Sei sulla pagina 1di 22

Differ Equ Dyn Syst

DOI 10.1007/s12591-016-0340-8
ORIGINAL RESEARCH

Existence of Solutions and Approximate Controllability


of Fractional Nonlocal Stochastic Differential Equations
of Order 1 < q 2 with Infinite Delay and Poisson Jumps
P. Muthukumar1 K. Thiagu1,2

Foundation for Scientific Research and Technological Innovation 2016

Abstract In this paper, we investigate the existence of mild solutions and the approximate
controllability of a class of nonlinear fractional stochastic differential equations of order
1 < q 2 with infinite delay and Poisson jumps which satisfies the nonlocal conditions
in Hilbert space. The existence of mild solutions is proved by using Sadovskiis fixed point
theorem. Also the approximate controllability of the nonlinear fractional nonlocal stochastic
differential equations of order 1 < q 2 with infinite delay and Poisson jumps is checked by
using Lebesgue dominated convergence theorem. Finally an example is included to illustrate
the results.
Keywords Approximate controllability Fixed point theorem Fractional stochastic
differential systems Hilbert space Poisson jumps
Mathematics Subject Classification 34K50 93B05 93E03

Introduction
Fractional differential equations draw a great deal of attention and they play an important
role in many applied fields such as electrochemistry control, seepage flow in porous media,

This work was supported by Science Engineering Research Board (SERB), DST, Government of India under
YSS Project F. No: YSS/2014/000447 dated 20.11.2015.

P. Muthukumar
pmuthukumargri@gmail.com
K. Thiagu
thiyaguelakk@gmail.com

Department of Mathematics, The Gandhigram Rural InstituteDeemed University, Gandhigram


624302, Tamil Nadu, India

Department of Mathematics, Kandaswami Kandars College, Velur, Namakkal 638182, Tamil Nadu,
India

123

Differ Equ Dyn Syst

fluid dynamics, traffic models, electro magnetic, etc. The most important advantage of using
fractional differential equations in these and other applications is their nonlocal property. This
means that the next state of a system depends not only upon its current state of a system, but
also upon all its historical states. Recently fractional nonlocal stochastic integro-differential
equations with infinite delay have attractive much attention, because their applications in
many areas such as physics, population dynamics, electrical engineering. Very recently,
fractional neutral stochastic integro-differential equations with infinite delay have attracted
in interest [7]. Toufik Guendouzi et al. [11] established a set of sufficient conditions for
the existence and controllability result for fractional neutral stochastic integro-differential
equations with infinite delay.
Controllability of integer order stochastic differential systems in infinite dimensional
spaces are derived using different kinds of approaches [24,25]. Chang et al. [5] studied the
controllability of semilinear differential systems with nonlocal initial conditions in Banach
spaces. In recent years, various controllability problem for different kinds of dynamical
control systems have been studied [12,13,18,34]. There are many different notions of controllability both for linear and non-linear systems [9,16,21,23]. Klamka [15] studied stochastic
controllability and minimum energy control of systems with multiple delays in control. Further it is important to note that event driven dynamics become very useful in most fields of
application and lead to stochastic differential equations with jumps [8,27,28,32,35]. Controllability of deterministic and stochastic functional differential equations was discussed
in [22,32]. Muthukumar et al. [26] established approximate controllability of second-order
neutral stochastic differential equations with infinite delay and Poisson jumps. In modern
years, much attention has been paid to establish sufficient conditions for the controllability
of nonlinear fractional differential systems of order 1 < q < 2 by several authors including
monograph [17]. Shu et al. [33] discussed the existence and uniqueness of mild solutions
for fractional differential equations with nonlocal conditions of order 1 < q < 2. Yan [36]
established the controllability of fractional order partial neutral functional integro-differential
inclusions with infinite delay. On the problem of controllability of fractional neutral stochastic
systems with infinite delay is also established (see [30]).
Also in addition, fractional stochastic differential equations with Poisson jumps have
become very popular in modeling the phenomena arising in the fields, such as economics,
where the jump processes were widely used to describe the commodity price dynamics (see
[6]). Sakthivel et al. [31] studied a class of fractional control systems of nonlinear fractional
nonlocal dynamical systems of order 1 < q < 2, they established a new set of sufficient
conditions for the controllability of nonlinear fractional systems by using fixed point analysis
approach and extended the result to study the approximate controllability of the corresponding systems with nonlocal systems, which is an untreated topic in the literature and this fact
is the motivation of the present work. Kexue et al. [19] studied controllability of nonlocal
fractional differential systems of order 1 < q 2 in Banach spaces. Neutral differential
systems with impulses aries in many areas of applied mathematics and these systems have
been studied during the last decades. Chalishajar et al. [3] studied the existence results for
impulsive perturbed partial neutral functional differential equations in Frechet spaces. Chalishajar et al. [4] established the approximate controllability of abstract impulsive fractional
neutral evolution equations with infinite delay in Banach spaces. Hence we also discussed
the extension of our result for fractional nonlocal neutral impulsive stochastic differential
equations with infinite delay and Poisson jumps of orders 1 < q 2. Very recently, controllability of nonlocal second-order impulsive neutral stochastic functional integro-differential
equations with delay and Poisson jumps have attracted in interest [14]. Chalishajar [2] studied

123

Differ Equ Dyn Syst

the controllability of second order impulsive neutral functional differential inclusions with
infinite delay.
Most of the papers above about the fractional differential systems are concerned with the
fractional derivative whose order between zero and one. The purpose of this paper is to establish sufficient conditions for the approximate controllability of fractional nonlocal stochastic
differential equations of order q (1, 2]. When q = 2, the case corresponding to the second
order differential systems, the results in this paper are generalization of classical results.
In this paper, we address the existence of mild solutions and approximate controllability of
fractional nonlocal stochastic differential equations of order 1 < q 2 with infinite delay
and Poisson jumps in Hilbert space described by
c

Dt x(t)=Ax(t) + Bu(t) + f (t, xt )+

g(, x )d( )+

(dt, d),
h(t, x t , ) N
Z

t J := [0, b]
x0 (t) = (t) + m 1 (xt1 , ..., xtm )(t), 0 < t1 < t2 < < tm < b, m N,


x (0) + n 1 (x) = .

t (, 0],

(1)

Here the state variable x() takes values in a real separable Hilbert space H with the inner
q
product ,  and norm   H . The fractional derivative c Dt , 1 < q 2 is understood
in the Caputo sense. The control function u() is given in L2 (J, U ) of admissible control
functions with U as a Hilbert space. B is a bounded linear operator from U into H . Also
A : D(A) H H is the infinitesimal generator of a strongly continuous q-order
cosine family Cq (t) on H . Let K be another separable Hilbert space. Let {(t)}t0 be a
given K -valued Wiener process with a finite trace nuclear covariance operator Q 0. Let

q = (
q (t)), t D
q be a stationary Ft Poisson point process with characteristic measure .
Let
N
(dt,
d)
be
the
Poisson counting measure associated with 
q . Hence we have N (t, Z ) =

