Research Article Volume 1 Issue 3
1Department of Mathematics, Sri Ramakrishna Mission Vidyalaya College of Arts and Science, India
2Department of Mathematics,King Abdulaziz University, Saudi Arabia
3Department of Mathematics, Mansoura University, Egypt
Correspondence: Elsayed EM, Department of Mathematics,Faculty of Science, Mansoura University, Mansoura 35516,Egypt
Received: February 03, 2018 | Published: May 1, 2018
Citation: Janaki M, Elsayed EM, Kanagarajan K. Katugampola-type fractional differential equations with delay and impulses. Open Acc J Math Theor Phy. 2018;1(3):73-77. DOI: 10.15406/oajmtp.2018.01.00012
Our aim in this note is to study the existence of solutions of a Katugampola-type fractional impulsive differential equation with delay. We use successive approximation method to show the existence of solutions. In the end, an example is given to verify the hypothetical results.
Keywords: katugampola fractional derivative, impulsive equations, time delay
Because of its wide applicability in biology, medicine and in more and more fields, the theory of fractional differential equations(FDEs) has recently been attracting increasing interest, see for instance1‒8 and references therein. Impulsive differential equations have played an important role in modelling phenomena, especially in describing dynamics of populations subject to abrupt changes as well as other phenomena such as harvesting, diseases, and so forth, some authors have used impulsive differential systems to describe the model since the last century. For the basic theory on impulsive differential equations, the reader can refer to the books9‒12 and the papers.1,13‒16 In addition, some modelling is done via impulsive functional differential equations when these processes involve hereditary phenomena such as biological and social macrosystems. For fractional functional differential equations, the initial value problem, for a class of nonlinear fractional functional differential equations is discussed. For more details, see.17‒24 Motivated by the papers,25,26 the aim of this note is to discuss the existence and uniqueness of solutions of Katugampola-type FDEs with delay and impulses.
Consider the Katugampola-type FDEs with delay and impulse of the form,
(ρcDω0+ℨ(t)=ℌ(t,ℨt), t≠tk; t∈ℑ:=[0,T];Δℨ(tk)=Ik(ℨ(tk)), k=1,2,...,m;ℨ(t)=ψ(t), t∈[−μ,0], (1)
where is the generalized fractional derivative in Caputo sense, and are given functions satisfying some assumptions that will be specified later. and represent the right and left limits of at respectively, and they satisfy that . If , then for any , define by for , here represents the history of the state from time to present time and .
The rest of this paper is organized as follows. In Section 2, we give some notations and recall some concepts and preliminary results. In Section 3, the existence and uniqueness of the problem(1) are obtained by successive approximation method. In Section 4, an example is given to demonstrate the effectiveness of the main results.
In this section, we recollect several definitions of fractional derivatives and integals from the papers27‒30
