pq(θ+ω−τ)Π(τ,x(τ))dτ−∫θ0(θ−τ)q−1Pq(θ−τ)Π(τ,x(τ))dτ||<limθ→∞Lϰ.ωqq(1+θ)2q+L(1+||x||)limθ→∞∫θ0(θ−τ)q−1(1+θ−τ)2qh2(τ)dτ+c1c2Llimθ→∞(θ−τ)−q−θ−qq+ϵ. | (3.23) |
Similarly,
||∫si+ω0(si+ω−τ)q−1Pq(si+ω−τ)Π(τ,x(τ))dτ−∫si0(si−τ)q−1Pq(si−τ)Π(τ,x(τ))dτ||<Lϰ.ωqq(1+si)2q+L(1+||x||)∫si0(si−τ)q−1(1+si−τ)2qh2(τ)dτ+c1c2Lθ−q−(θ−θϵ)−qq+ϵ. | (3.24) |
Note that si→∞ when θ→∞. Therefore, using (3.16)–(3.18), (3.13) and (3.24), we derive limθ→∞||Φ(x)(θ+ω)−Φ(x)(θ)||=0.
Step 2. We show that, for any x∈SAPωPCb(J,E), Φ(x) is bounded.
Let θ∈J.
(ⅰ) Let θ∈[0,θ1]. Then, applying Lemma (1.2) (ⅲ), (3.9) and (3.14), one gets
||Φ(x)(θ)||≤M||x0||+Mω||x1||+L∫θ0(θ−τ)q−1(1+θ−τ)2q||Π(τ,x(τ))||dτ. | (3.25) |
On the hand, from (3.1), we get
||Π(τ,x(τ))||≤≤||Π(τ,0||+h1(τ)||x(τ)||≤ξ+h1(τ)||x||SAPωPCb(J,E). | (3.26) |
By (3.25) and (3.26), we get
||Φ(x)(θ)||≤M||x0||+Mω||x1||+Lξ∫θ0(θ−τ)q−1(1+θ−τ)2qdτ+ςL||x||SAPωPCb(J,E)≤M||x0||+Mω||x1||+Lξ∫θ0δq−1(1+δ)2qdδ+ςL||x||SAPωPCb(J,E)≤M||x0||+Mω||x1||+Lξ∫∞0δq−1(1+δ)2qdδ+ςL||x||SAPωPCb(J,E)=M||x0||+Mω||x1||+LξB(q,q)+ςL||x||SAPωPCb(J,E), | (3.27) |
where B is the beta function. Hence, y is bounded on [0,θ1].
(ⅱ) If θ∈(θi,si], i∈N, then, by (3.7), it yields
||Φ(x)(θ)||=||gi(si,x(θ−i))||≤κ1(||x||+1),∀z∈E. | (3.28) |
(ⅲ) If θ∈(si,θi+1], then it follows from (3.8) and Lemma (1.2) that
||Cq(θ−si)gi(si,x(θ−i))+Kq(θ−si)g′i(si,x(θ−i))||≤Mκ1(1+||x(θ−i)||)+M ωκ1(1+||x(θ−i)||). | (3.29) |
Moreover, as in (3.27),
||∫θ0(θ−τ)q−1Pq(θ−τ)Π(τ,x(τ))dτ||≤L∫θ0(θ−τ)q−1(1+θ−τ)2qΠ(τ,x(τ))dτ||≤LξB(q,q)+ςL||x||SAPωPCb(J,E). | (3.30) |
Similarly, we can derive
∫si0(si−τ)q−1Pq(si−τ)Π(τ,x(τ))dτ≤LξB(q,q)+ςL||x||SAPωPCb(J,E). | (3.31) |
As a result of (3.27)–(3.31), we conclude that y is bounded on J.
Now, Φ(x) is continuous on Jii∈{0}∪N, and, hence, from Steps 1 and 2, we confirm that Φ(x)∈SAPωPCb(J,E). Thus, Φ is a function from SAPωPCb(J,E) to itself.
Step 3. We show in this step that Φ is a contraction mapping from SAPωPCb(J,E) to SAPωPCb(J,E).
To show this, let x,y∈SAPωPCb(J,E).We have three cases.
Case 1. θ∈[0,θ1]
Using (3.14), it yields
||Φ(x)(θ)−Φ(y)(θ)||≤||∫θ0(θ−τ)q−1Pq(θ−τ)Π(τ,x(τ))dτ−∫θ0(θ−τ)q−1Pq(θ−τ)Π(τ,y(τ))dτ||. | (3.32) |
Using Lemma 1.2, (3.2), (3.4), (3.8) and (3.9), relation (3.32) becomes
||Φ(x)(θ)−Φ(y)(θ)||≤L||x−y||SAPωPCb(J,E) ∫θ0(θ−τ)q−1(1+θ−τ)2qh1(τ)dτ≤Lς||x−y||SAPωPCb(J,E). | (3.33) |
Case 2. θ∈(θi,si], i∈N. Relations (3.5) and (3.14) lead to
||Φ(x)(θ)−Φ(y)(θ)||=||gi(θ,x(θ−i))−gi(θ,y(θ−i))||≤N||x(θ−i)−y(θ−i)||≤N||x−y||SAPωPCb(J,E), | (3.34) |
where N=max1≤i≤m{Ni}.
Case 3. θ∈[si,θi+1],i∈N. It yields from (3.5), (3.6) and (3.14)
||Cq(θ−si)gi(si,x(θ−i))−Cq(θ−si)gi(si,y(θ−i))||≤MN||x−y||SAPωPCb(J,E), | (3.35) |
and
||Kq(θ−si)g′i(si,x(θ−i))−Kq(θ−si)g′i(si,y(θ−i))||≤MωN||x(θ−i)−y(θ−i)||≤MωN||x−y||SAPωPCb(J,E), | (3.36) |
where N=max1≤i≤m{Ni}.
