Mediterr. J. Math.

13 (2016), 2539–2550
DOI 10.1007/s00009-015-0638-8
1660-5446/16/052539-12
published online October 22, 2015
c Springer Basel 2015

Approximate Controllability of Semilinear

Fractional Control Systems of Order
α ∈ (1, 2] with Infinite Delay
A. Shukla, N. Sukavanam and D. N. Pandey

Abstract. The objective of this paper is to present some sufficient con-

ditions for approximate controllability of semilinear fractional control
system of order α ∈ (1, 2] with infinite delay. The results are obtained
by the theory of strongly continuous α-order cosine family and sequence
method. At the end, an example is given to illustrate the theory.
Mathematics Subject Classification. 34A08, 34K35, 93B05.
Keywords. Fractional derivatives and integrals, approximate controlla-
bility, α-order Cosine family, Sequence method, infinite delay.

1. Introduction
Fractional order systems are generalization of the integer order systems. The
potential applications of fractional calculus are in diffusion process, electri-
cal science, electrochemistry, viscoelasticity, control science, electro magnetic
theory and several more. As delay can be observed in many real-life prob-
lems, fractional order delay differential systems seem natural to model the
real-world systems. Basics of fractional calculus and applications in control
theory can be see in [25,26].
In setting of first-order systems, Kalman [7] introduced the concept
of controllability for finite-dimensional deterministic linear control systems.
The basic concepts of control theory in finite and infinite dimensional spaces
has been introduced in [3] and [4], respectively. In [1] Balachandran et al.
presented a survey for controllability of nonlinear control systems in Banach
spaces using Schauder’s fixed point. In [6,13] Park et al. obtained sufficient
conditions for approximate controllability of delay systems using fixed point
theorems. Recently, Anurag et al. [17] obtained some sufficient conditions for
approximate controllability of retarded semilinear stochastic system with non
local conditions using Banach fixed point theorem. In all discussed papers,
authors used fixed point theorems for obtaining the sufficient conditions for
2540 A. Shukla et al. MJOM

controllability. In [23] Zhou used a different approach, we call it sequential

approach, using this approach Zhou obtained some sufficient conditions for
approximate controllability for a class of semilinear abstract equations.
In setting of second-order systems,; it is advantageous to treat the
second-order abstract differential equations directly rather than to convert
them to first-order systems. A useful tool for the study of second-order equa-
tions is the theory of strongly continuous cosine families. We will make use
of some of the basic ideas from cosine family theory [20]. Motivation for the
second-order systems can be found in [2,5]. Recently, Kumar et al. [10] es-
tablished sufficient conditions for controllability of second-order systems with
nonlocal conditions in Banach spaces using Sadovskii’s fixed point theorem
and strongly continuous cosine family.
In setting of fractional order α ∈ (0, 1] systems, in [21,22,24] authors ob-
tained sufficient conditions for existence of mild solutions for fractional-order
evolution equations. Using existence results for mild solutions, in [8,15,16,21]
authors established sufficient conditions for the approximate controllability of
a class of fractional-order semilinear control systems using contraction prin-
ciple and fixed point theorems. Recently, Kumar et al. [9] used the sequence
method as introduced by [23] and obtained sufficient conditions for approx-
imate controllability of fractional-order α ∈ (0, 1] control system with finite
delay in state.
In setting of fractional-order α ∈ (1, 2], there are not so many papers
on existence of mild solution and controllability of fractional order α ∈ (1, 2]
semilinear systems. Recently, Li et al. [11] obtained sufficient conditions for
the existence of mild solution and exact controllability of fractional differen-
tial systems with nonlocal conditions in infinite-dimensional Banach spaces
using Sadovskii fixed point theorem.
Up to now, to the best of our knowledge, there are no results on the
approximate controllability of fractional semilinear control system of order
α ∈ (1, 2] in Banach spaces. In [11] the author obtained sufficient condition
for existence of mild solution for fractional-order α ∈ (1, 2] systems. Approx-
imate controllability results can been obtained using fixed point theorems
or sequence method. So, the present paper is devoted to the study of ap-
proximate controllability of the fractional semilinear control system of order
α ∈ (1, 2] in Banach spaces with infinite delay in state using sequence method.

