Abstract
In the present work, we consider an impulsive fractional differential equation with a deviated argument in an arbitrary separable Hilbert space H. We obtain an associated integral equation and then consider a sequence of approximate integral equations. The existence and uniqueness of solutions to every approximate integral equation is obtained by using analytic semigroup and Banach fixed point theorem. Next we demonstrate the convergence of the solutions of the approximate integral equations to the solution of the associated integral equation. We study the Faedo–Galerkin approximation of the solution and establish some convergence results. Finally, we consider an example to show the effectiveness of obtained theory.
Similar content being viewed by others
Avoid common mistakes on your manuscript.
Introduction
In recent few decades, researcher has developed great interest in fractional calculus due to its wide applicability in science and engineering. Tools of fractional calculus have been available and applicable to deal with many physical and real world problems such as anomalous diffusion process, traffic flow, nonlinear oscillation of earthquake, real system characterized by power laws, critical phenomena, scale free process, describe viscoelastic materials and many others. The details on the theory and its applications can be found in [1–4] and papers [5–9] and references cited therein.
On the other hand, many real world processes and phenomena which are subjected during their development to short-term external influences can be modeled as impulsive differential equation with fractional order which have been used efficiently in modelling many practical problems. Their duration is negligible compared with the total duration of the entire process and phenomena. Such process is investigated in various fields such as biology, physics, control theory, population dynamics, economics, chemical technology, medicine and so on. For the study for impulsive differential equation, we refer to monograph [10, 11], and papers [12–23] and references given therein.
The purpose of this work is to establish the approximation of the solution to following differential equation with deviated argument in a separable Hilbert space \((H,\Vert \;\cdot \Vert ,(\cdot ,\cdot ))\)
where \(0<q<1\), \(^c D^q_{0^+}\) is the fractional derivative in Caputo sense with single base point 0, \(0=t_0<t_1<\cdots <t_p<t_{p+1}=T_0\) are pre-fixed numbers, \(\Delta x|_{t=t_i}=x(t_i^+)-x(t_i^-)\) and \(x(t_i^+)=\lim _{h\rightarrow 0+}x(t_i+h)\) and \(x(t_i^-)=\lim _{h\rightarrow 0-}x(t_i+h)\) denote the right and left limits of x(t) at \(t=t_i\), respectively. In (1), \(A:D(A)\subset H\rightarrow H\) is a closed, positive definite and self adjoint linear operator with dense domain D(A). We assume that \(-A\) is the infinitesimal generator of an analytic semigroup of bounded linear operators on H. The functions \(f:[0,T_0]\times H^2\rightarrow H\), \(a:H\times [0,T_0]\rightarrow {\mathbb {R}}\), \(I_i:H\rightarrow H\) are appropriate functions to be mentioned later. For more details of differential equation with deviating argument, we refer to papers [24–26] and references given therein.
In the present work, we investigate the Faedo–Galerkin approximations of the solutions for (1)–(3). The Faedo–Galerkin approximations of the solutions in a separable Hilbert space to the following system
has been studied first by Miletta [27] under the assumption that \(-A\) is the infinitesimal generator of an analytic semigroup and the nonlinear function M is Lipschitz continuous on a ball in \(D(A^\alpha )\), \(0<\alpha <1\). Bahuguna and Srivastava [28] has discussed the more general cases. For a nice introduction on existence of an approximate solution and associated study of different problems are broadly talked about in the references [28–34].
The organization of the article is as follows: In Sect. 2, We provide some basic definitions, lemmas and theorems as preliminaries as these are useful for proving our results. In Sect. 3, we prove the existence and uniqueness of the approximate solutions by using analytic semigroup and Banach fixed point theorem. In Sect. 4, we show the convergence of the solution to each of the approximate integral equations with the limiting function which satisfies the associated integral equation and the convergence of the approximate Feado-Galerkin solutions will be shown in Sect. 5. In Sect. 6, we provide an example to illustrate the obtained theory.
Preliminaries and Assumptions
In this segment, some basic definitions, preliminaries, Theorems and Lemmas and assumptions which will be used to prove existence result, is stated.
Throughout the work, we assume that \((H,\Vert \;\cdot \Vert ,<\cdot ,\cdot >)\) is a separable Hilbert space. The symbol \(C([0,T_0];H)\) stands for the Banach space of all the continuous functions from \([0,T_0]\) into H equipped with the norm \(\Vert \;z(t)\Vert _C=\sup _{t\in [0,T_0]}\Vert \;z(t)\Vert _H\) and \(L^p((0,T_0);H)\) stands for Banach space of all Bochner-measurable functions from \((0,T_0)\) to H with the norm
Since \(-A\) is the infinitesimal generator of an analytic semigroup of bounded linear operators \(\{ {\mathcal {T}}(t);t\ge 0\}\). Therefore, there exist constants \(C\ge 1\) and \(\delta \ge 0\) such that \(\Vert \;{\mathcal {T}}(t)\Vert \le Ce^{\delta t}\), \(t\ge 0\). In addition, we note that
where \(M_j\) are some positive constants. Henceforth, without loss of generality, we might accept that \({\mathcal {T}}(t)\) is uniformly bounded by M i.e., \(\Vert \;{\mathcal {T}}(t)\Vert \le M\) and \(0\in \rho (-A)\) i.e., \(-A\) is invertible. This permits us to define the positive fractional power \(A^\alpha \) as closed linear operator with domain \(D(A^\alpha )\subseteq H\) for \(\alpha \in (0,1]\). Moreover, \(D(A^\alpha )\) is dense in H with the norm
Hence, we signify the space \(D(A^\alpha )\) by \(H_\alpha \) endowed with the \(\alpha \)-norm \((\Vert \;\cdot \Vert _\alpha )\). Also, we have that \(H_\kappa \hookrightarrow H_\alpha \) for \(0<\alpha <\kappa \) and therefore, the embedding is continuous. Then, we define \(H_{-\alpha }=(H_\alpha )^*\), for each \(\alpha >0\). The space \(H_{-\alpha }\) stands for the dual space of \(H_\alpha \), is a Banach space with the norm \(\Vert \;z\Vert _{-\alpha }=\Vert \;A^{-\alpha }z\Vert \). For study on the fractional powers of closed linear operators, we refer to book by Pazy [35].