I (
q (s)) with measurable set Z 
B(K {0}), which denotes the Borel -field
sD
q ,st Z
(dt, d) = N (dt, d) dt(d) be the compensated Poisson measure
of K {0}. Let N
that is independent of (t). Let P2 ([0, b] Z ; H ) be the space all predictable mapping
b
: [0, b] Z H for which 0 Z E(t, )2H dt(d) < . We can define the
b
(dt, d), which is a centred square integrable
H-valued stochastic integral 0 Z (t, ) N
martingale. We can also employing the same notation   for the norm of L(K , H ), which
denotes the space of all bounded operators from K into H . Simply L(H ) if K = H . The
histories xt represents the function defined by xt : (, 0] H , xt ( ) = x(t + ), for
t 0 belong to some phase space Ch described axiomatically. Further f : J Ch H ,
g : J Ch L Q (K , H ) and h : J Ch Z H are nonlinear functions. Here L Q (K , H )
denotes the space of all Q-Hilbert Schmidt operators from K into H . Moreover m 1 : Chm
Ch and n 1 : H H are continuous functions. The initial data = {(t) : t (, 0]} is
an F0 measurable Ch -valued stochastic process independent of Brownian motion {(t)} and
Poisson point process 
q () with finite second moment. (t) is an F0 -measurable H -valued
random variable independent of wiener process (t) and Poisson point process 
q () with
finite second moment.

Preliminaries
Let (, F, P) be a complete probability space furnished with complete family of right continuous increasing sub -algebras {Ft : t J } satisfying Ft F. An H -valued random
variable is an F-measurable function x(t) : H and a collection of random variables

123

Differ Equ Dyn Syst

S = {x(t, w) : H |tJ } is called a stochastic process. Usually, we suppress the dependence on w and write x(t) instead of x(t, w) and x(t) : J H in the place of S.
Let n (t) (n = 1, 2, . . .) be a sequence of real valued one-dimensional
standard Brown
ian motions mutually independent over (, F, P). Set (t) =
n=1 n n (t)en , t 0,
where n 0 (n = 1, 2, . . .) are non-negative real numbers and {en } (n = 1, 2, . . .) is
complete orthonormal
basis in K . Let Q L(K , K ) be an operator defined by Qen = n en
with finite T r (Q) =
n=1 n < , T r (Q) denotes the trace of the operator Q. Then the
above K -valued stochastic process (t) is called a Q-Wiener process. Let us assume that
Ft = ((s) : 0 s t) is the 
-algebra
by and Ft = F. Let L(K , H )
generated
2
and define 2Q =T r ( Q  ) =
n=1  n en  . If  Q < , then is called a QHilbert Schmidt operator. Let L Q (K , H ) denotes the space of all Q-Hilbert Schmidt operators
: K H . The completion L Q (K , H ) of L(K , H ) with respect to the topology induced
by a norm   Q , where 2Q = ,  is a Hilbert space with the above norm topology. Let
C be the closed subspace of all continuously differentiable process that belong to the space
C (J, L2 (; H )) consisting of Ft -adapted measurable process such that F0 -adapted processes
0
, L2 (, Ch ) . Assume that h : (, 0] (0, +) with l = h(t)dt < + is a
continuous function.
Now we shall see an axiomatic definition of the abstract phase space Ch and is defined
by Ch = { : (, 0] H, for any a > 0, (E | ( ) |2 )1/2 is bounded and measurable
0
function on [a, 0] with (0) = 0 and h(s) sups 0 (E | ( ) |2 )1/2 ds < }. If
0
Ch is endowed with the norm Ch = h(s) sups 0 (E | ( ) |2 )1/2 ds, Ch . Then
(Ch ,  Ch ) is a Banach space.
Let B(H ) be the space of all bounded linear operators on H . Let I be the identity operator
on H . If A is a linear operator on H , then R(, A) = (I A)1 denotes the resolvent
operator of A. We can use the notation
k (t) =

t 1
, t > 0, > 0
()

where () is the Gamma function. If = 0, we set k0 (t) = (t), the delta distribution.
Definition 1 [1] The RiemannLiouville fractional integral of order q > 0 is defined by
 t
q
Jt x(t) =
kq (t s)x(s)ds
0

where x(t) L1 ([0, b]; H ).


Definition 2 [1] The RiemannLiouville fractional derivative of order 1 < q 2 is defined
by
q

Dt x(t) =

d 2 2q
J
x(t)
dt 2 t

where x(t) L1 ([0, b]; H ), Dt x(t) L1 ([0, b]; H )


Definition 3 [1] The Caputo fractional derivative of order 1 < q 2 is defined by
c

Dt x(t) = Dt (x(t) x(0) x  (0)t)


q

where x(t) L1 ([0, b]; H ) C 1 ([0, b]; H ), Dt x(t) L1 ([0, b]; H ).

123

Differ Equ Dyn Syst

The Laplace transform for the RiemannLiouville fractional integral is given by


1
x L ()
q
where x L () is the Laplace transform of x given by

x L () =
et x(t)dt, Re > 0 .
q

L[Jt x(t)] =

The Laplace transform of the Caputo derivative is given by


L[c Dt x(t)] = q x L () x(0)q1 x  (0)q2
q

Consider a class of fractional differential system with infinite delay


c

Dt x(t) = Ax(t) + Bu(t), t [0, b],


x(0) = x0 H,
x  (0) = x1 H,

(2)

c Dq
t

is the Caputo fractional derivative, A is the infinitesimal generator


where q (1, 2],
of a strongly continuous q-order cosine family {Cq (t)}t0 on the Hilbert space H , the state
values x() takes values in H , the control function u() is given in L2 (J, U ) of admissible
control functions with U , a Hilbert space, B is a bounded linear operator form U into H .
Definition 4 [10] The infinitesimal generator A : H H of {C(t) : t R} is given
d2
2
by Ax = dt
2 C(t)x|t=0 for all x D(A) = {x H : C()x C (R; H )}. It is known
that the infinitesimal generator A is closed, densely defined operator on H for cosine and
corresponding sine families and their generators.
Definition 5 [19] Let q (1, 2]. A family {Cq (t)}t0 B(H ) is called a solution operator
(or a strongly continuous q-order fractional cosine family) for the above problem (2) if the
following conditions are satisfied
(a)
(b)
(c)
(d)

Cq (0) = I , I is the identity operator in H ,


Cq (t)x is strongly continuous for t 0, for every x H ,
Cq (t)D(A) D(A) and ACq (t)x0 = Cq (t)Ax0 for all x0 D(A), t 0,
t
Cq (t)x0 is a solution of x(t) = x0 + 0 kq (t s)Ax(s)ds for all x0 D(A), t 0.