Definition The fractional (arbitrary) order integral of the function of order is defined byIωaℌ(t)=∫ta(t−s)ω−1Γ(ω)ℌ(s)ds, (2)
where Γ is the gamma function.
Definition For a function given on the interval , the Caputo fractional order derivative of , is defined by
(cDωa+ℌ)(t)=1Γ(n−ω)∫ta(t−s)n−ω−1ℌ(n)(s)ds, (3)
where n=[ω]+1 . Definition 4.3 The generalized left-sided fractional integral of order is defined by(ρIωa+ ℌ)(t)=ρ1−ωΓ(ω)∫ta(tρ−sρ)ω−1sρ−1 ℌ(s)ds, (4)
for t>a , if the integral exists.Definition Thegeneralized fractional derivative, corresponding to the generalized fractional integral (4), is defined for 0≤a<t , by
(ρDωa+ ℌ)(t)=ρω−n+1Γ(n−1)(t1−ρddt)n∫ta(tρ−sρ)n−ω−1sρ−1 ℌ(s)ds, (5)
if the integral exists.Definition The Caputo-type generalized fractional derivative, ρDωa+ ℌ(t) is defined via the above generalized fractional derivative (5) as follows
(ρcDωa+ ℌ)(t)=(ρDωa+[ ℌ(s)−n−1∑k=0 ℌk(a)k!(μ−a)k])(t), (6)
where n=⌈Re(ω)⌉ . Definition The generalized fractional derivative in Caputo sense, corresponding to the generalized fractional integral in Caputo sense (6), is defined for 0≤a<t , by(ρcDωa+ ℌ)(t)=(t1−ρddt)n(ρcIωa+ ℌ)(t)
=ρω−n+1Γ(n−ω)(t1−ρddt)n∫ta(tρ−sρ)n−ω−1sρ−1 ℌ(s)ds. (7)
Remark In Caputo sense, the Katugampola fractional derivative operator ρcDωt is a left inverse of the integral operator ρcIωt but in general is not a right inverse,
ρcDωt(Iωtℨ(t))=ℨ(t)
and the following holdsρcIωt(Iωtℨ(t))=ℨ(t)−n−1∑k=0(tρ−a)kk!ℨ(k)(a), t∈[a,b]. (8)
For readers’ understanding, we introduce the following notations for the following lemma and theorem. LetWe denote exist and . Obviously,is a Banach space with the norm
Lemma 4.8 Assume that A function is a solution of the initial value problem
(ρcDω0+ℨ(t)=ℌ(t), t≠tk, t∈J:=[0,T];Δℨ(tk)=Ik(ℨ(tk)), k=1,2,...,m;ℨ(t)=ψ(t), t∈[−μ,0] (9)
if and only if ℨ satisfies the following integral equation
ℨ (t)=(ψ(t), t∈[−μ,0];ρ1−ωΓ(ω)∫ttk(tρ−sρ)ω−1sρ−1ℌ(s)ds+k∑j=1Ij(ℨ(tj)) +k−1∑i=0ρ1−ωΓ(ω)∫ti+1ti(tρi+1−sρ)ω−1sρ−1ℌ(s)ds, t∈(tk,tk+1], k=0,1,2,...,m. (10)
Proof. Assume that ℨ satisfies (9). One can see, from Remark 2.7 and ψ(0)=0 , thatℨ(t)=ρ1−ωΓ(ω)∫t0(tρ−sρ)ω−1sρ−1ℌ(s)ds, for t∈J0=[t0,t1].
In view of ℨ(t+1)−ℨ(t−1)=I1(ℨ(t1)) , we get thatℨ(t+1)=I1(ℨ(t1))+ρ1−ωΓ(ω)∫t10(tρ1−sρ)ω−1sρ−1ℌ(s)ds.
It follows that, for t∈(t1,t2] ,ℨ(t)=ℨ(t+1)+ρ1−ωΓ(ω)∫tt1(tρ−sρ)ω−1sρ−1ℌ(s)ds
=ρ1−ωΓ(ω)∫tt1(tρ−sρ)ω−1sρ−1ℌ(s)ds+ρ1−ωΓ(ω)∫t10(tρ1−sρ)ω−1sρ−1ℌ(s)ds+I1(ℨ(t1)).
In consequence, we can see, by means of ℨ(t+2)=ℨ(t−2)+I2(ℨ(t2)) , thatℨ(t+2)=1∑i=0ρ1−ωΓ(ω)∫ti+1ti(tρi+1−sρ)ω−1sρ−1ℨ(s)ds+2∑j=1Ij(ℨ(tj)),
which implies that for t∈(t2,t3] ,ℨ(t)=ρ1−ωΓ(ω)∫tt2(tρ−sρ)ω−1sρ−1ℌ (s)ds
ℨ(t)=ρ1−ωΓ(ω)∫tt2(tρ−sρ)ω−1sρ−1ℌ (s)ds
Repeating the above process, the solution ℨ (t) for t∈(tk,tk+1] can be written asℨ (t)=ρ1−ωΓ(ω)∫ttk(tρ−sρ)ω−1sρ−1ℌ(s)ds
+k−1∑i=0ρ1−ωΓ(ω)∫ti+1ti(tρi+1−sρ)ω−1sρ−1ℌ(s)ds+k∑j=1Ij(ℨ(tj)).
Conversely, if ℨ is a solution of (10), one can obtain by a direct computation, that ρcDω0+ℨ(t)=ℌ(t) , t≠tk , t∈[0,T] , and Δℨ(tk)=ℨ(t+k)−ℨ(t−k)=Ik(ℨ(tk)) , whereℨ(t+k)=k−1∑i=0ρ1−ωΓ(ω)∫ti+1ti(tρi+1−sρ)ω−1sρ−1ℌ(s)ds+k∑j=1Ij(ℨ(tj)),
andℨ(t−k)=ρ1−ωΓ(ω)∫tktk−1(tρk−sρ)ω−1sρ−1ℌ(s)ds+k−2∑i=0ρ1−ωΓ(ω)∫ti+1ti(tρi+1−sρ)ω−1sρ−1ℌ(s)ds
+k−1∑j=1Ij(ℨ(tj)).