Lemma 2.1
Let \(-A\) be the infinitesimal generator of an analytic semigroup \(\{ {\mathcal {T}}(t):{t\ge 0}\}\) such that \(\Vert \;{\mathcal {T}}(t)\Vert \le M\), for \(t\ge 0\) and \(0\in \rho (-A)\). Then,
-
(i)
For \(0\le \alpha \le 1\), \(H_\alpha \) is a Hilbert space.
-
(ii)
The operator \(A^\alpha {\mathcal {T}}(t)\) is bounded for every \(t>0\) and
$$\begin{aligned} \Vert \;A{\mathcal {T}}(t)\Vert\le & {} M t^{-1}, \end{aligned}$$(7)$$\begin{aligned} \Vert \;A^\alpha {\mathcal {T}}(t)\Vert\le & {} M_\alpha t^{-\alpha }. \end{aligned}$$(8)
Now, we state some basic definitions and properties of fractional calculus.
Definition 2.1
The Riemann–Liouville fractional integral operator J is defined as
where \(F\in L^1((0,T_0);H)\) and \(q>0\) is the order of the fractional integration.
Definition 2.2
The Riemann-Liouville fractional derivative is given as
where \(D^m_t=\frac{d^m}{dt^m},\;F\in L^1((0,T_0);H),\;J^{m-q}_{0+}\in W^{m,1}((0,T_0);H)\).
Definition 2.3
The Caputo fractional derivative is given as
where \(F\in C^{m-1}((0,T_0);H)\cap L^1((0,T_0);H)\).
We denote by \({\mathcal {C}}^\alpha _t=PC([0,t];H_\alpha ),\;t\in (0,T_0]\) the space of all \(H_\alpha \)-valued functions on [0, t] such that x(t) is continuous on \(t\ne t_i\), left continuous at \(t=t_i\) and the right limit \(x(t_i^{+})\) exists for \(i=1,\cdots ,p\). It is clear that \({\mathcal {C}}^\alpha _t\) is a Banach space endowed with the norm
For \(0\le \alpha <1\), we define
where \({\mathcal {L}}>0\) is a appropriate constant to be defined later.
Now, we introduce the following assumptions on A, f, a and \(I_i\;(i=1,\cdots ,p)\):
-
(A1)
A is a closed, densely defined, positive definite and self-adjoint linear operator from \(D(A)\subset H\) into H. We assume that operator A has the pure point spectrum
$$\begin{aligned} 0<\lambda _0\le \lambda _1\le \lambda _2\le \cdots \le \lambda _m\le \cdots , \end{aligned}$$(13)with \(\lambda _m\rightarrow \infty \) as \(m\rightarrow \infty \) and a corresponding complete orthonormal system of eigenfunctions \(\{\phi _j\}\), i.e.,
$$\begin{aligned} A\phi _j=\lambda _j \phi _j,\;\;\text {and}\;\;<\phi _l,\phi _j>=\delta _{lj}, \end{aligned}$$(14)where
$$\begin{aligned} \delta _{lj}={\left\{ \begin{array}{ll}1,\;\;j=l,\\ 0,\;\; \text {otherwise}. \end{array}\right. } \end{aligned}$$ -
(A2)
Let \(W_1\subset Dom(f)\) be an open subset of \({\mathbb {R}}_+\times H_\alpha \times H_{\alpha -1}\), where \(\alpha \in [0,1)\). For \((\tau ,x,y)\in W_1\), there is a neighborhood \(U_1\subset W_1\) of \((\tau ,x,y)\) and positive constants \({\mathcal {L}}_f={\mathcal {L}}_f({\tau ,x,y,U_1})\) such that
$$\begin{aligned} \Vert \;f(t,x_1,y_1)-f(s,x_2,y_2)\Vert \le {\mathcal {L}}_f[|t-s|^{\mu _1}+\Vert \;x_1-x_2\Vert _\alpha +\Vert \;y_1-y_2\Vert _{\alpha -1}],\nonumber \\ \end{aligned}$$(15)for all \((t, x_1, y_1),(s, x_2, y_2)\in U_1\) and \(0<\mu _1\le 1\).