Moreover, similar to (3.33),
||∫θ0(θ−s)q−1Pq(θ−τ)Π(τ,x(τ))dτ−∫θ0(θ−τ)q−1Pq(θ−τ)Π(τ,y(τ))dτ||≤Lς||x−y||SAPωPCb(J,E), | (3.37) |
and
||∫si0(si−τ)q−1Pq(si−τ)Π(τ,x(τ))dτ|−∫si0(si−τ)q−1Pq(si−τ)Π(τ,y(τ))dτ||≤Lς||x−y||SAPωPCb(J,E). | (3.38) |
Due to (3.33)–(3.38), we conclude that
||Φ(x)−Φ(y)||≤||x−y||SAPωPCb(J,E)(MN+MωN+2Lς). | (3.39) |
It yields from (3.11) and (3.39) that Φ is contraction. Applying the Banach fixed point theorem, we have that Φ has a unique fixed-point which is an S-asymptotically ω-periodic solution to Problem (1.1).
Remark 3.2. If h1is bounded on J, then relation (3.10) is verified. In fact, suppose that h1(τ)≤κ,∀τ∈J. We have
ς=∫θ0(θ−τ)q−1(1+θ−τ)2qh1(τ)dτ≤κ∫θ0(θ−τ)q−1(1+θ−τ)2qdτ≤κ∫θ0δq−1(1+δ)2qdδ.≤κ∫∞0δq−1(1+δ)2qdδ=κB(q,q)<∞, | (3.40) |
where B is the beta function. Thus, (3.10) is verified.
Remark 3.3. If limθ→∞∫θ0h2(τ)dτ=0, then relation (3.13) is verified. In fact, we have
limθ→∞∫θ0(θ−τ)q−1(1+θ−τ)2qh2(τ)dτ≤limθ→∞∫θ0h2(τ)dτ=0. | (3.41) |
Corollary 3.1. Suppose that conditions (HA) and (Hgi) ) are satisfied. If (HΠ) is verified with h1(τ)≤κ,∀τ∈J, and limθ→∞∫θ0h2(τ)dτ=0 then, by (3.41) and Theorem (1.1), Problem (1.1) has a unique S-asymptotically ω-periodic provided that
MN+MωN+2LκB(q,q)<1. | (3.42) |
Remark 3.4. If there is no impulses effect, then N=N=0. Hence, relations (3.42) becomes 2LκB(q,q)<1.
In this section, we demonstrate the existence of S-asymptotically ω-periodic mild solutions for 1.2. We denote by Pck(E) the family of non-empty, convex and compact subsets of E.
Consider the following assumptions:
(HF)F:J×E→Pck(E) is a multi-valued function such that:
(ⅰ) For any z∈E, the multi-valued function θ→ F(.,z) is strongly measurable.
(ⅱ) For any x∈PC(J,E), the set
S1F(..x(.)):={φ :J→E, φ is locally integrable, and φ(τ)∈F(τ,x(τ)),a.e.θ∈J} |
is not empty.
(ⅲ) There is a measurable bounded, almost everywhere, function L1:J→J such that
h(F(θ,z1),F(θ,z2))≤L1(θ)||z1−z2||,∀θ∈J, u,z2∈E, | (4.1) |
where h is the Hausdorff distance.
(ⅳ) There is L2∈C(J,R+) such that
h(F(θ+ω,z),F(θ,z))≤L2(θ)||1+z||,∀θ∈J, z∈E. | (4.2) |
(ⅴ) The function
t⟼σ(τ):=||F(τ,0)|=supz∈F(τ,0)||z|| | (4.3) |
is bounded almost everywhere on J.
We need the following Lemma, which is due to Covitz and Nadler [34].
Lemma 4.1. Let (X,d) be a metric space and G be a contraction multi-valuedfunction from X to the family of non-empty closed subsets of X. Then, G has a fixed point.
Theorem 4.1. Under conditions (HA)∗,(HF),(Hgi) and (H), Problem (1.2) has an S-asymptotically ω-periodic mild solution provided that
limθ→∞∫θ0(θ−τ)q−1(1+θ−τ)2qσ(τ)dτ=0, | (4.4) |
limθ→∞∫θ0(θ−τ)q−1(1+θ−τ)2qL1(τ)dτ=0, | (4.5) |
limθ→∞∫θ0(θ−τ)q−1(1+θ−τ)2qL2(τ)dτ=0, | (4.6) |
and
MN+MωN+2Lω1B(q,q)<1, | (4.7) |
where |L1(t)|≤λ1, a.e.t∈J.
Proof. Due to (HF)(ii), for any x∈SAPωPCb(J,E), the set S1F(.,x(.)) is not empty. Therefore, for any x∈SAPωPCb(J,E), we can define a multi-valued function R(x) as follows: an element y∈R(x) if and only if
y(θ)={Cq(θ)x0+Kq(θ)x1+∫θ0(θ−τ)q−1Pq(θ−τ)f(τ)dτ,θ∈[0,θ1],gi(θ,x(θ−i)),θ∈(θi,si],i∈N,Cq(θ−si)gi(si,x(θ−i))+Kq(θ−si)g′i(si,x(θ−i))−∫si0(si−τ)q−1Pq(si−τ)f(τ)dτ+∫θ0(θ−τ)q−1Pq(θ−τ)f(τ)dτ,θ∈[si,θi+1],i∈N, | (4.8) |
where f∈S1F(.,x(.)). Since the proof is similar to what was shown in Theorem 1.1, we will illustrate only the differences.
Step 1. We show that if x∈SAPωPCb(J,E) and y∈R(x), then limθ→∞||y(θ+ω)−y(θ)||=0.
Let ϵ>0. Because x∈SAPωPCb(J,E), then limθ→∞||x(θ+ω)−x(θ)||=0 and, hence, there is θϵ>θ1 such that (3.15) is verified.