This completes the proof.Initially, set C0={v|v∈ℭ (J ,ℝ), v(0)=0} . For each,v∈C0 ,we denote by ˉv(t)=v(t), 0≤t≤T and ˉv(t)=0, −μ≤t≤0. the function defined by
ˉv(t)=v(t), 0≤t≤T and ˉv(t)=0, −μ≤t≤0. (11)
If ℨ is a solution of (1), thenℨ(.) can be decomposed as ℨ(t)=ˉv(t)+ϕ(t) for −μ≤t≤T , which implies that ℨt=ˉvt+ϕt for 0≤t≤T , whereϕ(t)=0, 0≤t≤T, and ϕ(t)=ψ(t), −μ≤t≤0. (12)
Therefore, the problem (1) can be transformed into the following fixed point problem of the operator N:ℭ 0→ℝ,Nv(t)=ρ1−ωΓ(ω)∫ttk(tρ−sρ)ω−1sρ−1ℌ(s,ˉvs+ϕs)ds
+k−1∑i=0ρ1−ωΓ(ω)∫ti+1ti(tρi+1−sρ)ω−1sρ−1ℌ(s,ˉvs+ϕs)ds
+k∑j=1Ij(ˉv(tj)), t∈(tk,tk+1], k=0,1,2,...,m. (13)
Now, let us present our main results.Theorem For the functions ℌ∈ℭ(J×ℝ,ℝ) and Ik:ℝ→ℝ , assume the following conditions hold • There exists a continuous function α:[0,T]→ℝ+ satisfying |ℌ(t,pt)−ℌ(t,qt)|≤α(t) sup s∈[0,t]|p(s)−q(s)|, p,q∈ℝ, t∈[0,T];
• There exists a constant Lk>0 such that |Ik(p)−Ik(q)|≤Lk|p−q|, k=1,2,...,m ;
m+1∑i=1αiTρωρωΓ(ω+1)+m∑j=1Lj<1, where αk= sup t∈(tk,tk+1)α(t) ; • There exists a constant M>0 such that |ℌ(t,ϕt)|≤M , where ϕ is defined in (12).
Proof
To complete the proof, we shall use the method of successive approximations. Define a sequence of functions
vn:[0,T]→ℝ ,
n=1,2,...
as follows:
v0(t)=0, vn(t)=Nvn−1(t). (14)
Since v0(t)=0 , it is easy to see from(11) that (ˉv0)s=0 fors∈[0,T] . Thus we have,|v1(t)−v0(t)|≤ρ1−ωΓ(ω)∫ttk(tρ−sρ)ω−1sρ−1|ℌ(s,ϕ(s))|ds+k∑j=1Ij(0)
+k−1∑i=0ρ1−ωΓ(ω)∫ti+1ti(tρi+1−sρ)ω−1sρ−1|ℌ(s,ϕ(s))|ds
≤M(tρω−tρωk)ρωΓ(ω+1)+k∑i=1M(tρωi−tρωi−1)ρωΓ(ω+1)+k∑j=1|Ij(0)|
≤m+1∑i=1M(tρωi−tρωi−1)ρωΓ(ω+1)+k∑j=1|Ij(0)|:=N0, k=1,2,..,m,
it follows that ‖v1(t)−v0(t)‖≤N0. Furthermore,|vn(t)−vn−1(t)|≤ρ1−ωΓ(ω)∫ttk(tρ−sρ)ω−1sρ−1|ℌ(s,(ˉvn−1)s+ϕs)−ℌ(s,(ˉvn−2)s+ϕs)|ds
+k−1∑i=0ρ1−ωΓ(ω)∫ti+1ti(tρi+1−sρ)ω−1sρ−1|ℌ(s,(ˉvn−1)s+ϕs)−ℌ(s,(ˉvn−2)s+ϕs)|ds
+k∑j=1|Ij(ˉvn−1)(tj)−Ij(ˉvn−2)(tj)|
≤ρ1−ωΓ(ω)∫ttk(tρ−sρ)ω−1sρ−1α(s) sup x∈[0,s]|ˉvn−1(x)−ˉvn−2(x)|ds
+k−1∑i=0ρ1−ωΓ(ω)∫ti+1ti(tρi+1−sρ)ω−1sρ−1α(s) sup x∈[0,s]|ˉvn−1(x)−ˉvn−2(x)|ds
+k∑j=1Ij|ˉvn−1(tj)−ˉvn−2(tj)|
≤(αk(tρω−tρωk)Γ(ω+1)+k∑i=1αi(tρωi−tρωi−1)Γ(ω+1)+k∑j=1Lj).‖vn−1−vn−2‖