A is called the infinitesimal generator of Cq (t). The strongly continuous q-order fractional
cosine family is also called q-order cosine family.
The corresponding fractional sine family Sq : R+ B(H ) associated with Cq is defined
t
by Sq (t)x = 0 Cq (s)xds for t 0, x H .
Definition 6 The fractional RiemannLiouville family Tq : R+ B(H ) associated with
Cq is defined by
q1

Tq (t) = Jt

Cq (t).

Definition 7 The q-order cosine family Cq is called exponentially bounded if there are
constants M 1 and 0 0 such that
Cq (t) Me0 t , t 0.
An operator A is said to belong to C q (M, 0 ), if the problem (2) has an q-order cosine family
Cq (t) satisfying the above inequality.

123

Differ Equ Dyn Syst

Assume A C q (M, 0 ) and let Cq (t) be the corresponding q-order cosine family. Then we
have

q1 R(q , A)x0 =
et Cq (t)x0 dt, Re > 0 , x0 H,
0

et Sq (t)x0 dt, Re > 0 , x0 H,
q2 R(q , A)x0 =
0

R(q , A)x0 =
et Tq (t)x0 dt, Re > 0 , x0 H.
0

Based on the result in [13], the function x C([0, b]; H ) is called a mild solution of (2) if
x satisfies
 t
x(t) = Cq (t)x0 + Sq (t)x1 +
Tq (t s)Bu(s)ds, t J.
0

Existence of Mild Solutions


In order to see the existence of a mild solution and controllability of the fractional stochastic
dynamical system (1), we need to consider the basic definitions.
The Kuratowski measure of noncompactness is defined on each bounded subset B of a
Banach space X by


n

Bi , diam(Bi ) r
(B) = inf r > 0, B
i=1

Definition 8 (Kuratowskis measure of noncompactness [20]) Kuratowski introduced a measure (A) of non-compactness for subsets A of a Banach space X. Although is real valued
and fulfils: A B (A) (B), is not a measure in the measure theoretical sense. It
gives however an indication of the deviation closed subsets A of X have from being compact,
in the sense that A is compact if and only if (A) = 0.
Lemma 1 (Sadovskiis fixed point theorem [29]) Let 1 be a condensing operator on a
Banach space X , that is 1 is continuous and takes bounded sets into bounded sets, and
(1 (B)) (B) for every bounded set B of X with (B) > 0. If 1 (A) A for a
convex, closed and bounded set A of X , then 1 has a fixed point in X , where () denotes
Kuratowskis measure of non-compactness.
1

Lemma 2 [11] Assume that x Ch . Then for all t J , xt Ch , l(Ex(t)2 ) 2 xt Ch
0
1
l sups[0,t] (Ex(s)2 ) 2 + x0 Ch , where l = h(s)ds < .
Definition 9 [19] An H -valued stochastic process {x(t) : t (, b]} is said to be a mild
solution of the system (1) if
(i) x(t) is Ft -adapted and measurable for t 0,
(ii) x(t) is continuous on [0, b] almost surely and for each s [0, t), the function C(t)(t
s)q1 Tq (t s) f (s, xs ) is integrable such that the following stochastic integral equation

123

Differ Equ Dyn Syst

is satisfied
x(t) = Cq (t)((0) + m 1 (xt1 , ..., xtm )(0)) + Sq (t)( n 1 (x))
t

t
Tq (t s) f (s, xs )ds +

+
0

t
+

Tq (t s)

s

g(, x )d( ) ds

t 

Tq (t s)Bu(s)ds

(ds, d),
Tq (t s)h(s, xs , ) N

(3)

0 Z

(iii) x0 (t) = (t) + m 1 (xt1 , ..., xtm )(t), t (, 0], x  (0) + n 1 (x) = .
In order to prove the theorems, we need the following assumptions:
(H1) A is an infinitesimal operator of a strongly continuous q-order cosine family {Cq (t) :
t 0} on H . Also Cq (t) MC and Sq (t) M S . Moreover the operators Cq (t)
and Sq (t) are compact for t > 0.
(H2) The bounded linear operator Tq (t) is compact for t > 0 and there exists a positive
constant M such that
sup Tq (t) M.
tJ

(H3) The functions m 1 : Chm Ch and n 1 : H H are continuous, satisfy the Lipschitz
conditions and there exists positive constant Mn 1 > 0 such that
En 1 (x) n 1 (y)2H Mn 1 x y2Ch

for all x, y Ch , and t J .


(H4) The function f : J Ch H is continuous there exists a positive constant M f such
that
E f (t, x t )2H M f (1 + x t 2Ch ) for t J, x Ch ,
E f (t, x 1 ) f (t, x 2 )2H M f x 1 x 2 2Ch for t J and x 1 , x 2 Ch .

0
a a

(H5) For each Ch , K 1 (t) = lim

g(s, )d(s) exists and is continuous. Also there

exists a positive constant Mk such that EK 1 (t)2H Mk .


(H6) The functions g : J Ch L(K , H ) and h : J Ch Z H satisfies the following
properties:
(i) For each (t, s) J J , g(t, s, .) : Ch L(K , H ) is continuous and for each x Ch ,
g(., ., x) : J J L(K , H ) is strongly measurable.

123

Differ Equ Dyn Syst

(ii) The non-linear function h is a Borel measurable function which satisfy the Lipschitz
continuity condition and there exists positive constants Mh > 0 and L h > 0 such that



h (t, x, ) 2H (d) Mh 1 + x2Ch ,
Z
h (t, x1 , ) h (t.x2 , ) 2H (d) Mh x1 x2 2Ch ,
Z



h (t, x, ) 4H (d) L h 1 + x4Ch ,
Z
h (t, x1 , ) h (t.x2 , ) 4H (d) L h x1 x2 4Ch , for all x, x1 , x2 Ch , t J.
Z

(iii) there are positive integrable functions m, n L1 ([0, b]) and continuous non-decreasing
functions A g , Ah : [0, ) (0, ) such that for every (t, x) J Ch , we have
t
Eg(t, x)2H ds m(t)A g x2Ch , lim inf

A g ( )
= 1 < ,

Eh(t, x, )2H (d)ds n(t)Ah x2Ch , lim inf

Ah ( )
= 2 < .

t 
0

In order to study the approximate controllability of the fractional control system (1), we need
to introduce the relevant operator

0b

=
0

Tq (b s) B B Tq (b s) ds,

R(, 0b ) = ( I + 0b )1 for  > 0,


where B denotes the adjoint of B and Tq (T ) is the adjoint of Tq (t). It is clear that the
operator 0b is a linear bounded operator.
(H7)  R(, 0b ) 0 as  0+ in the strong operator topology.
Note that the assumption (H 7) is equivalent to the fact that the fractional linear control
system (2) is approximately controllable on J.
Definition 10 [11] Let xb (, ; u) be the state value of (1) at the terminal time b corresponding to the control u and the initial value . Introduce the set
R(b, , ) = {xb (, ; u)(0) : u() L2 (J, U )}
which is called the reachable set of (1) at the terminal time b and its closure in H is denoted
by R(b, , ). The system (1) is said to be approximately controllable on the interval J if
R(b, , ) = H.
 L2 (, L2 (0, b : L0 )) such that
Lemma 3 [11] For any 
x L2 (Fb , H ) there exists
2
F
b
(s)d(s).