-
(A3)
For each \((x,\tau )\in W_2\), where \(W_2\subset Dom(a)\) is an open subset of \(H_\alpha \times {\mathbb {R}}_+\), there is a neighborhood \(U_2\subset W_2\) of \((x,\tau )\) and positive constant \({\mathcal {L}}_a={\mathcal {L}}_a(x, \tau , U_2)\) such that \(a(\cdot ,\cdot ):H_\alpha \times {\mathbb {R}}_+\rightarrow {\mathbb {R}}_+\), \(a(\cdot , 0)=0\),
$$\begin{aligned} |a(x_1,t_1)-(x_2,t_2)|\le {\mathcal {L}}_a\left[ \Vert \;x_1-x_2\Vert _{\alpha }+|t_1-t_2|^{\mu _2}\right] , \end{aligned}$$(16)for all \((x_1,t_1),(x_2,t_2)\in U_2\), \(0<\mu _2\le 1\).
-
(A4)
All the function \(I_i:H_\alpha \rightarrow H_\alpha , (i=1,\cdots ,p)\) are continuous function such that
-
(i)
\(\Vert \;I_i(u)\Vert _\alpha \le L_i\), for all \(\alpha \in (0,1)\).
-
(ii)
\(\Vert \;I_i(u_1)-I_i(u_2)\Vert _\alpha \le N_i\Vert \;u_1-u_2\Vert _\alpha \), for all \(u_1,u_2\in H_\alpha \).
-
(i)
where \(L_{i}\) and \(N_{i}, i\) = 1, \(\ldots \), p are positive constants.
From [12], we adopt the following thought of solution.
Definition 2.4
A piecewise continuous function \(x:[0,T_0]\rightarrow H\) is said to be a mild solution for the system (1)–(3) if \(x\in {\mathcal {C}}^\alpha _{T_0}\cap {\mathcal {C}}^{\alpha -1}_{T_0}\) and satisfy the following impulsive integral equation
The operator \({\mathcal {S}}_q(t)\) and \({\mathcal {T}}_q(t)\) are defined as follows:
where \(\zeta _q(\xi )=\frac{1}{q}\xi ^{1-1/q}\times \psi _q(\xi ^{-\frac{1}{q}})\) is a a probability density function defined on \((0,\infty )\) i.e., \(\zeta _q(\xi )\ge 0\), \(\int ^\infty _0\zeta _q(\xi )d\xi =1\) and
For more details of probability function and generalized functions, we refer to papers [36–39].
Lemma 2.2
The operator \({\mathcal {S}}_q(t),\;t\ge 0\) and \({\mathcal {T}}_q(t),\;t\ge 0\) are bounded linear operators and satisfy
-
(i)
\(\Vert \;{\mathcal {S}}_q(t)y\Vert \le M\Vert \;y\Vert ,\;\Vert \;{\mathcal {T}}_q(t)y\Vert \le \frac{qM}{\Gamma (1+q)}\Vert y\Vert \;\text {and}\;\Vert \;A^\alpha {\mathcal {T}}_q(t)y\Vert \le \frac{qM_\alpha \Gamma (2-\alpha )t^{-q\alpha }}{\Gamma (1+q(1-\alpha ))}\) \(\Vert \;y\Vert \), for any \(y\in H\).
-
(ii)
The families \(\{ {\mathcal {S}}_q(t):t\ge 0\}\) and \(\{ {\mathcal {T}}_q(t):t\ge 0\}\) are strongly continuous.
-
(iii)
If \({\mathcal {T}}(t)\) is compact, then \({\mathcal {S}}_q(t)\) and \({\mathcal {T}}_q(t)\) are compact operators for any \(t>0\).
Approximate Solutions and Convergence
The existence of approximate solutions for the problem (1) is established in this section.
Let \({\mathcal {H}}_n\) be the finite dimensional subspace of H which is spanned by \(\{\phi _0,\phi _1,\cdots ,\phi _n\}\) and \(P^n:H\rightarrow {\mathcal {H}}_n\) be the corresponding projection operator for \(n=0,1,2,\cdots ,\). We define
by
and
for \(i=1,2,\ldots ,p,\) respectively. We choose T, \(0<T\le T_0\) sufficiently small such that
Now, we consider
By the assumptions \((A2)-(A3)\), we have that f is continuous on [0, T]. Therefore, there exist a constant \(N_f>0\) such that
with
Now, we define the operator \(Q_n\) on \({\mathcal {B}}\) as follows
for \(\;t\in [0,T]\) and \(x\in {\mathcal {B}}\).
Theorem 3.1
Suppose \((A1)-(A4)\) holds and \(u_0\in D(A^\alpha )\), for \(0\le \alpha <1\). Then, there exists a unique fixed point \(x_n\in {\mathcal {C}}^\alpha _{T}\cap {\mathcal {C}}^{\alpha -1}_T\) of the map Q i.e., \(Q_n x_n=x_n\) for each \(n=0,1,2,\cdots ,\) and \(x_n\) satisfies the following approximate integral equation
for \(\;t\in [0,T]\).