Let y∈R(x) and θ∈[si,θi+1]. According to (4.8), we have
||∫θ+ω0(θ+ω−τ)q−1Pq(θ+ω−τ)f(τ)dτ−∫θ0(θ−τ)q−1Pq(θ−τ)f(τ)dτ||=||∫θ−ω(θ−τ)q−1Pq(θ−τ)f(τ+ω)dτ−∫θ0(θ−τ)q−1Pq(θ−τ)f(τ)dτ||≤||∫0−ω(θ−τ)q−1Pq(θ−τ)f(τ+ω)dτ||+||∫θ0(θ−τ)q−1Pq(θ−τ)f(τ+ω)dτ−∫θ0(θ−τ)q−1Pq(θ−τ)f(τ)dτ||=I1+I2. | (4.9) |
Let τ∈[−ω,0] be fixed. Since F(τ+ω,0) is compact, there is vτ+ω∈ F(τ+ω,0) such that
||f(τ+ω)−vτ+ω||=d(f(τ+ω),F(τ+ω,0)≤h(F(τ+ω,x(τ+ω)),F(τ+ω,0)). | (4.10) |
From (4.1), (4.3) and (4.10), we get
||f(τ+ω)||≤h(F(τ+ω,x(τ+ω)),F(τ+ω,0))+||vτ+ω||≤L1(τ+ω)||x(τ+ω)||+σ(τ+ω)≤||x||L1(τ+ω)+σ(τ+ω),∀τ∈[−ω,0]. | (4.11) |
Then, by (3.9) and (4.11), it follows that
limθ→∞I1=limθ→∞||∫0−ω(θ−τ)q−1Pq(θ−τ)f(τ+ω)dτ||≤limθ→∞(ω1||x||+ω2)L∫0−ω(θ−τ)q−1(1+θ−τ)2qdτ≤(λ1||x||+λ2)Llimθ→∞1(1+θ)2q∫0−ω(θ−τ)q−1dτ=(λ1||x||+λ2)Llimθ→∞(θ+ω)q−θqq(1+θ)2q≤(λ1||x||+λ2)Llimθ→∞ωqq(1+θ)2q=0, | (4.12) |
where λ2 is a positive number such that σ(θ)≤λ2,a.e.,θ∈J.
Next, let τ∈[0,θ] be fixed. From the fact that F(τ+ω,x(τ)) is compact, there are zτ+ω,zτ∈F(τ,x(τ+ω)) such that d(f(τ+ω),zτ+ω)=d(f(τ+ω),F(τ,x(τ+ω))) and d(f(τ),zτ)=d(f(τ),F(τ,x(τ+ω))). Then, by (4.1) and (4.2), it yields
||f(τ+ω)−f(τ)||≤||f(τ+ω)−zτ+ω||+||zτ+ω−zτ||+||zτ−f(τ)||≤d(f(τ+ω),F(τ+x(τ+ω)))+||zτ+ω−zτ||+d(f(τ,F(τ,x(τ+ω)))≤h(F((τ+ω),x(τ+ω)),F(τ+x(τ+ω)))+2||F(τ,x(τ+ω))||+h(F(τ,x(τ+ω))),F(τ,x(τ))≤L1(τ)||x((τ+ω)−x(τ)||+2||F(τ,x(τ+ω))||+L2(τ)||1+x(τ)||≤2||x||L1(τ)+2||F(τ,x(τ+ω))||+L2(τ)||1+x(τ)||. | (4.13) |
Moreover, according to (4.1) and (4.3), we get
||F(τ,x(τ+ω))||≤||F(τ,0)||+L1(τ)||x(τ+ω)||=σ(τ)+L1(τ)||x||. | (4.14) |
Then, by (4.13) and (4.14), one obtains
I2≤||∫θ0(θ−τ)q−1||Pq(θ−τ)|| f(τ+ω)−f(τ)dτ||≤4||x||L∫θ0(θ−τ)q−1(1+θ−τ)2qL1(τ)dτ+2L∫θ0(θ−τ)q−1(1+θ−τ)2qσ(τ)dτ+L(1+||x||)∫θ0(θ−τ)q−1(1+θ−τ)2qL2(τ)dτ. | (4.15) |
Using (4.4)–(4.6) and (4.15), it yields
limθ→∞I2=limθ→∞||∫θ+ω0(θ+θ−τ)q−1Pq(θ+ω−τ)f(τ+ω)dτ−∫θ0(θ−τ)q−1Pq(θ−τ)f(τ)ds||=0. | (4.16) |
Note that τi→∞ when θ→∞. Hence, as above, we derive
limθ→∞||∫si+ω0(si+ω−τ)q−1Pq(si+ω−τ)f(τ)dτ−∫si0(si−τ)q−1Pq(si−τ)f(τ)dτ||=0. | (4.17) |
Then, due to (3.16)–(3.18), (4.9), (4.12), (4.16) and (4.17), we conclude that
limθ→∞||y(θ+ω)−y(θ)||=0 |
Step 2. In this step, we show that if x∈SAPωPCb(J,E) and y∈R(x), then y is bounded.
Let θ∈[0,θ1]. Then, using Lemma 1.2, (3.9) and (4.8), one has
||y(θ)||≤M||x0||+Mω||x1||+L∫θ0(θ−τ)q−1(1+θ−τ)2q||f(τ)||dτ. | (4.18) |
On the hand, from (4.1), we get
||f(τ)||≤||F(τ,x(τ))||≤||F(τ,0)||+L1(τ)||x(τ)||≤σ(τ)+L1(τ)||x||,∀τ∈J. | (4.19) |
By (4.18) and (4.19), it yields
||y(θ)||≤M||x0||+Mω||x1||+L(λ2+||x||λ1)∫θ0(θ−τ)q−1(1+θ−τ)2qdτ=M||x0||+Mω||x1||+L(ω2+||x||ω1)∫θ0δq−1(1+δ)2qdτ=M||x0||+Mω||x1||+L(ω2+||x||ω1)B(q,q), | (4.20) |
where B is the beta function. Therefore, y is bounded on [0,θ1]. Similarly, one can show that if θ∈[si,θi+1], then
||∫θ0(θ−τ)q−1Pq(θ−τ)f(τ))dτ||≤L(ω2+||x||ω1)B(q,q), | (4.21) |
and
∫si0(si−τ)q−1Pq(si−τ)Π(τ,x(τ))dτ≤L(ω2+||x||ω1)B(q,q). | (4.22) |
Then, by (4.20)–(4.22) and by arguing as in (3.28) and (3.29), we deduce that y is bounded on J, and our claim in this step is proved.