2. Problem Formulation
We consider the following semilinear fractional-order control system with
delay of the following form:
C
Dtα x(t) = [Ax(t) + Bu(t) + f (t, xt )], 0 < t ≤ T;
x(t) = x0 = φ ∈ B, t ∈ (−∞, 0]
x (0) = ψ ∈ Y, (2.1)
where α ∈ (1, 2], C Dtα is the Caputo fractional derivative and the positive
constant T < ∞. Here, the state x(·) takes values in a Banach space Y and
Vol. 13 (2016) Approximate Controllability of Semilinear 2541

the control function u(·) is given in L2 ([0, T ]; U ), a Banach space of admissible

control functions, with U a Banach space, A is the infinitesimal generator of
strongly continuous α-order cosine family {Cα (t)}t≥0 on the Banach space Y .
B : U → Y is a bounded linear operator. The history function xt : (−∞, 0] →
Y is defined by xt (θ) = x(t + θ), for t ≥ 0 belongs to some abstract phase
space B described axiomatically. The map f : [0, T ] × B → Y is a nonlinear
operator.

3. Preliminaries
We will use an axiomatic definition for the phase space B which is similar to
the one used in [5]. Specifically, B will be a linear space of functions mapping
(−∞, 0] into Y endowed with a seminorm || · ||B , which satisfies the following
axioms:
(A) If x : (−∞, σ+b] → Y , b > 0, is such that x|[σ,σ+b] ∈ C([σ, σ+b] : Y )
and xσ ∈ B, then for every t ∈ [σ, σ + b] the following conditions hold:
(1) xt is in B,
(2) ||x(t)|| ≤ H||xt ||B
(3) ||xt ||B ≤ K(t − σ) sup{||x(s)|| : σ ≤ s ≤ t} + M (t − σ)||xσ ||B ,
where H > 0 is a constant. K, M : [0, ∞) → [1, ∞), K is continuous, M is
locally bounded, and H, K, M are independent of x(·). (A1) For the function
x(·) in (A), the function t → xt is continuous from [σ, σ + b] into B. (B) The
space B is complete.
Let Z be the closed subspace of all continuous functions x from (−∞, T ]
to the space L2 ((−∞, T ]; B) and the restriction x : [0, T ] → L2 ((−∞, T ]; B)
is continuous. Let || · ||Z be a seminorm in Z defined by
||x||Z = sup ||xt ||B ,
t∈[0,T ]

where ||xt || ≤ K||φ||B + M sups∈[0,T ] {||x(s)||}, K = supt∈[0,T ] {K(t)} and

M = supt∈[0,T ] {M (t)}.
Let Y be the Banach space with norm || · || and L(Y ) denote the space
of all bounded linear operators on Y . Let Lp ([0, T ]; Y ), 1 ≤ p < ∞ denote
the space of Y -valued Bochner integrable functions f : [0, T ] → Y with the
norm
⎛ T ⎞1/p

||f ||Lp ([0,T ];Y ) = ⎝ ||f (t)||p dt⎠ (3.1)
0
1
C([0, T ]; Y ) and C ([0, T ]; Y ), we denote the space of functions which are
continuous and 1-time continuously differentiable, respectively.
Definition 3.1. If x(t) ∈ L1 ([0, T ]; Y ), then the Riemann–Liouville fractional
integral of order α > 0 is defined by
t
1
Jtα x(t) = (t − s)α−1 x(s)ds,
Γ(α)
t0
2542 A. Shukla et al. MJOM
∞
where Γ(·) is the gamma function, defined as Γ(α) = 0
e−t tα−1 dt.
Definition 3.2. The Riemann–Liouville fractional derivative of a function
x(t) ∈ L1 ([0, T ]; Y ) of order α ∈ (1, 2) is defined by
t
1 d2
Dtα x(t) =D J 2 2−α
x(t) = (t − s)1−α x(s)ds.
Γ(2 − α) dt2
t0