Proof
To demonstrate the theorem, we first need to show that \(Q_nx\in {\mathcal {C}}^\alpha _T\cap {\mathcal {C}}^{\alpha -1}_T\). It is clear that \(Q_n:{\mathcal {C}}^\alpha _T\rightarrow {\mathcal {C}}^\alpha _T\). Now, it remains to show that \(Q_n x\in {\mathcal {C}}^{\alpha -1}_T\). For \(x\in {\mathcal {C}}^{\alpha -1}_T\), \(0<\tau <t<T\), then we have
From the first term of above inequality, we have
Also, we have that for each \(z\in H\)
Therefore, we estimate the first term as
where \(K_1=M_1\Vert \;u_0\Vert _{\alpha -1}\). The second integrals is estimated as
where \(K_2=\Vert \;A^{\alpha -2}\Vert M_2 N_f T\). The third integrals is estimated as
where \(K_3=M_1\Vert \;A^{\alpha -2}\Vert N_f\). Similarly, we estimate
where \(K_4=M_1\Vert \;A^{-1}\Vert \sum _{i=1}^p L_i\).
Thus, from the inequality (33)–(36), we obtain that
for a positive suitable constant \({\mathcal {L}}\). Therefore, we conclude that \((Q_nx)\in {\mathcal {C}}^{\alpha -1}_T\). Hence, the \(Q_n: {\mathcal {C}}^{\alpha -1}_T\rightarrow {\mathcal {C}}^{\alpha -1}_T\) is a well defined map.
Next, we prove that \(Q_n:{\mathcal {B}}\rightarrow {\mathcal {B}}\). For \(0\le t\le T\) and \(x\in {\mathcal {B}}\), we get that \(\Vert \;(Q_n x)(t) -u_0\Vert _\alpha \)
Therefore, it gives that \(Q_n({\mathcal {B}})\subset {\mathcal {B}}\). At long last, we will assert that \(Q_n\) is a contraction map. For \(x,\;y\in {\mathcal {B}}\) and \(0\le t\le T\), we get that
We have the following inequalities:
Similarly, we have
Using (40), (41) in (39)and obtain that
From the inequality (25), we get
Therefore, it implies that the map \(Q_n\) is a contraction map and has a unique fixed point \(x_n\in {\mathcal {B}}\) i.e., \(Q_n x_n=x_n\) and \(x_n\) satisfies the approximate integral equation
for \(t\in [0,T]\). \(\square \)
Lemma 3.2
Assume that hypotheses \((A1)-(A4)\) hold. If \(u_0\in D(A^\alpha )\), where \(0< \alpha <1\), then \(x_n(t)\in D(A^\upsilon )\) for all \(t\in (0,T]\) with \(0\le \upsilon <1\). Furthermore, if \(u_0\in D(A)\) then \(x_n(t)\in D(A^\upsilon )\) for all \(t\in [0,T]\) with \(0\le \upsilon <1\).
Proof
From Theorem (3.1), we have that there exists a unique \(x_n\in {\mathcal {B}}\subset {\mathcal {C}}^{\alpha -1}_T\) such that \(x_n\) satisfy the Eq. (30). Theorem 2.6.13 in Pazy [35] implies that \(T(t):H\rightarrow D(A^\upsilon )\) for \(t>0\) and \(0\le \upsilon <1\) and for \(0\le \upsilon \le \eta <1\), \(D(A^\eta )\subseteq D(A^\upsilon )\). It is not difficult to see that H\(\ddot{o}\)lder continuity of \(x_n\) might be made using the similar arguments from Eqs. (33)–(36). Additionally from Theorem 1.2.4 in Pazy [35], we have that \(T(t)y\in D(A)\) if \(y\in D(A)\). The result follows from these facts and the fact that \(D(A)\subseteq D(A^\upsilon )\) for \(0\le \upsilon \le 1\). This finishes the proof of Lemma. \(\square \)
Corollary 3.1
Suppose that the hypotheses \((A1)-(A4)\) hold. If \(u_0\in D(A^\alpha )\) with \(0<\alpha <1\), then for any \(t_0\in (0,T]\), there exists a constant \(U_{t_0}\) such that
for all \(t_0\le t\le T\) independent of n, where \(0<\alpha <\upsilon <\beta \). Furthermore, if \(u_0\in D(A)\), there exist a positive constant \(U_0\) such that \(\Vert A^\upsilon x_n(t)\Vert \le U_{0}\), \(t\in [0,T]\), \(n=1,2,\cdots ,\).
Proof
Let \(u_0\in D(A^\alpha )\). Applying \(A^\upsilon \) on the both the sides of (30) and \(t_0\le t\le T\), we get
Again, for \(0\le t\le T\) and \(u_0\in D(A^\alpha )\), we have
Since, we might displace the first term in (44) by \(M\Vert \;u_0\Vert _\upsilon \).
Moreover, if \(u_0\in D(A)\), then \(u_0\in D(A^\upsilon )\) for \(0\le \upsilon <1\). Therefore, we can effortlessly get the required result. This finishes the proof of Lemma. \(\square \)
Convergence of Solutions
The convergence of the solution \(x_n\in H_\alpha \) of the approximate integral Eq. (30) to a unique solution \(x(\cdot )\) of the Eq. (17) on [0, T] is discussed in this section.
Theorem 4.1
Suppose that \((A1)-(A4)\) are satisfied. If \(u_0\in D(A^\alpha )\), then
for every \(t_0\in (0,T]\).