As a result of Eqs 1.1 and 1.2, R is a multivalued function from SAPωPCb(J,E) to the non-empty subsets of SAPωPCb(J,E).
Next, in order to apply Lemma 3.2 and show that R has a fixed point, we have to show that R is a contraction where its set of values is closed. We do this in two steps.
Step 3. The set of values of R is closed.
Let x∈SAPωPCb(J,E) and (yn)n≥1be a sequences in R(x) with yn→y in SAPωPCb(J,E). Then, there is fn ∈S1F(.,x(.)) such that
yn(θ)={Cq(θ)x0+Kq(θ)x1+∫θ0(θ−τ)q−1Pq(θ−τ)fn(τ)dτ,θ∈[0,θ1],gi(θ,x(θ−i)),θ∈(θi,si],i∈N,Cq(θ−si)gi(si,x(θ−i))+Kq(θ−si)g′i(si,x(θ−i))−∫si0(si−τ)q−1Pq(si−τ)fn(τ)dτ+∫θ0(θ−τ)q−1Pq(θ−τ)fn(τ)dτ,θ∈[si,θi+1],i∈N. | (4.23) |
We have to show that y∈R(x). By arguing as in (4.19) , one obtains
||fn(τ)||≤σ(τ)+L1(τ)||x||,∀τ∈J. | (4.24) |
Now, let θ be a fixed point in J, and Jθ=[0,θ]. From the fact that σ and L1 are bounded almost everywhere, we can deduce, from (4.24), that the family {fn:n≥1} is bounded in L2(Jθ,E) and, hence, it is weakly compact in L2(Jθ,E). Thus, it has a subsequence, denoted again by (fn)n≥1, such that fn→f weakly in L2(Jθ,E). According to Mazur's lemma, we can find a sequence (zn)n≥1 of convex combinations of fn with zn→f strongly in L2(Jθ,E). Then, we can assume, without loss of generality, that zn(τ)→f (τ),a.e.τ∈Jθ. Moreover, from (4.24) and Lemma 1.2, we get
(θ−τ)q−1||Pq(θ−τ)fn(τ)||≤MΓ(2q)(θ−τ)2q−1(λ2+λ1||x||),a.e.τ∈[0,θ]. |
Note that the function τ→(θ−τ)2q−1 belongs to L1([0,θ],E). Therefore, by the continuity of Pq(.) and applying the Lebesgue dominated convergence theorem, it yields
limn→∞∫θ0(θ−τ)q−1Pq(θ−τ)fn(τ)dτ=∫θ0(θ−τ)q−1Pq(θ−τ)f(τ)dτ. | (4.25) |
Thus, from (4.25) and the continuity Pq(.) , it follows, by taking the limit as n→∞ in (4.23), that
limn→∞yn(θ)={Cq(θ)x0+Kq(θ)x1+∫θ0(θ−τ)q−1Pq(θ−τ)f(τ)dτ,θ∈[0,θ1],gi(θ,x(θ−i)),θ∈(θi,si],i∈N,Cq(θ−si)gi(si,x(θ−i))+Kq(θ−si)g′i(si,x(θ−i))−∫si0(si−τ)q−1Pq(si−τ)f(τ)dτ+∫θ0(θ−τ)q−1Pq(θ−τ)f(τ)dτ,θ∈[si,θi+1],i∈N. | (4.26) |
Note that (HF)(iv) leads to f(s)∈F(s,x (s)),a.e. s∈J and, hence, (4.26) leads to
y(θ)={Cq(θ)(x0−g(x))+Kq(θ)(x1−p(x))+∫θ0(θ−τ)q−1Pq(θ−τ)f(τ)dτ,θ∈[0,θ1],gi(θ,x(θ−i)),θ∈(θi,si],i∈N,Cq(θ−si)gi(si,x(θ−i))+Kq(θ−si)g′i(si,x(θ−i))−∫si0(si−τ)q−1Pq(si−τ)f(τ)dτ+∫θ0(θ−τ)q−1Pq(θ−τ)f(τ)dτ,θ∈[si,θi+1],i∈N. |
Then, y∈R(x).
Step 4. We show that R is a contraction.
Let u1,u2∈SAPωPC(J,E) and y1∈R(u1). Then, there is f∈S1F(.,u(.)) such that
y1(θ)={Cq(θ)x0+Kq(θ)x1+∫θ0(θ−τ)q−1Pq(θ−τ)f1(τ)dτ,θ∈[0,θ1],gi(θ,u1(θ−i)),θ∈(θi,si],i∈N,Cq(θ−si)gi(si,u1(θ−i))+Kq(θ−si)g′i(si,u1(θ−i))−∫si0(si−τ)q−1Pq(si−τ)f1(τ)dτ+∫θ0(θ−τ)q−1Pq(θ−τ)f1(τ)dτ,θ∈[si,θi+1],i∈N. | (4.27) |
Consider the multivalued function Θ:J→2E defined by:
Θ(θ)={z∈E:||z−f1(θ)||≤L1(θ)||u1(θ)−u2(θ)||,a.e.θ∈J}. |
We show that the set of values of Θ is non-empty. Let θ∈J. From (4.1), we get
h(F(θ,u1(θ)),F(θ,u2(θ)))≤L1(θ)‖u1(θ)−u2(θ)‖. |
Thus, from the compactness of F(θ,u2(θ)), there is zθ∈F(θ,u2(θ)) such that
||f1(θ)−zθ|| ≤h(F(θ,u1(θ)),F(θ,u2(θ)))≤L1(θ)‖u1(θ)−u2(θ)‖, |
which leads to Θ(θ)≠ϕ,θ∈J. Moreover, the set Λ(θ)=Θ(θ)∩F(θ,u2(θ)),θ∈J is not empty. Because the functions f1,L1,u1,u2 are measurable, Proposition 3.4 in [35] (Corollary 1.3.1(a) in [36]) guarantees that the multivalued map θ→Λ(θ) is measurable. Note that Θ(θ),θ∈J is closed. Consequently, the set of values of Λ is non-empty and compact and, hence, by Theorem 3.1.1 in [37], there exists a measurable selection f2 for Λ with