≤(m+1∑i=1αiTωρρωΓ(ω+1)+m∑j=1Lj).‖vn−1−vn−2‖
:=N1‖vn−1−vn−2‖, (15)
which implies that ‖vn−vn−1‖≤N1‖vn−1−vn−2‖ with N1<1 . Note that for any r>n>0 , we have ‖vr−vn‖≤‖vn+1−vn‖+‖vn+2−vn+1‖+...+‖vr−vr−1‖≤(Nn1+Nn+11+...+Nr−11)‖v1−v0‖
‖vr−vn‖≤Nn11−N1‖v1−v0‖. (16)
for sufficiently large numbersr,n, it follows from the above inequalities with N1<1 that ‖vr−vn‖→0. Thus, {vn(t)} is a Cauchy sequence in Pℭ(J). Since Pℭ(J) is a complete Banach space, then ‖vn−v‖→0 (n→∞), for some v∈Pℭ(J) , which means that vn(t) is uniformly convergent to v(t) with respect to t. In what follows, we shall show that v(t) is a solution of the equation (1). Observe that|ρ1−ωΓ(ω)∫ttk(tρ−sρ)ω−1sρ−1ℌ(s,(ˉvn)s+ϕs)ds−ρ1−ωΓ(ω)∫ttk(tρ−sρ)ω−1sρ−1ℌ(s,(ˉvs+ϕs))ds|
≤ρ1−ωΓ(ω)∫ttk(tρ−sρ)ω−1sρ−1|ℌ(s,(ˉvn)s+ϕs)−ℌ(s,(ˉvs+ϕs))|ds
≤ρ1−ωΓ(ω)∫ttkα(t)(tρ−sρ)ω−1sρ−1 sup x∈[0,s]|ˉvn(x)−ˉv(x)|ds
=ρ1−ωΓ(ω)∫ttkα(t)(tρ−sρ)ω−1sρ−1 sup x∈[0,s]|vn(x)−v(x)|ds.
Since vn(t)→v(t) as n→+∞, for any ε>0, there exists a sufficiently large number n0>0 such that for all n>n0 , we have|vn(x)−v(x)|< min {ρωΓ(ω+1)m∑i=0αiTρωε,1m∑j=1Ljε}.
Therefore,|ρ1−ωΓ(ω)∫ttk(tρ−sρ)ω−1sρ−1ℌ(s,(ˉvn)s+ϕs)ds−ρ1−ωΓ(ω)∫ttk(tρ−sρ)ω−1sρ−1ℌ(s,(ˉvs+ϕs))ds|<ε, (17)
|k−1∑i=0ρ1−ωΓ(ω)∫ti+1ti(tρi+1−sρ)ω−1sρ−1ℌ(s,(ˉvn)s+ϕs)ds
−k−1∑i=0ρ1−ωΓ(ω)∫ti+1ti(tρi+1−sρ)ω−1sρ−1�(s,(ˉvs+φs))ds|
≤k−1∑i=0ρ1−ωΓ(ω)∫ti+1ti(tρi+1−sρ)ω−1sρ−1|ℌ(s,(ˉvn)s+ϕs)−ℌ(s,(ˉvs+ϕs))|ds
≤k−1∑i=0α(ti)(tρi−tρi−1)ωρωΓ(ω+1) sup x∈[0,s]|vn(x)−v(x)|ds<ε. (18)
and|k∑j=1Ij(ˉvn(tj))−k∑j=1Ij(ˉv(tj))|≤k∑j=1Lj|ˉvn(tj)−ˉv(tj)|
=k∑j=1Lj|vn(tj)−v(tj)|<ε. (19)
In consequence, we can see that for a sufficiently large number n>n0 ,|v(t)−Nv(t)|
≤|v(t)−vn+1(t)|+|vn+1(t)−Nvn(t)|+|Nvn(t)−Nv(t)|
≤|v(t)−vn+1(t)|+|vn+1(t)−[ρ1−ωΓ(ω)∫ttk(tρ−sρ)ω−1sρ−1ℌ(s,(ˉvn)s+ϕs)ds
+k−1∑i=0ρ1−ωΓ(ω)∫ti+1ti(tρi+1−sρ)ω−1sρ−1ℌ(s,(ˉvn)s+ϕs)ds+k∑j=1Ij(ˉvn(tj))]|
+|k−1∑i=0ρ1−ωΓ(ω)∫ttk(tρ−sρ)ω−1sρ−1ℌ(s,(ˉvn)s+ϕs)ds−ρ1−ωΓ(ω)∫ttk(tρ−sρ)ω−1sρ−1ℌ(s,ˉvs+ϕs)ds|
+|k−1∑i=0ρ1−ωΓ(ω)∫ti+1ti(tρi+1−sρ)ω−1sρ−1ℌ(s,(ˉvn)s+ϕs)ds
−k−1∑i=0ρ1−ωΓ(ω)∫ti+1ti(tρi+1−sρ)ω−1sρ−1ℌ(s,ˉvs+ϕs)ds|+|k∑j=1Ij(ˉvn(tj))−k∑j=1Ij(ˉv(tj))|.