xb = E
xb + 0

123

Differ Equ Dyn Syst

For any  > 0 and 


x L2 (Fb , H ), we define the control function


b
(s)d(s) Cq (b)((0) + m 1 (xt1 , ..., xtm )(0))
xb +
u  (t) = B Tq (b t)( I +0b )1 E
0

Sq (b)((0) n 1 (0)) B Tq (b t)

t

( I + 0b )1 Tq (b s) f (s, xs )ds

B Tq (b t)

t

( I + 0b )1 Tq (b s)

 s

B Tq (b t)

t 
0


g(, x )d( ) ds

(ds, d).
( I + 0b )1 Tq (b s)h(s, xs , ) N

(4)

First we shall see the existence of a mild solution to the fractional stochastic control system
(1).
Theorem 1 Assume that the assumptions (H1)(H6) hold. Then for each  > 0, the control
system (1) has a mild solution on [0, b] provided that




4M 2 M f b2 l 2 + 8M 2 b2 l 2 T r (Q)A g sup m(s) + (Mh + L h )Ah sup n(s)


30
5 + 2 (M M B )4 b2 < 1.


sJ

sJ

Proof Let C ((, b], H ) be the space of all continuous H -valued stochastic process { (t) :
t (, b]} and Cb = {x : x C ((, b], H ), x0 = Ch }. Let  b be the semi-norm
1
defined by xb = x0 Ch + sup0sb (Ex(s)2 ) 2 , x Cb . For any  > 0, define the
operator  : Ch Ch by

(t), t (, 0]

q (t)((0) + m 1 (x t1 , ..., x tm )(0)) + Sq (t)( n 1 (x))

t
t

T
(t

s)
f
(s,
x
)ds
+
Tq (t s)Bu  (s)ds
+
q
s

0
0


(  x)(t) =
t
s

g(, x )d( ) ds
+ Tq (t s)

t 

(ds, d), t J.
+ Z Tq (t s)h(s, xs , ) N
0

Using Holders inequality and assumption (H4), we get


2
 t

  t   t


2
 M2
E
T
(t

s)
f
(s,
x
)ds
ds
E

f
(s,
x
)
ds
q
s
s
H


0


M2 M f b
0

(1 + xs 2Ch )ds.

By Bochners theorem and Lemma 2, Tq (t s) f (s, xs ) is integrable on J .

123

Differ Equ Dyn Syst

Next we shall prove that  has a fixed point, by Sadovskiis theorem, which is then a
mild solution of the control system (1). Define
=

(t) t (, 0]
Cq (t)((0) + m 1 (xt1 , ..., xtm )(0)) + Sq (t)( n 1 (x)), t J

 Cb . Let x(t)=
(t) + z(t), t (, b]. Then x satisfies (1) if and only
It is clear that
if z 0 = 0 and
(t) + z(t))) +
z(t) = Sq (t)( n 1 (

t
0

t
+

Tq (t s)Bu  (s)ds +

t 
+

t
0

s + z s )ds
Tq (t s) f (s,

Tq (t s)

s

 + z )d( ) ds
g(,

(ds, d),
s + z s , ) N
Tq (t s)h(s,

0 Z

where u  (t) is defined by (4). Let Cb0 ={z Ch : z 0 = 0 Ch }. For any z Cb0 , we can have
1

zb = z 0 Ch + sup0sb (Ez(s)2 ) 2 = sup0s.b (Ez(s)2 ) 2 . Hence (Cb0 ,  b ) is a


Banach space. Set Br ={y Cb0 : y2b r } for each positive number r . Then for each r , Br
is a bounded closed convex set in Cb0 . Then by Lemma 2 for z Br , we have


t 2C
t + z t 2C 2 z t 2C + 

h
h
h

 (0)
4 l 2 sup Ez (s) 2 + z 0 2Ch + l 2 sup E
0st

0st

0 2C
+ m 1 (xt1 , ..., xtm ) (0) 2 + 
h

2
2 2
4 l r + l Mc E (0) + m 1 (xt1 , ..., xtm ) (0) 2H + 2Ch

(5)

Define the map on Cb0 as

0, t (, 0]

t

(t) + z(t))) + Tq (t s) f (s,


s + z s )ds
Sq (t)( n 1 (

0
 s

t
t
(z)(t) =
 (s)ds + T (t s)
 + z )d( ) ds
T
(t

s)Bu
g(s,

q
q

t 

(ds, d), t J
s + z s , ) N

+ Z Tq (t s)h(s,
0

Then is well-defined on Br for each r > 0. Also the operator  has a fixed point if and
only if has a fixed point.

123

Differ Equ Dyn Syst

Let = 1 + 2 . Then the operators 1 and 2 are defined on Br by


(t) + z(t))) +
(1 z)(t) = Sq (t)( n 1 (

t

s + z s )ds
Tq (t s) f (s,

t
(2 z)(t) =

 s

 + z )d( ) ds
Tq (t s)
g(,

t

Tq (t s)Bu (s)ds +
0

t

(ds, d)
s + z s , ) N
Tq (t s)h(s,

0 Z

In order to prove this theorem, we shall prove the next theorem.


Theorem 2 Assume that the above assumptions (H1)(H6) hold. Then 1 is a contraction
mapping and 2 is compact.
Proof The proof of this theorem needs following Lemmas.
Lemma 4 Consider the above assumptions (H1)(H6). For each  > 0, there exists a positive
number r such that (Br ) Br
Proof Suppose let us assume that (Br )  Br . Then for each positive number r , there exists a
function z r () Br , but (z r )
/ Br , that is E(z r )(t))2H > r for some t = t (r ) J .Then
for each  > 0, we have
r E(z r )(t))2H
 t
2


2



5ESq (t)( n 1 ( (t) + z(t))) H + 5E Tq (t s) f (s, s + z s )ds 


2
 t
 t
 s
 2








+ 5E
T
(t

s)Bu
(s)ds
+
5
E
T
(t

s)
g(,

+
z
)d(
)
ds 
q
q





H

2
 t 


 .


+ 5E
T
(t

s)
h
(s,

+
z
,
)
N
(ds,
d)
q
s
s


H

0 Z

If the right side terms are represented by Ii , i = 1, 2, . . . 5, then,


r 5

5


Ii .