||f1(θ)−f2(θ)||≤L1(θ)‖u1(θ)−u2(θ)‖,a.e.θ∈J. | (4.28) |
Set
y2(θ)={Cq(θ)x0+Kq(θ)x1+∫θ0(θ−τ)q−1Pq(θ−τ)f2(τ)dτ,θ∈[0,θ1],gi(θ,u2(θ−i)),θ∈(θi,si],i∈N,Cq(θ−si)gi(si,u2(θ−i))+Kq(θ−si)g′i(si,u2(θ−i))−∫si0(si−τ)q−1Pq(si−τ)f2(τ)dτ+∫θ0(θ−τ)q−1Pq(θ−τ)f2(τ)dτ,θ∈[si,θi+1],i∈N. | (4.29) |
Obviously, y2∈R(u1). Now, we estimate the value of ||y1−y2||. Let θ∈[0,θ1]. Using Lemma 1.2, (3.8), (3.9) and (4.27)–(4.29), we get
||y1(θ)−y2(θ)||≤||∫θ0(θ−τ)q−1Pq(θ−τ)||f1(τ)−f2(τ)||dτ≤Lω1||u1−u2|| ∫θ0(θ−τ)q−1(1+θ−τ)2qdτ≤Lω1||u1−u2|| ∫∞0(θ−τ)q−1(1+θ−τ)2qdτ≤||u1−u2||Lω1B(q,q)). | (4.30) |
Let θ∈[si,θi+1], i∈N. As in (4.30), one can show that
||∫θ0(θ−τ)q−1Pq(θ−τ)f1(τ)dτ−∫θ0(θ−τ)q−1Pq(θ−τ)f2(τ)dτ||≤Lω1||u1−u2|| ∫θ0(θ−τ)q−1(1+θ−τ)2qdτ≤Lω1B(q,q))||u1−u2||, | (4.31) |
and
||∫si0(si−τ)q−1Pq(si−τ)f1(τ)dτ|−∫si0(si−τ)q−1Pq(si−τ)f2(τ)dτ||≤Lω1||u1−u2|| ∫si0(si−τ)q−1(1+si−τ)2qdτ≤Lω1B(q,q)||u1−u2||. | (4.32) |
Combining relations (3.34)–(3.36) and (4.30)–(4.32), it yields
||y1(θ)−y2(θ)||≤||u1−u2||.(MN+MωN+2Lω1B(q,q)). | (4.33) |
Due to (4.7), relation (4.33) becomes
||y1(θ)−y2(θ)||<ϑ||u1−u2||, | (4.34) |
where ϑ=MN+MωN+2Lλ1B(q,q)<1. By interchanging the role of y1 and y2 in the above discussion and using (4.7) and (4.34), we conclude that R is a contraction.
As a result of Steps 1.1–3.1 and by applying Lemma (3.2), R has a fixed-point which is S-asymptotically ω-periodic solution to Problem(1.2).
Remark 4.1. As in Remark (2.1), if limτ→∞σ(τ)=limτ→∞L1(τ)=limτ→∞L2(τ)=0, then relations (4.4)–(4.6) are verified.
Remark 4.2. If there is no impulses effect, then N=N=0 and, hence, relation (4.7) becomes 2Lλ1B(q,q)<1.
In this section, we give two examples as applications of our results.
Example 5.1. Let α=32, q=34, E=L2[0,π], m=4, ω=2π, J=[0,∞), si=iπ2, i∈{0}∪N, and θi=(2i−1)π2;i∈N. Observe that s4=ω and for i∈N, si+m=si+4=(i+4)π2=iπ2+2π=si+ω, and θi+m=θi+4=(2i+7)π4=(2i−1)π4+2π=θi+2π=θi+ω.
Consider an operator A:D(A)⊂E→E defined as follows: Av=v′′and
D(A):={v∈L2[0,π]: vyy∈L2[0,1],v(0)=v(π)=0}. |
Note that the operator A has the representation
Ax=∞∑n=1−n2<x, xn>xn,x∈D(A), | (5.1) |
where xn(y)=√2 sinny,n=1,2,..., is the orthonormal set of eigenfunctions of A. Moreover, A is the infinitesimal generator of a strongly continuous cosine family C(t)t∈R which is given by
C(t)(x)=∞∑n=1cosnt<x , xn>xn,x∈E, |
and the associated sine family S(t)t∈R on E is defined by
S(t)(x):=∞∑n=1sinntn<x , xn>xn,x∈E. |
It is known that ||C(t)||≤e−π2t and ||S(t)||≤e−π2t for t≥0 (see [38], P.1307). Therefore, the family {C(θ): θ≥0} is exponentially stable and the operator A satisfies (HA)∗ with M=1. Consider a function Π:J×E→E defined by
Π(θ,u)(s):=κsinu(s)+cosθ; θ∈J, u∈E, s∈[0,π], | (5.2) |
where κ>0. We demonstrate that Π satisfies the conditions of Corollary (1.1). Let u,v∈E=L2[0,π]. One has
||Π(θ,u)−Π(θ,v)||L2[0,π]=(∫π0|Π(θ,u)(s)−Π(θ,v)(s)|2ds)12=κ(∫π0|sinu(s)−sinv(s)|2ds)12≤κ(∫π0|u(s)−v(s)|2ds)12=κ||u−v||L2[0,π]. | (5.3) |
Moreover,
||Π(θ+2π,u)−Π(θ,u)||L2[0,π]=(∫π0|Π(θ+2π,u)(s)−Π(θ,u)(s)|2ds)12=0. | (5.4) |
Relations (5.3) and (5.4) leads to (HΠ), where h1(θ)=κ and h2(θ)=0,θ∈J.