Definition 3.3. The Caputo fractional derivative of order α ∈ (1, 2) is defined

by
t 
C 1 d2
Dtα x(t) =J 2−α 2
D x(t) = (t − s)1−α x(s) ds,
Γ(2 − α) ds2
t0

1
where x(t) ∈ L1 ([0, T ]; Y ) C ([0, T ]; Y ).
Consider the following linear fractional order system:
C
Dtα x(t) = Ax(t), x(0) = ξ, x (0) = 0, (3.2)
where α ∈ (1, 2], A : D(A) ⊂ Y → Y is closed and densely defined operator
in Banach space Y . Applying Riemann–Liouville fractional integral of order
α on both sides of (3.1), we get
t
1
x(t) = ξ + (t − s)α−1 Ax(s)ds (3.3)
Γ(α)
0

Definition 3.4. Let α ∈ (1, 2]. A family {Cα (t)}t≥0 ⊂ L(Y ) is called the
solution operator (or strongly continuous α-order fractional cosine family)
for (3.1) and A is called the infinitesimal generator of Cα (t), if the following
conditions are satisfied:
(a) Cα (t) is strongly continuous for t ≥ 0 and Cα (0) = I;
(b) Cα (t)D(A) ⊂ D(A) and ACα (t)ξ = Cα (t)Aξ for all ξ ∈ D(A), t ≥ 0;
(c) Cα (t)ξ is a solution of (3.1) for all ξ ∈ D(A).
Definition 3.5. The fractional sine family Sα : [0, ∞) → L(Y ) associated with
Cα is defined by
t
Sα (t) = Cα (s)ds, t ≥ 0. (3.4)
0

Definition 3.6. The fractional Riemann–Liouville family Pα : [0, ∞) → L(Y )

associated with Cα is defined by
Pα (t) = J α−1 Cα (t) (3.5)
Definition 3.7. The α-order cosine family Cα (t) is called exponentially
bounded if there are constants M ≥ 1 and ω ≥ 0 such that
||Cα (t)|| ≤ M eωt , t ≥ 0. (3.6)
Vol. 13 (2016) Approximate Controllability of Semilinear 2543

An operator A is said to belong to C α (Y ; M, ω), if the problem (3.1)

has a solution operator Cα (t) satisfying (3.5).
Definition 3.8 [11,24]. A function xt (·) ∈ C((−∞, T ]; Y ) is said to be the
mild solution of (1.1) if it satisfies
t
xt (Bu) = Cα (t)φ + Sα (t)ψ + Pα (t − s){Bu(s) + f (s, xs )}ds, t ∈ (0, T ]
0
(3.7)
Definition 3.9. The set defined by
KT (f ) = {x(T ; φ, ψ, u); u ∈ U }; 0 ≤ t ≤ T, (3.8)
is called the reachable set which consists of all possible final states.
Definition 3.10. A control system is said to be approximately controllable on
[0, T ], iff
KT (f ) = Y,
where KT (f ) denotes the closure of KT (f )
Define a solution mapping ϕ from L2 ([0, T ]; Y ) to C([0, T ]; Y ) as
(ϕBu)(t) = xt (Bu)
Let x(T ; φ, ψ, u) be the state value of the system (2.1) at time t ∈ [0, T ],
corresponding to the control u ∈ U and the initial conditions (φ, ψ). The
system (2.1) is said to be approximately controllable in the time interval
[0, T ], if for every desired final state xF and
> 0, there exists a control
function u ∈ U such that the solution of (2.1) satisfies
||x(T ; φ, ψ, u) − xF || <