Proof
For \(0<\alpha <\upsilon \), \(n\ge m\) and \(t\in (0,T]\), we have
Let \(n>m\). Thus, \({\mathcal {H}}_m\subset {\mathcal {H}}_n\). Let \({\mathcal {H}}_m^{\top }\) be the orthogonal complement of \({\mathcal {H}}_m\) for each \(m=0,1,\ldots ,\). Thus, we have \({\mathcal {H}}_n^{\top }\subset {\mathcal {H}}_m^{\top }\). Also, we have \(H={\mathcal {H}}_m\oplus {\mathcal {H}}_m^{\top }={\mathcal {H}}_n\oplus {\mathcal {H}}_n^{\top }\). Let \(y\in H\) be an arbitrary element. Then, \(y=y_m+z_m\) with \(y_m\in {\mathcal {H}}_m\) and \(z_m\in {\mathcal {H}}_m{\top }\). Therefore, we have that \(y_m\in {\mathcal {H}}_m=P^m y\). It is easy to see that \(z_m\in {\mathcal {H}}_m^{\top }\rightarrow z_m=\sum _{i=m+1}^na_i\phi _i+z_m^{\prime }\), where \(z_m^{\prime }\in {\mathcal {H}}_n^{\top }\). Let us take \(y_m^{\prime }=\sum _{i=m+1}^na_i\phi _i\).
Therefore, \(y=y_m+y_m^{\prime }+z_m^{\prime }\) and \(P^n y=y_m +y_m^{\prime }\). Thus,
If, \(y=\sum _{i=1}^\infty a_i\phi _i\). Then, we get \(\Vert y\Vert ^2=\sum _{i=1}^\infty |a_i|^2\).
Since, \(A^{\alpha -\upsilon }\phi _i=\lambda _i^{\alpha -\upsilon }\phi \). Hence, we get
Thus, we have the following estimation
Thus, we obtain
Similarly, we estimate
We choose \(t_{0}^{\prime }\) such that \(0<t_0^{\prime }<t<T\), we have
we estimate the first integral as
The second integral is estimated as
Thus, we have
where
Since \(1-M\sum _{i=1}^pN_i>0\), we have
By Lemma 5.6.7 in [35], we have that there exists a constant \({\mathcal {K}}\) such that
taking supremum over \([t_0,T]\) and letting \(m\rightarrow \infty \), we obtain
As \(t_0^{\prime }\) is arbitrary, therefore the right hand side may be made as small as desired by taking \(t_0^{\prime }\) sufficiently small. This completes the proof of the Theorem. \(\square \)
Proposition 4.2
If \(u_0\in D(A)\), then there exist a Cauchy sequence \(x_n\in {\mathcal {B}}\) on [0, T] i.e.,
as \(m,n\rightarrow \infty \).
Proof
Taking \(t_0=0\) in the proof of Theorem 4.1, we replace the term \((t_0-t_0^{\prime })^{q(1-\alpha )-1}t_0^{\prime }\) by \((1-b_1)^{q(1-\alpha )-1}{t_0^{\prime }}^{1-\alpha }\) in Eq. 52 and the constant \(U_{t_0^{\prime }}\) by the constant \(U_0\) from the Lemma 3.1 and Corollary 3.1. \(\square \)
Theorem 4.3
Suppose that (A1)–(A4) are satisfied and \(u_0\in D(A^\alpha )\). Then, there exists a unique \(x_n\in {\mathcal {B}}\), satisfying
and \(x\in {\mathcal {B}}\), satisfying
such that \(x_n\) converges to x in \({\mathcal {B}}\) i.e., \(x_n\rightarrow x\) as \(n\rightarrow \infty \).
Proof
Let \(u_0\in D(A^\alpha )\). For \(0<t\le T\), \(A^\alpha x_n(t)\rightarrow A^\alpha x(t)\) as \(n\rightarrow \infty \) and \(x(0)=x_n(0)=u_0\) for all n. Also, for \(t\in [0,T]\), we have \(A^\alpha x_n(t)\rightarrow A^\alpha x(t)\) as \(n\rightarrow \infty \) in H. Since \(x_n\in {\mathcal {B}} \), therefore it follows that \(x\in {\mathcal {B}}\) and
Also, we have
as \(n\rightarrow \infty \). For \(0<t_0<t\), we rewrite 30 as
We may estimate the first integral as
Thus, we deduce that
Letting \(n\rightarrow \infty \) in the above inequality, we obtain
Since \(t_0\) is arbitrary, we deduce that x satisfies the integral Eq. (17).
Now, we shall show the uniqueness of the solution to Eq. (17). Let \(x_1\) and \(x_2\) be the two solutions of the (17). We have
By Lemma 5.6.7 in Pazy [35], we obtain that
Also, we have that
From (68) and (69), we deduce that \(u_1=u_2\) on [0, T]. Hence, the theorem is proved. \(\square \)
Faedo–Galerkin Approximations
In this section, we consider the Faedo–Galerkin Approximation of a solution and show the convergence results for such an approximation.
We know that for any \(0< T < T_0\), we have a unique \(x\in {\mathcal {C}}^\alpha _T\) satisfying the following integral equation
for \(0<t<T_0\).