(6)

i=1

Evaluating each term separately, we get





 (t) + z (t) 2H
I1 M S2 E n 1


M S2  2Ch + Z ,

(7)

where Z = 2Ch + m 1 (xt1 , ..., xtm )(t)2H +  2Ch + n 1 (x)2H


2
 t


r


I2 E Tq (t s) f (s, s + z s )ds 

0

123

Differ Equ Dyn Syst

t
M 2b

s + z rs )2H ds
E f (s,

t
M 2b

s + z rs 2H )ds
M f (1 + E

M 2 b2 M f (1+4(l 2 r + l 2 MC2 E(0)+m 1 (xt1 , ..., xtm )(0)2 + 2Ch ))


 2
 t



ds
T
(t

s)Bu
(s)
I3 E 
q



(8)

t
M b
2

EBu  (s)2 ds

M 2 b2 M B2 Eu  (s)2


Eu  (s)2 2 M 2 M B2 2E


x b 2 + 2


(9)
b

(s)2 ds + MC2 (2C + Z )


E
h

+ M S2 ( 2Ch + Z )
+ M 2 b2 M f (1 + 4(l 2 r + l 2 MC2 E(0) + m 1 (xt1 , ..., xtm )(0)2 + 2Ch ))

+ M 2 b2 2Mk + 2T r (Q)A g (4(l 2 r + l 2 MC2 E(0) + m 1 (xt1 , ..., xtm )(0)2

+ 2Ch )) sup m(s)
sJ


+ M b (2Mh + 2 L h )Ah (4(l 2 r + l 2 MC2 E(0) + m 1 (xt1 , ..., xtm )(0)2

+ 2Ch )) sup n(s)
(10)
2 2

sJ

Hence, we have


6
I3 (M 2 M B2 b2 ) 2 M 2 M B2 M R ,

where
b
xb  + 2
M R = 2E
2

(s)2 ds + MC2 (2C + Z ) + M S2 ( 2C + Z )


E
h
h

+ M 2 b2 M f (1 + 4(l 2 r + l 2 MC2 E(0) + m 1 (xt1 , ..., xtm )(0)2 + 2Ch ))



2 2
+ M b 2Mk + 2T r (Q)A g (4(l 2 r + l 2 MC2 E(0) + m 1 (xt1 , ..., xtm )(0)2

+ 2Ch )) sup m(s)
sJ

+ M 2 b2 (2Mh + 2 L h )Ah (4(l 2 r

123

+ l 2 MC2 E(0) + m 1 (xt1 , ..., xtm )(0))2

(11)

Differ Equ Dyn Syst

+ 2Ch )) sup n(s)


sJ

 t
 s
 2


r

T
(t

s)
g(,

+
z
)d(
)
ds 
I4 E
q




t 
M 2b

t
2Mk + 2T r (Q)

2





 + z r )d  ds
Eg(,
C


2 2
M b 2Mk + 2T r (Q)A g (4(l 2 r + l 2 MC2 E(0) + m 1 (xt1 , ..., xtm )(0)2

+ 2Ch )) sup m(s)
sJ

(12)

In the same way, we get


2
 b 

(ds, d)
s + z rs , ) N
I5 E
Tq (t s)h(s,

0

2M 2

 t
0

+ 2M 2




2


s + z rs , ) (d)ds
Eh(s,
H

 t 
0




M (2Mh + 2 L h )
2

4


s + z rs , ) (d)ds
Eh(s,
H

 t
0

2 



r


1 + s + z s 
ds

M b (2Mh + 2 L h )Ah (1 + 4(l r


2 2

 21

Ch
2 2
+ l MC E(0) + m 1 (xt1 , ..., xtm )(0)2

+ 2Ch )) sup n(s)

(13)

sJ

Combining these Eqs. (6)(13), we get



r L 0 + 20M M f b l r + 30(M b
2

2 2

2 2

M B2 )


1 2 2
M MB MR
2

+ 10M 2 b2 T r (Q)A g (4(l 2 r + l 2 MC2 E(0) + m 1 (xt1 , ..., xtm )(0)2 + 2Ch )) sup m(s)
+ 10M b (Mh +
2 2

sJ

L h )Ah (4(l r + l
2

MC2 E(0) + m 1 (xt1 , ..., xtm )(0)2

+ 2Ch )) sup n(s)


sJ

(14)

where


L 0 = 5M S2  2Ch + Z + 5M 2 b2 M f
(1 + 4(l 2 MC2 E(0) + m 1 (xt1 , ..., xtm ) (0) 2 + 2Ch )) + 10M 2 b2 Mk
 !
+ 10M 2 b2 Mh + L h Ah
(15)

123

Differ Equ Dyn Syst

Dividing both the sides of (14) by r and letting as r , we get



!

4M 2 M f b2 l 2 + 8M 2 b2 l 2 T r (Q)A g sup m(s) + (Mh + L h )Ah sup n(s)
sJ



30
5 + 2 (M M B )4 b2 > 1


sJ

which is a contradiction to our assumption. Thus for  > 0 there exists a positive number r
such that (Br ) Br .
Lemma 5 Let assumptions (H1)(H6) hold. Then 1 is a contraction mapping.
Proof Let v, v Br . Then
 t
E(1 v)(t) (1 v)(t)2H 2M 2 E

s + vs ) f (s,
s + v s )2H ds
 f (s,

t
2M M f

vs v s )2Ch ds

(2M 2 M f b) sup Ev v2H


0st

Let (2M 2 M f b) < 1. Then the assumption in Theorem 1 and (H4) gives the contraction of
the mapping 1 .
Lemma 6 Assume that r > 0, 2 (Br ) Br and let the assumptions (H1)(H6) be hold.
Then 2 maps bounded sets into bounded sets.
Proof For each t J, z Br and  > 0 from (5), we have
t + z t 2C 4(l 2 r + l 2 MC2 E(0) + m 1 (xt1 , ..., xtm )(0)2H + 2C ) = q 

h
h
2
 t




E(2 z)(t)2H 3E
 Tq (t s)Bu (s)ds 
H

 t
 s
 2




+ 3E Tq (t s)
g(, + z )d( ) ds 

0

2
 t 





+ 3E
Tq (t s)h(s, s + z s , ) N (ds, d)


0 Z

From Eqs. (10)(13), we have






6
E(2 z)(t)2H 3 2 (M M B )4 b2 M R + 3M 2 b2 2Mk + 2T r (Q)A g q  sup m(s)

sJ

2 2

+ 3M b (2Mh + 2 L h )Ah q sup n(s) =
sJ

Thus for each z Br , E2 z2b .