Define a continuous linear operator L from L2 ([0, T ]; Y ) to C([0, T ]; Y ) by

T
Lp = Pα (T − s)p(s)ds, f or p(·) ∈ L2 ([0, T ]; Y ).
0
We need the following hypothesis to prove our results:
(H1) The nonlinear function f (t, x) : [0, T ]×B → Y satisfies the Caratheodory
condition and there exists a positive constant L such that ||f (t, x) −
f (t, y)||Y ≤ L||x − y||B , ∀ x, y ∈ B and t ∈ [0, T ].
(H2) For any given
> 0 and p(·) ∈ L2 ([0, T ]; Y ), there exists some u(·) ∈ U
such that
||Lp − LBu||Y <

(H3) ||Bu(·)||L2 ([0,T ];Y ) ≤ λ||p(·)||L2 ([0,T ];Y ) where λ is a positive constant
independent of p(·).
2544 A. Shukla et al. MJOM
 
MT α T αL
(H4) The constant λ satisfies; Γ(α) λL exp M Γ(α) <1

4. Controllability Results
In this section, the approximate controllability for a class of second-order
semilinear control system (2.1) with infinite delay is proved.

Lemma 1. Assume hypothesis (H2 ); then linear system corresponding to sys-

tem (2.1) ( put f (t, xt ) = 0 in (2.1)) is approximately controllable.

Proof. Since the domain D(A) of the operator A is dense in Y , it is sufficient

to prove that D(A) ⊂ KT (0), that is , for any given
> 0 and ζ ∈ D(A),
there exists a control function u(·) ∈ L2 ([0, T ]; U ) such that
||ζ − Cα (T )φ − Sα (T )ψ − LBu|| <
(4.1)
Let us take ζ ∈ D(A); then (ζ − Cα (T )φ − Sα (T )ψ) ∈ D(A). It can be seen
that there exists some p ∈ C 1 ([0, T ]; Y ) such that
T
χ= Pα (T − s)p(s)ds,
0

where χ = ζ − Cα (T )φ − Sα (T )ψ. The assumption (H2 ) implies that for any

given
> 0, there exists a control function u(·) ∈ L2 ([0, T ]; U ) such that
||χ − LBu|| <
.
After putting the value of χ, we obtain (4.1). Since
is arbitrary, we infer
that KT (0) ⊂ D(A). The denseness of the domain D(A) in Y implies the
approximate controllability of linear system corresponding to (2.1). 

Lemma 2. Let u1 (·) and u2 (·) be in L2 ([0, T ]; U ). Then under hypothesis (H1 )
the solution mapping ϕ(Bu)(t) = xt (Bu) of (2.1) satisfies
 
M T α−1/2 M LT α
||xt (Bu1 ) − xt (Bu2 )||Y ≤ exp ||Bu1 − Bu2 ||L2 ([0,T ];U )
Γ(α) Γ(α)

Proof. The solution mapping ϕ(Bu)(t) = xt (Bu) is given by

t
xt (Bu) = Cα (t)φ + Sα (t)ψ + Pα (t − s)[Bu(s) + f (s, xs )]ds
0

Let us define y(·, φ, ψ) : [−∞, T ] → Y as

y(t, φ, ψ) = Cα (t)φ + S(t)ψ, f or t ∈ [0, T ]
y(t, φ, ψ) = φ, f or t ∈ [−∞, 0]
y  (0, φ, ψ) = ψ
Vol. 13 (2016) Approximate Controllability of Semilinear 2545

Let xt = yt + zt , t ∈ (−∞, T ]. It is easy to see that x(·) satisfies (2.1) if and

only if
z(0) = 0,z  (0) = 0 and for all t ∈ [0, T ], we have
t
z(t) = Pα (t − s)[Bu(s) + f (s, ys + zs (Bu))]ds.
0

Now, let us take xt (Bu1 ), xt (Bu2 ) ∈ B and u1 (·), u2 (·) ∈ U ; then

||xt (Bu1 ) − xt (Bu2 )|| = ||[zt (Bu1 ) + yt ] − [zt (Bu2 ) + yt ]|| = ||zt (Bu1 ) − zt (Bu2 )
t
T α−1
≤M ||Bu1 (s) − Bu2 (s)||ds
Γ(α)
0