Next, for any i∈N, let gi:[ti,si]×E→E, be defined as:
gi(θ,u)(s):=λ(siniθ)i2u(s) ;(θ,u)∈[ti,si]×E,s∈[0,π], | (5.5) |
where λ is a positive real number. Then,
g′i(si,u)(s):=λ(cosisi)iu(s);u∈E,s∈[0,π],i∈N. |
Obviously, gi is bounded on bounded subsets. Note that, for any i∈N, any θ∈J, and any u,v∈E, we have
limθ→∞i→∞(||gi+m(θ+2π,u)−gi(θ,u)||L2[0,π])2=limθ→∞i→∞∫π0|gi+m(θ+2π,u)(s)−gi(θ,u)(s)|2ds=limθ→∞i→∞λ2∫π0|(sin(i+m)(θ+2π))u(s)(i+m)2−(siniθ)u(s)i2|2ds=λ2limθ→∞i→∞∫π0|(sin(i+m)θ)u(s)(i+m)2−(siniθ)u(s)i2|2ds≤4λ2limθ→∞i→∞∫π0|u(s)(i+m)2+u(s)i2|2ds≤limθ→∞i→∞4λ2i4∫π0|u(s)|2ds=4λ2i4||u||2L2[0,π]=0, | (5.6) |
and
limi→∞(||g′i+m(si+2π,u)−g′i(si,u)||L2[0,π])2=limi→∞∫π0|g′i+m(si+2π,u)(s)−g′i(si,u)(s)|2ds=λ2limi→∞∫π0|(cos(i+m)(si+2π)u(s)i+m−(cosisi)u(s)i|2ds=λ2limi→∞∫π0|(cos(i+m)si)u(s)i+m−(cosisi)u(s)i|2ds≤4λ2limi→∞∫π0|u(s)i+m+u(s)i|2ds≤limi→∞4λi∫π0|u(s)|2dslimi→∞4λi||u||2L2[0,π]=0. | (5.7) |
In addition,
||gi(θ,u)−gi(θ,v)||L2[0,π]=λ(∫π0|(siniθ)u(s)i2−(siniθ)v(s)i2|2ds)12≤λ||u−v||, | (5.8) |
and
||g′i(si,u)−g′i(si,v)||L2[0,π]=λ(∫π0|(siniθ)u(s)i−(siniθ)v(s)i|2ds)12≤λ||u−v||. | (5.9) |
Furthermore,
||gi(θ,u)||L2[0,π]=λ(∫π0|(siniθ)u(s)i2|2ds)12≤λ||u||, | (5.10) |
and
||g′i(si,u)||L2[0,π]=λ(∫π0|(cosiθ)u(s)i|2ds)12≤λ||u||. | (5.11) |
As a result of relations (5.6)–(5.11), (Hgi) is satisfied where N=N=λ and κ1=κ2=λ. By applying Corollary (1.1), we conclude that Problem (1.1) has a unique S−asymptotically 2π−periodic mild solution provided that
λ(1+ω)+2κLB(q,q)<1, | (5.12) |
where A, Π,gi are given by (5.1), (5.2) and (5.5), respectively, and L appears in (3.9). By choosing λ and κ sufficiently small, we can derive (5.12).
Example 5.2. Assume that A,α,q, E, m, ω, J, si, θi,i∈N are as in Example (1.1). Let Z be a non-empty convex compact subset of E, L1:J→J be a measurable bounded almost everywhere function such that Limθ→∞L1(θ)=0 and F:J×E→Pck(E) be a multi-valued function defined by
F(θ,u)=L1(θ)||u||sinθ σ (1+||u||)Z;(θ,u)∈J×E, | (5.13) |
where σ is a constant such that Sup{ ||z || : z∈Z}≤σ. Clearly, for every u∈E,θ→F(θ,u) is strongly measurable and, for any x∈PC(J,E), the function f(θ)=L1(θ) ||x(θ)|| sinθ σ (1+||x(θ)||)z0, z0∈Z is locally integrable, and f(θ)∈F(θ,x(θ)),θ∈J. Moreover, using (5.13), for any u,v∈E and any θ∈J, we have
H(F(θ,u),F(θ,v))≤L1(θ)|sinθ| |||u||(1+||u||)−||v||(1+||v||)|≤L1(θ)||u−v||, | (5.14) |
H(F(θ+2π,u),F(θ,u))=0, | (5.15) |
and
supθ∈J||F(θ,0)||={0}. | (5.16) |
Then, from (5.14)–(5.16), it follows that assumption (HF) is verified where L2(θ)=σ(θ)=0, θ∈J. Thus, applying Theorem 1.2, Problem (1.2), where A, F,gi are given by (5.1), (5.13) and (5.5), respectively, and L appears in (3.9), has an S−asymptotically 2π−periodic mild solution provided that
λ+2πλ+2Lλ1B(q,q)<1, |
where λ1 is a positive number such that |L1(θ)|≤ λ1,a.e. for θ∈J.
Because, in some works, it was demonstrated that there are no non-stationary periodic solutions of fractional differential equations, studying the existence of S-asymptotically ω-periodic solutions for fractional differential equations is necessary and important. Sufficient conditions that assure the existence of S-asymptotically ω-periodic solutions for non-instantaneous impulsive semilinear differential equations of order 1<α<2 and generated by the infinitesimal generator of a strongly continuous cosine family of bounded linear operators have been obtained. Also, the case when the single-valued function in the right-hand side is replaced by a multi-valued function is investigated. Examples are given to demonstrate the possibility of applicability of our results. Moreover, our results generalize the obtained one in [12] into the case where the order is 1<α<2, there are non-instantaneous impulse effects, and the right-hand side is a multi-valued function instead of a single-valued-function. Furthermore, our technique can be used to extend many problems that are considered in the literatures such as [13,15,16,17,20,21,22,23,24,25,27,28,29] to the case where there are non-instantaneous impulse effects and the right-hand side is a multi-valued function instead of a single-valued-function.
This research has been funded by the Scientific Research Deanship at University of Ha'il-Saudi Arabia through project number RG-21 101.
The authors declare that they have no conflict of interest.