Also, we have a unique solution \(x_n\in {\mathcal {C}}^\alpha _T\) of the approximate integral equation
Applying the projection on above equation, then Faedo–Galerkin approximation is given by \(v_n(t)=P^n x_n(t)\) satisfying
Let solution \(x(\cdot )\) of (70) and \(v_n(\cdot )\) of (72), have the representation
Using (74) in (72) and taking inner product with \(\phi _i\), we obtain a system of fractional order integro-differential equation of the form
where
For the convergence of \(\alpha _i^n\) to \(\alpha _i\), we have the following convergence theorem.
Theorem 5.1
Let us assume that \((A1)--(A4)\) are satisfied and \(u_0\in D(A^\alpha )\). Then there exist a unique function \(v_n\in {\mathcal {B}}\) given as
and \(x\in {\mathcal {B}}\) satisfying
for \(t\in [0,T_0]\), such that \(v_n\rightarrow x\) as \(n\rightarrow \infty \) in \({\mathcal {B}}\) and x satisfies the Eq. (17) on \([0,T_0]\).
The system (75)–(77) determines the \(\alpha _i^n\)’s. It can easily be investigated that
Thus, we conclude that
Theorem 5.2
Let us assume that \((A1)--(A4)\) are satisfied. Then, we have the following results
-
(a)
If \(u_0\in D(A^\alpha )\), then
$$\begin{aligned} \lim _{n\rightarrow \infty }\sup _{t\in [t_0,T_0]} \left[ \sum _{i=0}^n\lambda _i^{2\alpha }(\alpha _i(t)-\alpha _i^n(t))^2\right] =0, \end{aligned}$$(86)for any \(0<t_0\le T_0\).
-
(b)
If \(u_0\in D(A)\), then
$$\begin{aligned} \lim _{n\rightarrow \infty }\sup _{t\in [0,T_0]} \left[ \sum _{i=0}^n\lambda _i^{2\alpha }(\alpha _i(t)-\alpha _i^n(t))^2 \right] =0, \end{aligned}$$(87)
for any \(0\le t\le T_0\).
The statement of this hypothesis takes after from the facts specified above and the following results.
Corollary 5.1
Assume that \((A1)--(A4)\) are satisfied. Then
-
(a)
If \(u_0\in D(A^\alpha )\), then
$$\begin{aligned} \sup _{t\in [t_0,T_0]}\Vert \;v_n(t)-v_m(t)\Vert _{\alpha }\rightarrow 0,\;\;\text {as}\;\;m,n\rightarrow \infty , \end{aligned}$$(88)for any \(0< t_0 \le {T_0}<T_{\max }\).
-
(b)
If \(u_0\in D(A)\), then
$$\begin{aligned} \sup _{t\in [0,T_0]}\Vert \;v_n(t)-v_m(t)\Vert _\alpha \rightarrow 0,\;\;\text {as}\;\;m,n\rightarrow \infty . \end{aligned}$$(89)
Proof
For \(n\ge m\) and \(0\le \alpha <\upsilon \), we get
If \(u_0\in D(A^\alpha )\) then the result in (a) follows from Theorem 4.1, If \(u_0\in D(A)\), (b) follows from Proposition 4.2. \(\square \)
Application
In this section, we present an example to show the feasibility of our abstract result.
Let us consider following fractional differential equation with impulsive conditions in the separable Hilbert space H
where \(0<q<1\), \(\widetilde{H}:{\mathbb {R}}^+\times [0,1]\times {\mathbb {R}}\rightarrow {\mathbb {R}}\) is a nonlinear function which is measurable in u, locally H\(\ddot{o}\)lder continuous in first argument t, locally Lipschitz continuous in w and uniformly in u. The function \(\widetilde{P}\) is given as
here, \(h:{\mathbb {R}}^+\rightarrow {\mathbb {R}}^+\) is assumed to be locally H\(\ddot{o}\)lder continuous in t with \(h(0)=0\) and \({\mathcal {G}}\in C^1([0,1]\times [0,1], {\mathbb {R}})\).
Now, we take \(H=L^2((0,1),{\mathbb {R}})\) and operator A as \(Aw=d^2 w/dx^2\) with domain \(D(A)=H^2(0,1)\cap H^1_0(0,1)\). Let \(\alpha =1/2\), then \(H_{1/2}=D(A^{1/2})=H^1_0(0,1)\) is a Banach space with norm \(\Vert w\Vert _{1/2}:=\Vert A^{1/2} w\Vert \), for \(w\in D(A^{1/2})\) and \(H_{-1/2}=(H^1_0(0,1))^*=H^{-1}(0,1)\equiv H^1(0,1)\) is dual space of the space \(H_{1/2}\).
Now, for each \(u\in (0,1)\), we may consider the function \(f:{\mathbb {R}}^+\times H_{1/2}\times H_{-1/2}\rightarrow H\) defined as
with \(\widetilde{P}:[0,1]\times H_{-1/2}\rightarrow H\) which is defined by
and \({\mathcal {H}}:{\mathbb {R}}^+\times [0,1]\times H_{1/2}\rightarrow H\) fulfills following conditions
where \({\mathcal {Q}}(\cdot , t)\in H\) and \({\mathcal {Q}}\) continuous in its second arguments.