α−1 t
T
+M ||f (s, ys + zs (Bu1 )) − f (s, ys + zs (Bu2 ))||ds
Γ(α)
0
t
T α−1 √ T α−1
≤M T ||Bu1 − Bu2 ||L2 ([0,T ];U ) + M L ||zs (Bu1 ) − zs (Bu2 )||B ds
Γ(α) Γ(α)
0

Using Gronwall’s inequality, we get

||xt (Bu1 ) − xt (Bu2 )||Z = ||zt (Bu1 ) − zt (Bu2 )||C
 
M T α−1/2 M LT α
≤ exp ||Bu1 − Bu2 ||L2 ([0,T ];U )
Γ(α) Γ(α)
This completes the proof of lemma. 
Theorem 4.1. Under the hypothesis (H1 ) − (H4 ) the fractional order semilin-
ear control system (2.1) is approximately controllable.

Proof. Since by the Lemma (2), KT (0) = Y , to prove the approximate con-
trollability of the fractional order system (2.1), it is sufficient to show that
KT (0) ⊂ KT (f ). For this we prove that for any given
> 0, and ζ ∈ KT (0),
there exists a control function u(·) ∈ L2 ([0, T ]; U ) such that

||ζ − Cα (T )φ − Sα (T )ψ − LBu|| < 3 , (4.2)

2
Now, we construct a sequence recursively as follows: let u1 (·) ∈ U be arbi-
trary. Then by hypothesis (H2 ), there exists some u2 (·) ∈ U such that

||L{Bu − f (s, xs (Bu1 ))} − LBu2 || < 3 , (4.3)

2
where xt (Bu1 ) = ϕ(Bu1 )(t), for all t ∈ [0, T ]. From (4.2) and (4.3), we have

||ζ − Cα (T )φ − Sα (T )ψ − Lf (s, xs (Bu1 )) − LBu2 || < 2 (4.4)

2
For u2 (·) ∈ L2 ([0, T ]; U ) thus obtained, we determine w2 (·) ∈ L2 ([0, T ]; U )
by hypothesis (H2 ) such that

||L[f (s, xs (Bu2 )) − f (s, xs (Bu1 ))] − LBw2 || < 3 ; (4.5)

2
2546 A. Shukla et al. MJOM

also by hypothesis (H3 ), we have

||Bw2 (·)||L2 ([0,T ];U ) ≤ λ||f (s, xs (Bu2 )) − f (s, xs (Bu1 ))||L2 ([0,T ];U )
 T 1/2
≤λ ||f (s, xs (Bu2 )) − f (s, xs (Bu1 ))||2 ds
0
 T 1/2
≤ λL ||xs (Bu2 ) − xs (Bu1 )||2B ds
0
√
≤ λL T ||xs (Bu2 ) − xs (Bu1 )||Z
 
MTα T αL
≤ λL exp M ||Bu1 − Bu2 ||L2 ([0,T ];U ) ,
Γ(α) Γ(α)
where xt (Bun ) = ϕ(Bun )(t), n = 1, 2 for all t ∈ [0, T ]. Now, we may define
u3 (·) = u2 (·) − w2 (·) in U which has the following properties:
||ζ − Cα (T )φ − Sα (T )ψ − Lf (s, xs (Bu2 )) − LBu3 ||


≤ ζ − Cα (T )φ − Sα (T )y0 − Lf (s, xs (Bu1 )) − LBu2


+ LBw2 − L[f (s, xs (Bu2 )) − f (s, xs (Bu1 ))]

≤ ||ζ − Cα (T )φ − Sα (T )ψ − Lf (s, xs (Bu1 )) − LBu2 ||

+ ||LBw2 − L[f (s, xs (Bu2 )) − f (s, xs (Bu1 ))]||
Using Eqs. (4.4) and (4.5), we get
 
1 1
||ζ − Cα (T )φ − Sα (T )ψ − Lf (s, xs (Bu2 )) − LBu3 || ≤ + 3

22 2
Mathematical induction implies that there exists a sequence un (·) ∈ U such
that
||ζ − Cα (T )φ − Sα (T )ψ − Lf (s, xs (Bun )) − LBun+1 ||
 