123

Differ Equ Dyn Syst

Lemma 7 Let assumptions (H1)(H6) holds. Then the set {2 z : z Br } is an equicontinuous family of functions on J .
Proof Let 0 < 1 < t < b and 1 > 0 such that Tq (s1 ) Tq (s2 ) < 1 with |s1 s2 | < 1
for s1 , s2 J . For z Br , 0 < |h 1 | < 1 , t + h 1 J , we have
E2 z(t + h 1 ) 2 z(t)2H

2
 t+h
 1




6M E
Bu (s)ds 

2

2
 t




+ 6E
[T
(t
+
h

s)

T
(t

s)]Bu
(s)ds
q
1
q



 t+h
 2
 1 s


 + z )d( ) ds 
+ 6M 2 E
g(,




 t
 s
 2



+ 6E
[T
(t
+
h

s)

T
(t

s)]
g(,

+
z
)d(
)
ds 
q
1
q




 t+h
2
 1 





+ 6M E
h(s, s + z s , ) N (ds, d) 

2

 t

2





+ 6E
[T
(t
+
h

s)

T
(t

s)]
h
(s,

+
z
,
)
N
(ds,
d)
q
1
q
s
s


0

Applying the assumptions (H1)(H6) and Holder inequality, we get


E|2 z (t + h 1 ) 2 z (t) 2H
t+h
 1

6M

t

Eu (s)  ds

M B2

+ 612 M B2

t+h
 1

+ 6M 2

Eu  (s) 2 ds


2Mk + 2T r (Q) m (s) A g q  ds

t 
+ 612

2Mk + 2T r (Q) sup m (s) A g q


t+h
 1
2
t

t
+ 612
0


ds

sJ

+ 6M

 !
2Mh + 2 L h sup n (s) Ah q  ds
sJ

 !
2Mh + 2 L h sup n (s) Ah q  ds.
sJ

123

Differ Equ Dyn Syst

Hence for 1 sufficiently small, the right side of the above inequality tends to zero as h 1 0.
Also the compactness of Tq (t), t > 0 implies the continuity in the uniform operator topology.
Hence the set {2 z : z Br } is equicontinuous.


Lemma 8 Let assumptions (H1)(H6) holds. Then 2 maps Br into a precompact set in Br .
Proof Let 0 < t b be fixed and 1 be a real number satisfying 0 < 1 < t. For 1 > 0
define the operator 21 ,1 on Br by
21 ,1 (t)

t
 1

Tq (t s)Bu  (s)ds

0
t
 1

+
0
t
 1

 s

 + z )d( ) ds
Tq (t s)
g(,



(ds, d) .
s + z s , ) N
Tq (t s)
h(s,

But for t > 0, Tq (t) is a compact operator. Hence the set {(21 ,1 z)(t) : z Br } is precompact
in H for every 1 (0, t), 1 > 0. Also for each z Br , we have
 ,1

E(2 z)(t) (21

 t


312 E


z)(t)2H

2

Tq (t s)Bu  (s)ds 


t1



+ 312 E


t


Tq (t s)



+ 312 E


 2

 + z )d( ) ds 
g(,


t1

t

s


2

(ds, d) 
s + z s , ) N
Tq (t s)
h(s,


t1

 ,
E(2 z)(t) (21 1 z)(t)2H
t 

3M 2 12

M B2

6 2 2
M M B bM R
q2

t1

+ b(2Mk + 2T r (Q)A g q  sup m(s)) + b(2Mh


sJ


+ 2 L h )Ah q  sup n(s) ds 0, as 1 , 1 0 .


sJ

Hence there are relatively compact sets arbitrary close to the set {(2 z)(t) : z Br } and
hence the set {(2 z)(t) : z Br } is also precompact in Br . Hence from the Lemmas 48
under the Theorem 1 and ArzelaAscoli theorem satisfies all the conditions of Sadovskiis
fixed point theorem (Lemma 1), the control system (1) has mild solutions on [0, b].

123

Differ Equ Dyn Syst

Approximate Controllability
Theorem 3 Assume that the assumptions of Theorem1 and (H1)(H7) are hold. Also the
functions f, g and h are uniformly bounded on their respective domains. Further, if Sq (t)
and Cq (t) are compact, then the control system (1) is approximately controllable on J .
Proof Let x  be a fixed point of the operator  . By using the stochastic Fubini theorem,
we get

b
(s)d(s) Cq (b)((0) + m 1 (xt1 , ..., xtm )(0))
x  (b) = 
xb ( I + 0b )1 E
xb +
0


b
Sq (b)((0) n 1 (0)) +  ( I + 0b )1 Tq (b s) f (s, xs )ds
0

b
( I

+

 s


s)
g(, x )d( ) ds

b

+ 0b )1 Tq (b

+
0

(ds, d)
( I + 0b )1 Tq (b s)h(s, xs , ) N

The properties of f , g and h implies that g(, x  )2 + h(s, xs , )2 K 1 and
 f (s, xs )2 K 2 . Then there is a subsequence denoted by { f (s, xs ), g(, x ), h(s, xs , )}
which converges weakly to, say, { f (s), g(s, ), h(s, )}.
From the above equation, we have
Ex  (b) 
x b 2

9E( I + 0b )1 [E
xb Cq (b)((0) + m 1 (xt1 , ..., xtm )(0))]2
 b 

2


+ 9E( I + 0b )1 Sq (b)((0) n 1 (0))2 + 9E
( I + 0b )1 (s) 0 ds
0

L2

 b 
 2


+ 9E
( I + 0b )1 Tq (b s)[ f (s, xs ) f (s)]ds
0

 b 
 2


b 1
I
+

+ 9E
)
T
(b

s)
f
(s)
ds
(
q
0
0

 b 
 s
 2


b 1


+ 9E
[g(, x ) g(s, )] 
( I + 0 ) Tq (b s)
ds
0

 b 
 s
 2


b 1

+ 9E
g(s, ) 
( I + 0 ) Tq (b s)
ds
0

123

Differ Equ Dyn Syst

 b  
2


b 1


+ 9E
( I + 0 ) Tq (b s)[h(s, xs , ) h(s, )] N (ds, d)
0

 b  
2


b 1

+ 9E
( I + 0 ) Tq (b s)h(s, ) N (ds, d)
0

Hence for all 0 s b, the operator ( I + 0b )1 0 strongly as  0 and ( I +


0b )1  1. Thus by the Lebesque dominated convergence theorem and the compactness of
Sq (t) and Cq (t) implies that Ex  (b)
xb 2 0. This gives the approximate controllability
of the control system (1).


Remark 1 In this remark we consider a nonlinear fractional nonlocal neutral impulsive stochastic differential equations of order 1 < q 2 with infinite delay and Poisson jumps of
the form
 t

c q
(dt, d)
Dt [x(t) f (t, xt )] = [Ax(t)+Bu(t)]dt +
g(, x )d( ) +
h(t, xt , ) N

t J := [0, b]\{t1 , t2 , . . . , tn }
x0 (t) = (t) + m 1 (xt1 , ..., xtm )(t),
0 < t1 < t2 < < tm < b, m > n, m, n N t (, 0],


x (0) + n 1 (x) = .
x(tk ) = Ik (xtk ),
x  (tk ) = I k (xtk ), k = 1, 2, ..., n.