[1] |
E. Hernandez, D. O'Regan, On a new class of abstract impulsive differential equation, Proc. Amer. Math. Soc., 141 (2013), 1641–1649. https://doi.org/10.1090/S0002-9939-2012-11613-2 doi: 10.1090/S0002-9939-2012-11613-2
![]() |
[2] |
A. G. Ibrahim, A. A. Elmandouh, Existence and stability of solutions of ψ-Hilfer fractional functional differential inclusions with non-instantaneous impulses, AIMS Math., 6 (2021), 10802–10832. https://doi.org/10.3934/math.2021628 doi: 10.3934/math.2021628
![]() |
[3] |
J. R. Wang, M. Li, D. O'Regan, M. Fečkan, Robustness for nonlinear evolution equation with non-instantaneous effects, B. Sci. Math., 159 (2020), 102827. https://doi.org/10.1016/j.bulsci.2019.102827 doi: 10.1016/j.bulsci.2019.102827
![]() |
[4] |
J. R. Wang, A. G. Ibrahim, D. O'Regan, Global attracting solutions to Hilfer fractional non-instantaneous impulsive semilinear differential inclusions of Sobolev type and with nonlocal conditions, Nonlinear Anal. Model., 24 (2019), 775–803. https://doi.org/10.15388/NA.2019.5.6 doi: 10.15388/NA.2019.5.6
![]() |
[5] |
J. R. Wang, A. G. Ibrahim, D. O'Regan, Hilfer type fractional differential switched inclusions with non-instantaneous impulsive and nonlocal conditions, Nonlinear Anal. Model., 23 (2018), 921–941. https://doi.org/10.15388/NA.2018.6.7 doi: 10.15388/NA.2018.6.7
![]() |
[6] |
J. R. Wang, A. G. Ibrahim, D. O'Regan, Y. Zhou, A general class of non-instantaneous impulsive semilinear differential inclusions in Banach spaces, Adv. Differ. Equ., 2017 (2017), 287. https://doi.org/10.1186/s13662-017-1342-8 doi: 10.1186/s13662-017-1342-8
![]() |
[7] | J. R. Wang, A. G. Ibrahim, D. O'Regan, Noeemptness and compactness of the solution set for fractional differential inclusions with non-instantaneous impulses, Electron. J. Differ. Eq., 2019 (2019), 37. |
[8] |
M. S. Tavazoei, M. Haeri, A proof for non existence of periodic solutions in time invariant fractional order systems, Automatica, 45 (2009), 1886–1890. https://doi.org/10.1016/j.automatica.2009.04.001 doi: 10.1016/j.automatica.2009.04.001
![]() |
[9] |
I. Area, J. Losada, J. J. Nieto, On fractional derivatives and primitives of periodic of periodic functions, Abstr. Appl. Anal., 2014 (2014), 392598. https://doi.org/10.1155/2014/392598 doi: 10.1155/2014/392598
![]() |
[10] |
E. Kaslik, S. Sivasundaram, Non-existence of periodic solutions in fractional-order dynamical systems and a remarkable difference between integer and fractional-order derivatives of periodic functions, Nonlinear Anal. Real, 13 (2012), 1489–1497. https://doi.org/10.1016/j.nonrwa.2011.11.013 doi: 10.1016/j.nonrwa.2011.11.013
![]() |
[11] |
M. D. Ortigueira, J. D. Machado, J. J. Trujillo, Fractional derivatives and periodic functions, Int. J. Dynam. Control, 5 (2017), 72–78. https://doi.org/10.1007/s40435-015-0215-9 doi: 10.1007/s40435-015-0215-9
![]() |
[12] |
L. Ren, J. Wang, M. Fečkan, Asymptotically periodic behavior solutions for Caputo type fractional evolution equations, Fract. Calc. Appl. Anal., 21 (2019), 1294–1312. https://doi.org/10.1515/fca-2018-0068 doi: 10.1515/fca-2018-0068
![]() |
[13] | S. Maghsoodi, A. Neamaty, Existence and uniqueness of asymptotically w-periodic solution for fractional semilinear problem, J. Appl. Comput. Math., 8 (2019), 1–5. |
[14] |
L. Ren, J. R. Wang, D. O'Regan, Asymptotically periodic behavior of solutions of fractional evolution equations of order 1<α<2, Math. Slovaca, 69 (2019), 599–610. https://doi.org/10.1515/ms-2017-0250 doi: 10.1515/ms-2017-0250
![]() |
[15] |
J. Mu, Y. Zhou, L. Peng, Periodic solutions and S-asymptotically periodic solutions to fractional evolution equations, Discrete Dyn. Nat. Soc., 2017 (2017), 1364532. https://doi.org/10.1155/2017/1364532 doi: 10.1155/2017/1364532
![]() |
[16] |
J. Q. Zhao, Y. K. Chang, G. M. N. Guérékata, Asymptotic behavior of mild solutions to semilinear fractional differential equations, J. Optim. Theory Appl., 156 (2013), 106–114. https://doi.org/10.1007/s10957-012-0202-7 doi: 10.1007/s10957-012-0202-7
![]() |
[17] |
H. Wang, F. Li, S-asymptotically T-periodic solutions for delay fractional differential equations with almost sectorial operator, Adv. Differ. Equ., 2016 (2016), 315. https://doi.org/10.1186/s13662-016-1043-8 doi: 10.1186/s13662-016-1043-8
![]() |
[18] |
M. Muslim, A. Kumar, M. Fečkan, Existence, uniqueness and stability of solutions to second order nonlinear differential equations with non-instantaneous impulses, J. King Saud Uni. Sci., 30 (2018), 204–213. https://doi.org/10.1016/j.jksus.2016.11.005 doi: 10.1016/j.jksus.2016.11.005
![]() |
[19] |
Z. Alsheekhhussain, J. Wang, A. G. Ibrahim, Asymptotically periodic behavior of solutions to fractional non-instantaneous impulsive semilinear differential inclusions with sectorial operators, Adv. Differ. Equ., 2021 (2021), 330. https://doi.org/10.1186/s13662-021-03475-w doi: 10.1186/s13662-021-03475-w
![]() |
[20] |
F. Li, J. Liang, H. Wang, S-Asymptotically ω-periodic solution for fractional differential equations of order q∈(0,1) with finite delay, Adv. Differ. Equ., 2017 (2017), 83. https://doi.org/10.1186/s13662-017-1137-y doi: 10.1186/s13662-017-1137-y