1 1
≤ + · · · + n+1
, (4.6)
22 2
where xt (Bun ) = ϕ(Bun )(t), n = 1, 2, ..., for all t ∈ [0, T ] and
||Bun+1 (·) − Bun (·)||L2 ([0,T ];U )
 
MTα T αL
≤ λLexp M ||Bu1 − Bu2 ||L2 ([0,T ];U )
Γ(α) Γ(α)
Clearly, by hypothesis (H4 ) the sequence {Bun ; n = 1, 2, ...} is a Cauchy
sequence in the Banach space L2 ([0, T ]; U ) and there exists some v(·) ∈
L2 ([0, T ]; U ) such that
lim Bun (t) = v(t), in L2 ([0, T ]; U )
n→∞
Thus for any given
> 0, there exists some positive integer N such that

||LBuN +1 − LBuN || < (4.7)

2
Vol. 13 (2016) Approximate Controllability of Semilinear 2547

Hence, we obtain
||ζ − Cα (T )φ − Sα (T )ψ − Lf (s, xs (BuN )) − LBuN ||
≤ ||ζ − Cα (T )φ − Sα (T )ψ − Lf (s, xs (BuN )) − LBuN +1 ||
+||LBuN +1 − LBuN ||,
where xt (BuN ) = ϕ(BuN )(t), for all t ∈ [0, T ]. Using inequality (4.6) and
(4.7), we get
||ζ − Cα (T )φ − Sα (T )ψ − Lf (s, xs (BuN )) − LBuN ||
 
1 1

≤ 2
+ · · · + N
+ ≤

2 2 2

This means that ζ ∈ KT (f ). Thus KT (0) ⊂ KT (f ); hence the fractional

order semilinear system (2.1) is approximately controllable on [0, T ]. This
completes the proof. 

Theorem 4.2. Suppose that the range of the operator B, i.e. R(B) is dense
in L2 ([0, b]; Y ). Then under hypothesis (H1 ) the semilinear system (1.1) is
approximately controllable.

Proof. Since the range of the operator B is dense in L2 ([0, T ]; Y ), for any
given point p(·) ∈ L2 ([0, T ]; Y ) and every δ > 0, there exists some point
Bu(·) ∈ R(B), where u(·) ∈ U such that
||Bu(·) − p(·)||L2 ([0,b];Y ) < δ||p(·)||L2 ([0,b];Y ) (4.8)
Now, we have
T
M T α−1
||Lp − LBu|| ≤ ||p(s) − Bu(s)||ds
Γ(α)
0
α−1/2
MT
≤ ||p(·) − Bu(·)||L2 ([0,T ];Y )
Γ(α)
M T α−1/2
≤ δ||p(·)||L2 ([0,T ];Y )
Γ(α)
<

Thus from (4.8), we have

||Bu(·)||L2 ([0,T ];Y ) = ||Bu(·) − p(·) + p(·)||L2 ([0,T ];Y )
≤ ||Bu(·) − p(·)||L2 ([0,T ];Y ) + ||p(·)||L2 ([0,T ];Y )
≤ δ||p(·)||L2 ([0,T ];Y ) + ||p(·)||L2 ([0,T ];Y )
≤ (δ + 1)||p(·)||L2 ([0,T ];Y )
This implies that the conditions (H2 ) and (H3 ) are satisfied if we choose δ > 0
in such a manner that (H4 ) is verified. Then the approximate controllability
of (2.1) follows from Theorem (4.1) 
2548 A. Shukla et al. MJOM