For \(w\in D(A)\) and \(\lambda \in {\mathbb {R}}\) with \(-Aw=\lambda w\), we obtain
and \(\Vert w'\Vert _{L^2}=\lambda \Vert w\Vert _{L^2}\). This gives that \(\lambda >0\). Let \(w(u)=C_1\sin (\sqrt{\lambda }u)+C_2\cos (\sqrt{\lambda }u)\) be the solution of the equation \(-Aw=w''=\lambda w\). We use the boundary condition and get \(C_2=0\) and \(\lambda =\lambda _n=n^2\pi ^2\) for each \(n\in {\mathbb {N}}\). Therefore, we get
and \(<w_n, w_m>=0,\;\;m\ne n\), \(<w_n, w_n>=1\).
For \(w\in D(A)\), there exists a sequence of real numbers \(\{\beta _n\}\) such that
We also have
with \(\sum _{n\in {\mathbb {N}}}\lambda _n(\beta _n)^2<+\infty \). The semigroup \({\mathcal {S}}(t)\) have the following expression as
here, \(\{w_n\}\), \(n=1, 2, \cdots \) denotes the orthogonal set of eigenfunctions of A defined by the (99). Now, we will show that (A2)-(A3) are verified. For (A2), we have that \(\widetilde{P}:[0,1]\times H_{-1/2}\rightarrow H\) defined by
and \(z(u,t)=z(y, h(t)|z(y,t)|)\). Thus, for each \(u\in [0,1]\), we obtain
Since \(z_1, z_2\in H^1(0,1)\). Therefore, applying the Minkowski’s integral inequality and getting
Since we have
Thus, we estimate
Therefore,
The assumption on \(\widetilde{H}\) gives that there exist constants \(B_2>0\) and \(\mu \in (0, 1]\) such that
for all \(t,s\in [0,1]\), \(u\in (0,1)\) and \(w_1,\;w_2\in H^1_0(0,1)\). Therefore, \(f:[0,1]\times H^1_0(0,1)\times H^1(0,1)\rightarrow L^2(0,1)\) defined by \(f=\widetilde{P}+\widetilde{H}\) fulfills the assumption (A2).
Next, we will show that \(a: H^1_0(0,1)\times {\mathbb {R}}^+\rightarrow {\mathbb {R}}^+\) which is defined as \(a(w(u,t), t)=h(t)|w(u,t)|\), fulfill the assumption (A3). For \(t\in [0,1]\)
In the above inequality, we have used the following embedding \(H^1_0(0,1)\subset C[0,1]\). By the H\(\ddot{\mathrm{o}}\)lder continuity of h, we have that there exist \(L_h>0\) and \(\theta _1\in (0,1]\) such that
Furthermore, for \(w_1,w_2\in H^1_0(0,1)\), we have
For \(w_1, w_2\in D(A^{1/2})\), we have
Thus, all the results of this section to obtain the main results can be applied.
For the particular case, we can take following example
where \(L_a=\frac{1}{2}\) \(L_F=3/16\) and \(L_I=1/9\).
References
Miller, K.S., Ross, B.: An Introduction to the Fractional Calculus and Fractional Differential Equations. Wiley, New York (1993)
Samko, S.G., Kilbas, A.A., Marichev, O.I.: Fractional Integrals and Derivatives: Theory and Applications. Gordon and Breach Science Publisher, Yverdon (1993)
Podlubny, I.: Fractional Differential Equations. Mathematics in Science and Engineering, vol. 198. Academic Press, San Diego (1999)
Kilbas, A.A., Srivastava, H.M., Trujillo, J.J.: Theory and Applications of Fractional Differential Equations. Elsevier, Amsterdam (2006)
Farahi, S., Guendouzi, T.: Approximate controllability of fractional neutral stochastic evolution equations with nonlocal conditions. Results Math. 2014, 21 (2014)
Zhou, Y., Jiao, F.: Nonlocal Cauchy problem for fractional evolution equations. Nonlinear Anal. Real World Appl. 11, 4465–4475 (2010)
Shu, X.-B., Lai, Y., Chen, Y.: The existence of mild solutions for impulsive fractional partial differential equations. Nonlinear Anal. 74, 2003–2011 (2011)
El-Borai, M., Abujabal, Hamza, A.S.: On the Cauchy problem for some abstract nonlinear differential equations. Korean J. Comput. Appl. Math. 3, 279–290 (1996)
El-Borai, M.: Some probability densities and fundamental solutions of fractional evolution equations. Chaos Solitons Fractals. 14, 433–440 (2002)
Lakshmikantham, V., Bainov, D., Simeonov, P.S.: Theory of Impulsive Differential Equations. Series in Modern Applied Mathematics. World Scientic, Teaneck (1989)
Benchohra, M., Henderson, J., Ntouyas, S.K.: Impulsive Differential Equations and Inclusions. Contemporary Mathematics and Its Applications, vol. 2. Hindawi Publishing Corporation, New York (2006)
Wang, J.-R., Fečkan, M., Zhou, Y.: On the new concept of solution and existence results for impulsive fractional evolution equations. Dyn. PDE 8, 345–361 (2011)
Fečkan, M., Zhou, Y., Wang, J.-R.: On the concept and existence of solution for impulsive fractional differential equations. Commun. Nonlinear Sci. Numer. Simul. 17, 3050–3060 (2012)
Dabas, J., Chauhan, A.: Existence and uniqueness of mild solutions for an impulsive neutral fractional integro-differential equation with infinite delay. Math. Comput. Model. 57, 754–763 (2013)