(16)

where A, B, g and h are defined as before. The function f : J Ch H and the appropriate
functions Ik , I k : H H satisfies the following conditions.
(H8) The function f : J Ch H is continuous there exists a positive constant M f such
that
E f (t, x t )2H M f (1 + x t 2Ch ) for t J, x Ch ,
E f (t, x 1 ) f (t, x 2 )2H M f x 1 x 2 2Ch for t J and x 1 , x 2 Ch .

(H9) The functions Ik and I k : H H are continuous and there exists positive constants
M Ik ,M I k , k = 1, 2, ..., n such that
EIk (x) Ik (y)2H M Ik x y2Ch , x, y Ch , k = 1, 2, ..., n,
EI k (x) I k (y)2H M I k x y2Ch , x, y Ch , k = 1, 2, ..., n.
(H10) The functions Ik and I k : H H are continuous and there exists positive nondecreasing functions k , k : [0, ) (0, ), k = 1, 2, ..., n such that for every
x Ch and k = 1, 2, ..., n, we have
A Ik ( )
= 4 < ,

A I k ( )
k A I k x2Ch , lim inf
= 5 < .

EIk (x)2H k A Ik x2Ch , lim inf



EI k (x)2H

123

Differ Equ Dyn Syst

An H -valued stochastic process {x(t) : t (, b]} is said to be a mild solution of the


system (16) if the following stochastic integral equation is satisfied
x(t) = Cq (t)((0) + m 1 (xt1 , ..., xtm )(0))
+ Sq (t)(( n 1 (x)) f (0, ( + m 1 (xt1 , ..., xtm )))
t

t
Tq (t s) f (s, xs )ds +

+
0

 s

Tq (t s)
g(, x )d( ) ds

t
+

t 
+

(ds, d)
Tq (t s)h(s, xs , ) N

0 Z

Tq (t s)Bu(s)ds

Tq (t tk )Ik (xtk ) +

0<tk <t

Tq (t tk )I k (xtk ), k = 1, 2, ..., n.

(17)

0<tk <t

Example
Consider the following controlled fractional system in the form
c

q
Dt z(t, x)

2
=
z(t, x) + f (t, z(t h 1 , x)) + (t, x) +
x2

t
g(s, t, z(t h 1 , x))d(s)


+

(dt, dy),
h(t, z(t h 1 , x), y) N

z(t, 0) = z(t, ) = 0, z  (t, 0) = z  (t, ) = 0, t J,


z(t, x) = (t, x), z  (t, x) = (t, x), t (, 0],


m
ci z(ti h 1 , x) (t) = x0 (t),
z(t, 0) +
i=1

z  (t, 0) = x  (0) +

m


di x, 0 < x < ,

(18)

i=1
q

where c Dt is the Caputo fractional partial derivative of order 1 < q 2 , : [0, 1]


(0, ) (0, ) is continuous, h 1 > 0 and ci , di > 0 for i = 1, 2, .., m. Also (t) denotes a
standard one-dimensional Wiener process defined on a stochastic basis (, F, {Ft }t0 , P).To
write the above system (18) into the abstract form (1), we can choose the space H = U =
2

L2 ([0, ]). Define A : H H by Ax=x , (Az)(x)= z(x)
with the domain D(A) = {x
x2


for z D(A). Also


H : x, x  are absolutely continuous, x H and x(0) = x() = 0} and
2
the set {en , n N} is an orthonormal basis of H . In particular Ax =
n=1 n x, en en ,
x D(A).
Then
A
is
the
infinitesimal
generator
of
a
strongly
continuous
cosine
family

C(t)x=
cos(nt)x,
e
e
,
t

R
and
the
associated
sine
family
in
H
defined
by
S(t)x
=
n n
n=1

123

Differ Equ Dyn Syst

sin(nt)
n x, en en , t

R. From [27], for all x H , t R, Sq (t) 1 and Cq (t) 1.


Hence the
spectrum
A
consists
of eigen values n for n N, with associated eigen vectors
"
n=1

en (x) =

2
sin(nx),

n N.

0
Now we present a phase space Ch . Let h(s) = e2s , s < 0, then l = h(s)ds = 21 . Let
0
1
Ch = h(s) sup(E(0)2 ) 2 ds. Then (Ch ,  Ch ) is a Banach space. We define an
#

# 2
infinite dimensional space U by U = {u =
u
e
(x)
# n=2 u n < }, 0 x , and
n=2 n n
 2 1
the norm in U is u = ( n=2 u n ) 2 . The bounded linear operator B : U H is defined
by Bu(t)(x)=(t, x), 0 x . The functions f, g and h are continuous. Let (, F, P)
be a complete probability space and {K (t) : t J } is a Poisson point process taking values
in the space K = [0, ) with a -finite intensity measure (dy). The Poisson counting
(dt, dy) is induced by K () and the compensating martingale measure is denoted
measure N
(dt, dy) := N (dt, dy) dt(dy).
by N
For q (1, 2), A generates a strongly continuous cosine family C(t), it follows from
the subordinate principle (see [1, Theorem 3.1]) that A generates a strongly continuous
exponentially bounded fractional cosine family Cq (t) such that Cq (0) = I and

Cq (t) =

t, q (s)C(s)ds, t > 0,
2
!
q
q

and
where t, q (s) = t 2  q st 2
0

 (x) =


n=0

(x)n
, (0 < < 1).
n! (n + 1 )

It is clear that the fractional differential system (2) is approximately controllable on J for
q = 2. Hence for q = 2 with the above choices, the system (18) can be rewritten to the
abstract form (1) and all the conditions of Theorem 1 are satisfied. Thus there exists mild
solutions for the system (18). Moreover all the conditions of Theorem 3 are satisfied and
hence the fractional stochastic differential equations with Poisson jumps (18) is approximately controllable on [0, b]. Similarly we can prove the existence of mild solutions and the
approximate controllability of the fractional stochastic system (16).

Conclusion
In this paper, the existence of mild solutions and approximate controllability of nonlinear
fractional nonlocal stochastic differential equations of order 1 < q 2 with infinite delay and
Poisson jumps is studied by using fixed point theory with the corresponding linear system is
assumed as controllable. We also discussed the extension of our result for fractional nonlocal
neutral impulsive stochastic differential equations with infinite delay and Poisson jumps of
orders 1 < q 2. For the future work, the controllability results could be extended to
study sufficient conditions for stochastic nature of wave equations, heat equations, Burgers
equations and Navier stochastic equations with infinite delay and Poisson jumps satisfying
the nonlocal conditions as in (1).
Acknowledgements The authors would like to express their sincere thanks to the editor and anonymous
reviewers for helpful comments and suggestions to improve the quality of this manuscript.