5. Example
Consider the control system governed by the following partial differential
equation:
C
Dtα y(t, x) = yxx (t, x) + μ(t, x) + f (t, y(t + v)); t ∈ [0, T ]
y(t, 0) = y(t, π) = 0; f or t ∈ [0, T ]
y(t, x) = φ(t, x); f or t ∈ (−∞, 0]; and v ∈ (−∞, 0]; (5.1)
∂y
(0, x) = ψ(x); f or t ∈ (0, T ), 0 < x < π.
∂t
where α ∈ (1, 2], one can take arbitrary nonlinear function f satisfying the
conditions (H1 ) and (H4 ). Let Y = L2 [0, π] and C = C((−∞, 0]; Y ] be as in
introduction. Let control function μ : [0, T ] × (0, π) → R be continuous in t.
Define A : D(A) ⊆ Y → Y by
Aw = w ; w ∈ D(A),
where D(A) = {w ∈ V : w, w are absolutely continuous, w ∈ Y, w(0) =
w(π) = 0}. Then, A has the spectral representation
+∞

Aw = −n2 (w, wn )wn , w ∈ D(A),
n=1

where wn (s) = π2 sin ns and n = 1, 2, 3, ... is the orthogonal set of eigen-
functions of A. Further, it can be shown that A is the infinitesimal generator
of a strongly continuous cosine family C(t) : t ∈ R, defined on V which is
given by
+∞

C(t)w = cos nt(w, wn )wn , w ∈ V,
n=1
and, the associated sine family is given by
+∞
 1
S(t)w = sin nt(w, wn )wn , w ∈ V,
n=1
n
Let the control operator Bu : [0, T ] → V be defined by
(Bu)(t)(y) = μ(t, y); y ∈ (0, π)
For α = 2, the problem (5.1) can be rewritten as
y  (t) = Ay(t) + Bu(t) + f (t, xt ), t ∈ [0, T ]
y(t) = y0 = φ ∈ B, t ∈ (−∞, 0]
y  (0) = ψ ∈ Y, (5.2)
Therefore, by Theorem (4.1), if the hypothesis (H1 ) − (H4 ) are satisfied,the
differential system (2.1) is controllable on [0, T ].
For α ∈ (1, 2], since A is the infinitesimal generator of a strongly con-
tinuous cosine family C(t), from the subordinate principle ([11]), it follows
that A is the infinitesimal generator of a strongly continuous exponentially
bounded fractional cosine family Cα (t) such that Cα (0) = I, and
Vol. 13 (2016) Approximate Controllability of Semilinear 2549

∞
Cα (t) = ϕt,α/2 (s)C(s)ds, t > 0,
0
−α/2
where ϕt,α/2 (s) = t φα/2 (st−α/2 ), and
∞
 (−x)n
φγ (x) = , 0 < γ < 1.
n=0
n!Γ(−γn + 1 − γ)
The problem (5.1) can be rewritten as
C
Dtα y(t) = Ay(t) + Bu(t) + f (t, xt ), t ∈ [0, T ]
y(t) = y0 = φ ∈ B, t ∈ (−∞, 0]
y  (0) = ψ ∈ Y, (5.3)
Therefore, by Theorem (4.1), if the hypotheses (H1 ) − (H4 ) are satisfied, the
differential system (2.1) is controllable on [0, T ].

References
[1] Balachandran, K., Dauer, J.P.: Controllability of nonlinear systems in Banach
spaces: a survey. J. Optim. Theory Appl. 115(1), 7–28 (2002)
[2] Balachandran, K., Anthoni, S.M.: Controllability of second-order semilin-
ear neutral functional differential systems in Banach spaces. Comput. Math.
Appl. 41(10-11), 1223–1235 (2001)
[3] Barnett, S.: Introduction to Mathematical Control Theory. Clarendon
Press, Oxford (1975)
[4] Curtain, R.F., Zwart, H.J.: An Introduction to Infinite Dimensional Linear
Systems Theory. Spinger, New York (1995)
[5] Henrı́quez, H.R., Hernández, E.M.: Approximate controllability of second-
order distributed implicit functional systems. Nonlinear Anal. 70(2), 1023–
1039 (2009)
[6] Jeong, J.M., Kwun, Y.C., Park, J.Y.: Approximate controllability of semilin-
ear retarded functional differential equations. J. Dyn. Control Syst. 5, 329–
346 (1999)
[7] Kalman, R.E.: Controllability of linear systems,contributions to differential
equations. 1, 190–213 (1963)
[8] Kumar, S., Sukavanam, N.: Approximate controllability of fractional or-
der semilinear systems with bounded delay. J. Differ. Equ. 252(11), 6163–
6174 (2012)
[9] Kumar, S., Sukavanam, N.: On the approximate controllability of fractional
order control systems with delay. Nonlinear Dyn. Syst. Theory 13(1), 69–
78 (2013)
[10] Kumar, S., Sukavanam, N.: Controllability of second-order systems with non-
local conditions in Banach spaces. Numer. Funct. Anal. Optim. 35(4), 423–
431 (2014)
[11] Li, K., Peng, J., Gao, J.: Controllability of nonlocal fractional differential sys-
tems of order α ∈ (1, 2] in Banach spaces. Rep. Math. Phys. 71(1), 33–43 (2013)
2550 A. Shukla et al. MJOM