Thus, in view of the convergence of the two previous and (17)-??, one obtains that |v(t)−Nv(t)|→0 , which implies that v is a solution of (1). Finally, we prove the uniqueness of the solution. Assume that v1,v2:[0,T]→ℝ are two solutions of (1). Note that
|v1(t)−v2(t)|≤ρ1−ωΓ(ω)∫ttk(tρ−sρ)ω−1sρ−1α(s) sup x∈[0,s]|ˉv1(x)−ˉv2(x)|ds
+k−1∑i=0ρ1−ωΓ(ω)∫ti+1ti(tρi+1−sρ)ω−1sρ−1α(s) sup x∈[0,s]|ˉv1(x)−ˉv2(x)|ds
+k∑j=1Ij|ˉv1(tj)−ˉv2(tj)|
≤(p+1∑i=1αiTωρρωΓ(ω+1)+p∑j=1Lj)‖v1−v2‖.
According to the conditions (A3) , the uniqueness of the problem (1) follows immediately, which completes the proof.
Consider the following Katugampola-type fractional impulsive differential equation with delay of the form
(ρcDω0+ℨ(t)=e−t|ℨt|(9+et)(1+|ℨt)|, t∈[0,1], t≠12, 0<ω<1;Δℨ(12)=|ℨ(12−)|3+|ℨ(12−)|;ℨ(t)=ψ(t)=e−t−12, −μ≤t≤0. (20)
Let us take, ω=12 , ρ=1, Γ(ω+1)>410 , μ is a non-negative constant. ℨt(θ)=ℨ(t+θ) for−μ≤θ≤0 and 0≤t≤1 . Set ℌ(t,ℨ)=e−tℨ(9+et)(1+ℨ) , I(ℨ)=ℨ3+ℨ , for (t,ℨ)∈[0,1]×[0,+∞) . Now, we can see that|ℌ(t,pt)−ℌ(t,qt)|=e−t(9+et)|pt−qt|(1+|pt|)(1+|qt|)
≤e−t(9+et)|pt−qt|
≤α(t) sup s∈[0,t]|p(s)−q(s)|, where α(t)=e−t(9+et) and α= sup t∈[0,1]α(t)=110 , so the condition (A1) is satisfied. On the other hand, we get that
|I(p)−I(q)|=3|p−q|(3+p)(3+q)≤13|p−q|, p,q>0,
which satisfies the condition (A1) of Theorem 3.1 with L=13 . By a direct computation, we obtain that m+1∑i=1αiTρωρωΓ(ω+1)+m∑j=1Lj=2101Γ(ω+1)+13<1, and|ℌ(t,ℨt)|=e−t(9+et)|ℨt|(1+|ℨt|)≤e−t9+et≤110, t∈[0,1].
As a result, the equations in (20) satisfy all the hypotheses in Theorem 3.1 which guarantees that (20) has a unique solution.
In this note, the existences of solutions of a Katugampola-type fractional impulsive differential equation with delay were investigated. The successive approximation method was employed to show the existence of solutions. The example reflects the applicability of the proposed method.
None
Author declares that there is no conflict of interest.
©2018 Janaki, et al. This is an open access article distributed under the terms of the, which permits unrestricted use, distribution, and build upon your work non-commercially.