123

Differ Equ Dyn Syst

References
1. Bazhlekova, E.G.: Fractional evolution equations in Banach spaces. Dissertation, University Press Facilities, Eindhoven University of Technology (2001)
2. Chalishajar, D.N.: Controllability of second order impulsive neutral functional differential inclusions with
infinite delay. J. Optim. Theory Appl. 154(2), 672684 (2012)
3. Chalishajar, D.N., Karthikeyan, K., Anguraj, A.: Existence results for impulsive perturbed partial neutral
functional differential equations in Frechet spaces Dynamics of Continuous. Discrete Impuls. Syst. Ser.
A Math. Anal. 22, 2545 (2015)
4. Chalishajar, D.N., Malar, K., Karthikeyan, K.: Approximate controllability of abstract impulsive fractional
neutral evolution equations with infinite delay in Banach spaces. Electron. J. Differ. Equ. 275(2013), 121
(2013)
5. Chang, Y.K., Nieto, J.J., Li, W.S.: Controllability of semilinear differential systems with nonlocal initial
conditions in Banach spaces. J. Optim. Theory Appl. 142(2), 267273 (2009)
6. Cont, R., Tankov, P.: Financial modelling with jump processes. In: Financial Mathematics Series. Chapman
and Hall/CRC, Boca Raton (2004)
7. Cui, J., Yan, L.: Existence result for fractional neutral stochastic differential equations with infinite delay.
J. Phys. A. Math. Theor. 44(33), 335201 (2011)
8. Cui, J., Yan, L.: Successive approximation of neutral stochastic evolution equations with infinite delay
and Poisson jumps. Comput. Math. Appl. 218(2), 67766784 (2012)
9. Debbouche, A., Torres, D.F.: Approximate controllability of fractional delay dynamic inclusions with
nonlocal control conditions. Comput. Math. Appl. 243, 161175 (2014)
10. Fattorimi, H.O.: Second Order Linear Differential Equations in Banach Spaces. North Holland, Amsterdam (1985)
11. Guendouzi, T., Hamada, I.: Existence and cotrollability result for fractional neutral stochastic integrodifferential equations with infinite delay. Adv. Model. Optim. 15(2), 281299 (2013)
12. Guendouzi, T., Hamada, I.: Relative controllability of fractional stochastic dynamical systems with multiple delays in control. Malaya J. Mat. 1(1), 8697 (2013)
13. Guendouzi, T., Idrissi, S.: Approximate controllability of fractional stochastic functional evolutions driven
by a fractional Brownian motion. Romain J. 8(2), 103117 (2012)
14. Huan, D.D., Gao, H.: Controllability of nonlocal second-order impulsive neutral stochastic functional
integro-differential equations with delay and Poisson jumps. Cogent Eng. 2(1), 1065585 (2015)
15. Klamka, J.: Stochastic controllability and minimum energy control of systems with multiple delays in
control. Comput. Math. Appl. 206(2), 704715 (2008)
16. Klamka, J.: Constrained controllability of semilinear systems. Nonlinear Anal. 47(5), 29392949 (2001)
17. Kaczorek, T.: Selected Problems of Fractional Systems Theory, vol. 411. Springer, New York (2011)
18. Kumar, K., Kumar, R.: Controllability of Sobolev type nonlocal impulsive mixed functional integrodifferential evolution systems. Electron. J. Math. Anal. Appl. 1(3), 122132 (2015)
19. Kexue, L., Jigen, P., Jinghuai, G.: Controllability of nonlocal fractional differential systems of order
(1, 2]. Rep. Math. Phys. 71(1), 3343 (2013)
20. Lowen, R.: Kuratowskis measure of non compactness. Rev. Quit. J. Math. (Oxford) 39(2), 235254
(1988)
21. Mahmudov, N.I.: Controllability of linear stochastic systems. IEEE Trans. Autom. Control 46(5), 724731
(2001)
22. Mahmudov, N.I.: Approximate controllability of nonlinear deterministic and stochastic evolution equations in abstract spaces. SIAM J. Control Optim. 42(5), 16041622 (2003)
23. Mokkedem, F.Z., Fu, X.: Approximate controllability of semi-linear neutral integro-differential systems
with finite delay. Comput. Math. Appl. 242, 202215 (2014)
24. Muthukumar, P., Balasubramaniam, P.: Approximate controllability of second order damped Mckean
Vlasov stochastic evolution equations. Comput. Math. Appl. 60(10), 27882796 (2010)
25. Muthukumar, P., Balasubramaniam, P.: Approximate controllability of mixed stochastic Volterra
Fredholm type integro-differential systems in Hilbert space. J. Franklin Inst. 348(10), 29112922 (2011)
26. Muthukumar, P., Rajivgandhi, C.: Approximate controllability of second-order neutral stochastic differetial equations with infinite delay and Poisson jumps. J. Syst. Sci. Complex 28(5), 10331048 (2015)
27. Ren, Y., Sakthivel, R.: Existence, uniqueness, and stability of mild solutions for second-order neutral
stochastic evolution equations with infinite delay and Poisson jumps. J. Math. Phys. 53(7), 073517 (2012)
28. Rajivganthi, C., Thiagu, K., Muthukumar, P., Balasubramaniam, P.: Existence of solutions and approximate controllability of impulsive fractional stochastic differential systems with infinite delay and Poisson
jumps. Appl. Math. 60(4), 395419 (2015)
29. Sadovskii, B.N.: On a fixed point principle. Funct. Anal. Appl. 1(2), 151153 (1967)

123

Differ Equ Dyn Syst


30. Sakthivel, R., Ganesh, R., Suganya, S.: Approximate controllability of fractional neutral stochastic systems
with infinite delay. Rep. Math. Phys. 70(3), 291311 (2012)
31. Sakthivel, R., Ganesh, R., Ren, Y., Anthoni, S.M.: Approximate controllability of nonlinear fractional
dynamical systems. Commun. Nonlinear Sci. Numer. Simul. 18(2), 34983508 (2013)
32. Sakthivel, R., Ren, Y.: Complete controllability of stochastic evolution equations with jumps. Rep. Math.
Phys. 68(2), 163174 (2011)
33. Shu, X.B., Wang, Q.: The existence and uniqueness of mild solutions for fractional differential equations
with nonlocal conditions of order 1<<2. Comput. Math. Appl. 64(6), 21002110 (2012)
34. Sukavanam, N., Tafesse, S.: Approximate controllability of a delayed non-linear control system with
growing non-linear term. Nonlinear Anal. TMA 74(18), 68686875 (2011)
35. Tan, J., Wong, H., Guo, Y.: Existence and uniqueness of solutions to neutral stochastic functional differential equations with Poisson jumps. Abstr. App. Anal. (article ID 371239) (2012)
36. Yan, Z.: Controllability of fractional order partial neutral functional integro-differential inclusions with
infinite delay. J. Franklin Inst. 348(8), 21562173 (2011)

123

Potrebbero piacerti anche