[12] Naito, K.: Controllability of semilinear control systems dominated by the linear
part. SIAM J. Control Optim. 25, 715–722 (1987)
[13] Naito, K., Park, J.Y.: Approximate controllability for trajectories of a delay
volterra control system. J. Optim. Theory Appl. 61(2), 271–279 (1989)
[14] Pazy, A.: Semigroup of Linear Operators and Application to Partial Differential
Equations. Spinger, New York (1983)
[15] Sakthivel, R., Ren, Y., Mahmudov, N.I.: On the approximate controllability of
semilinear fractional differential systems. Comput. Math. Appl. 62(3), 1451–
1459 (2011)
[16] Sakthivel, R., Mahmudov, N.I., Nieto, J.J.: Controllability for a class
of fractional-order neutral evolution control systems. Appl. Math. Com-
put. 218(20), 10334–10340 (2012)
[17] Shukla, A., Arora, U., Sukavanam, N.: Approximate controllability of retarded
semilinear stochastic system with non local conditions. J. Appl. Math. Comput.
49(1–2), 513–527 (2015)
[18] Sukavanam, N.: Approximate controllability of semilinear control systems with
growing nonlinearity. In: Mathematical theory of control proceedings of inter-
national conference, Marcel Dekker, New York, pp 353–357 (1993).
[19] Sukavanam, N., Tafesse, S.: Approximate controllability of a delayed semilinear
control system with growing nonlinear term. Nonlinear Anal. 74(18), 6868–
6875 (2011)
[20] Travis, C.C., Webb, G.F.: Cosine families and abstract nonlinear second order
differential equations. Acta Math. Acad. Sci. Hungar. 32(1-2), 75–96 (1978)
[21] Wang, J., Zhou, Y.: Existence and controllability results for fractional semi-
linear differential inclusions. Nonlinear Anal. Real World Appl. 12(6), 3642–
3653 (2011)
[22] Wang, J., Zhou, Y.: Analysis of nonlinear fractional control systems in Banach
spaces. Nonlinear Anal. 74(17), 5929–5942 (2011)
[23] Zhou, H.X.: Approximate controllability for a class of semilinear abstract equa-
tions. SIAM J. Control Optim. 21(4), 551–565 (1983)
[24] Zhou, Y., Jiao, F.: Existence of mild solutions for fractional neutral evolution
equations. Comput. Math. Appl. 59(3), 1063–1077 (2010)
[25] Zhou, Y.: Basic Theory of Fractional Differential Equations. World Scientific
Publishing Co. Pte. Ltd, Hackensack, NJ (2014)
[26] Zhou, Y.: Fractional Evolution Equations and Inclusions: Analysis and Con-
trol. Elsevier, New York (2015)

A. Shukla, N. Sukavanam and D. N. Pandey

Department of Mathematics
Indian Institute of Technology Roorkee
Roorkee, 247667
India
e-mail: [email protected]

Received: January 6, 2015.

Revised: September 4, 2015.
Accepted: September 25, 2015.