![]() |
[21] |
A. M. Abou-El-Elai, A. L. Sadek, A. M. Mahmoud, E. Farghalyi, Asymptotic stability of solutions for a certain non-autonomous second-order stochastic delay differential equation, Turk. J. Math., 41 (2017), 576–584. https://doi.org/10.3906/mat-1508-62 doi: 10.3906/mat-1508-62
![]() |
[22] |
T. Zhang, L. Xiong, Periodic motion for impulsive fractional functional equations with piecewise Caputo derivative, Appl. Math. Lett., 101 (2020), 106072. https://doi.org/10.1016/j.aml.2019.106072 doi: 10.1016/j.aml.2019.106072
![]() |
[23] |
J. Andra, Coexistence of periodic solutions with various periods of impulsive differential equations and inclusions on tori via Poincare operators, Topol. Appl., 255 (2019), 128–140. https://doi.org/10.1016/j.topol.2019.01.008 doi: 10.1016/j.topol.2019.01.008
![]() |
[24] |
M. Fecčkan, R. J. Wang, Periodic impulsive fractional differential equations, Adv. Nonlinear Anal., 8 (2019), 482–496. https://doi.org/10.1515/anona-2017-0015 doi: 10.1515/anona-2017-0015
![]() |
[25] |
H. R. Henrique, Periodic solutions of abstract neutral functional differential equations with infinite delay, Acta Math. Hung., 121 (2008), 203–227. https://doi.org/10.1007/s10474-008-7009-x doi: 10.1007/s10474-008-7009-x
![]() |
[26] | A. A. Kilbas, H. M. Srivastava, J. J. Trujillo, Theory and applications of fractional differential equations, Elsevier, 2006. |
[27] |
T. Zhang, Y. Li, S-Asymptotically periodic fractional functional differential equations with off-diagonal matrix Mittag-Leffller functional kernels, Math. Comput. Simul., 193 (2022), 313–347. https://doi.org/10.1016/j.matcom.2021.10.006 doi: 10.1016/j.matcom.2021.10.006
![]() |
[28] |
T. Zhang, Y. Li, Exponential Euler scheme of multi-delay Caput-Fabrizio fractional-order differential quations, Appl. Math. Lett., 124 (2022), 107709. https://doi.org/10.1016/j.aml.2021.107709 doi: 10.1016/j.aml.2021.107709
![]() |
[29] |
T. Zhang, J. Zhou, Y. Liao, Exponentially stable periodic oscillation and Mittag-Leffler stabilization for fractional-order impulsive control neutral networks with piecewise Caputo derivatives, IEEE T. Cybernetics, 52 (2022), 9670–9683. https://doi.org/10.1109/TCYB.2021.3054946 doi: 10.1109/TCYB.2021.3054946
![]() |
[30] |
C. C. Travis, G. F. Webb, Cosine families abstract nonlinear second order differential equations, Acta Math. Acad. Sci. H., 32 (1978), 75–96. https://doi.org/10.1007/BF01902205 doi: 10.1007/BF01902205
![]() |
[31] |
J. W. He, Y. Liang, B. Ahmed, Y. Zhou, Nonlocal fractional evolution inclusions of order α∈(1,2), Mathematics, (2019) 2019, 7. https://doi.org/10.3390/math7020209 doi: 10.3390/math7020209
![]() |
[32] |
T. Ke, N. Lu, V. Obukhovskii, Decay solutions for a class of reactional differential varational inequalities, Fract. Calc. Appl. Anal., 18 (2015), 531–553. https://doi.org/10.1515/fca-2015-0033 doi: 10.1515/fca-2015-0033
![]() |
[33] |
J. R. Wang, Y. Zhou, Existence and controllability results for fractional semilinear differential inclusions, Nonlinear Anal. Real, 12 (2011), 3642–3653. https://doi.org/10.1016/j.nonrwa.2011.06.021 doi: 10.1016/j.nonrwa.2011.06.021
![]() |
[34] |
H. Covitz, S. B. Nadler, Multivalued contraction mapping in generalized metric space, Israel J. Math., 8 (1970), 5–11. https://doi.org/10.1007/BF02771543 doi: 10.1007/BF02771543
![]() |
[35] | C. Castaing, M. Valadier, Convex analysis and measurable multifunctions, Springer-Verlag, 1977. |
[36] |
F. Hiai, H. Umegaki, Integrals conditional expectation and martingales of multivalued functions, J. Multivariate Anal., 7 (1977), 149–182. https://doi.org/10.1016/0047-259X(77)90037-9 doi: 10.1016/0047-259X(77)90037-9
![]() |
[37] | M. Kamenskii, V. Obukhowskii, P. Zecca, Condensing multivalued maps and semilinear differential inclusions in Banach spaces, New York: Walter de Gruyter, 2001. https://doi.org/10.1515/9783110870893 |
[38] |
G. Arthi, Ju H. Park, H. Y. Jung, Exponential stability for second-order neutral stochastic differential equations with impulses, Int. J, Control, 88 (2015), 1300–1309. https://doi.org/10.1080/00207179.2015.1006683 doi: 10.1080/00207179.2015.1006683
![]() |
1. | Mamadou Moustapha Mbaye, Amadou Diop, Gaston Mandata N’Guérékata, Square-mean S-Asymptotically $$\omega $$-Periodic Solutions for Some Stochastic Delayed Integrodifferential Inclusions, 2024, 2305-221X, 10.1007/s10013-024-00709-8 | |
2. | Zainab Alsheekhhussain, Ahmed Gamal Ibrahim, Yousef Jawarneh, Topological Properties of Solution Sets for τ-Fractional Non-Instantaneous Impulsive Semi-Linear Differential Inclusions with Infinite Delay, 2023, 7, 2504-3110, 545, 10.3390/fractalfract7070545 | |
3. | Feryal Aladsani, Ahmed Gamal Ibrahim, Existence and Stability of Solutions for p-Proportional ω-Weighted κ-Hilfer Fractional Differential Inclusions in the Presence of Non-Instantaneous Impulses in Banach Spaces, 2024, 8, 2504-3110, 475, 10.3390/fractalfract8080475 |