Wang, G., Ahmad, B., Zhang, L., Nieto, J.J.: Comments on the concept of existence of solution for impulsive fractional differential equations. Commun. Nonlinear Sci. Numer. Simul. 19, 401–403 (2014)
Fečkan, M., Zhou, Y., Wang, J.-R.: Response to Comments on the concept of existence of solution for impulsive fractional differential equations [Commun Nonlinear Sci Numer Simul 2014;19:401–403.]. Commun. Nonlinear Sci. Numer. Simul. 19, 4213–4215 (2014)
Liang, S., Mei, R.: Existence of mild solutions for fractional impulsive neutral evolution equations with nonlocal conditions. Adv. Differ. Equ. 2014, 16 (2014)
Chadha, A., Pandey, D.N.: Existence results for an impulsive neutral fractional integro-differential equation with infinite delay. Int. J. Differ. Equ. 2014, 10 (2014)
Chadha, A., Pandey, D.N.: Existence of a mild solution for an impulsive neutral fractional integro-differential equation with nonlocal conditions. J. Fract. Calc. Appl. 6, 5–20 (2015)
Mardanov, M.J., Mahmudov, N.I., Sharifov, Y.A.: Existence and uniqueness theorems for impulsive fractional differential equations with the two-point and integral boundary conditions. Sci. World J. 2014, 8 (2014)
Xie, S.: Existence results of mild solutions for impulsive fractional integro-differential evolution equations with infinite delay. Fract. Calc. Appl. Anal. 17, 1158–1174 (2014)
Ge, H., Xin, J.: On the existence of a mild solution for impulsive hybrid fractional differential equations. Adv. Differ. Equ. 2014, 14 (2014)
Bouzaroura, A., Mazouzi, S.: Existence results for certain multi-orders impulsive fractional boundary value problem. Results. Math. 2014, 20 (2014)
Gal, C.G.: Nonlinear abstract differential equations with deviated argument. J. Math. Anal. Appl. 333, 971–983 (2007)
Gal, C.G.: Semilinear abstract differential equations with deviated argument. Int. J. Evol. Equ. 2, 381–386 (2008)
Muslim, M., Bahuguna, D.: Existence of solutions to neutral differential equations with deviated argument. Electron. J. Qual. Theory Differ. Equ. 27, 1–12 (2008)
Miletta, P.D.: Approximation of solutions to evolution equations. Math. Methods Appl. Sci. 17, 753–763 (1994)
Bahuguna, D., Srivastava, S.K.: Approximation of solutions to evolution integro-differential equations. J. Appl. Math. Stoch. Anal. 9, 315–322 (1996)
Bahuguna, D., Srivastava, S.K., Singh, S.: Approximations of solutions to semilinear integro-differential equations. Numer. Funct. Anal. Optim. 22, 487–504 (2001)
Bahuguna, D., Shukla, R.: Approximations of solutions to nonlinear Sobolev type evolution equations. Electron. J. Differ. Equ. 31, 1–16 (2003)
Bahuguna, D., Muslim, M.: Approximation of solutions to retarded differential equations with applications to populations dynamics. J. Appl. Math. Stoch. Anal. 2005, 1–11 (2005)
Muslim, M., Agarwal, R.P., Nandakumaran, A.K.: Existence, uniqueness and convergence of approximate solutions of impulsive neutral differential equations. Funct. Differ. Equ. 16, 529–544 (2009)
Muslim, M., Nandakumaran, A.K.: Existence and approximations of solutions to some fractional order functional integral equations. J. Int. Equ. Appl. 22, 95–114 (2010)
Kumar, P., Pandey, D.N., Bahuguna, D.: Approximations of solutions to a fractional differential equations with a deviating argument. Differ. Equ. Dyn. Syst. 2013, 20 (2013)
Pazy, A.: Semigroups of Linear Operators and Aplications to Partial Differential Equations. Springer, New York (1983)
Gelfand, I.M., Shilov, G.E.: Generalized Functions, vol. 1. Nauka, Moscow (1959)
Mainardi, F.: On the initial value problem for the fractional diffusion-wave equation: waves and stability in continuous media (Bologna, 1993). Ser. Adv. Math. Appl. Sci. 23, 246–251 (1994)
Mainardi, F.: On a Special Function Arising in the Time Fractional Diffusion-Wave Equation: Transform Methods and Special Functions, 171st edn. Science Culture Technology, Singapore (1994)
Pollard, H.: The representation of e\(^{-x\lambda }\) 604 as a Laplace integral. Bull. Am. Math. Soc. 52, 908–910 (1946)
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Chaddha, A., Pandey, D.N. Approximations of Solutions for an Impulsive Fractional Differential Equation with a Deviated Argument. Int. J. Appl. Comput. Math 2, 269–289 (2016). https://doi.org/10.1007/s40819-015-0059-1
Published:
Issue Date:
DOI: https://doi.org/10.1007/s40819-015-0059-1
Keywords
- Analytic semigroup
- Banach fixed point theorem
- Caputo derivative
- Impulsive differential equation
- Faedo–Galerkin approximation