Physica A 536 (2019) 122602

Contents lists available at ScienceDirect

Physica A
journal homepage: www.elsevier.com/locate/physa

Impulsive synchronization of complex-variable network with

distributed time delays
∗
Hui Leng, Zhaoyan Wu
College of Mathematics and Information Science, Jiangxi Normal University, Nanchang 330022, China

article info a b s t r a c t

Article history: Impulsive control scheme is adopted to consider the synchronization of complex-
Received 25 March 2019 variable dynamical network with distributed time delays. Based on the Lyapunov
Received in revised form 7 July 2019 function method and mathematical analysis technique, simple synchronization condi-
Available online 6 September 2019
tions with respect to system parameters, impulsive gains and intervals are firstly derived.
Keywords: For any given network, the impulsive intervals can be estimated from the derived
Complex-variable network conditions when the impulsive gains are fixed, and vice versa. Further, adaptive strategy
Synchronization is introduced to design unified impulsive controllers for different networks. Noticeably,
Impulsive control in the proposed adaptive impulsive control scheme, the impulsive gains can be adjusted
to the needed values and the impulsive instants can be estimated adaptively. Finally,
numerical examples are provided to verify the effectiveness of the theoretical results.
© 2019 Elsevier B.V. All rights reserved.

1. Introduction

Dynamical networks, including real-variable and complex-variable networks, have been extensively investigated within
the past decades and broadly applied to describe physical and social systems coupled with a great deal of interactive
individuals. The nodes represent the individuals of the dynamical networks and the edges represent the interactions
among individuals. Therefore, it is necessary to deliberate on node dynamics and the topological structure of dynamical
network for better studying its dynamical behaviors, e.g., synchronization.
In many practical situations, there are usually propagation delays, which have been observed in lasers, neuronal models,
electronic circuits and so on [1,2]. Many efforts have been devoted to study the complex dynamical networks with delays
including constant and time-varying delays [3–6]. That is, the current state of a node is affected by the outdated states
of its neighbor nodes. Many phenomena existing in the real world indicate that the current state of a node is affected
by the states in the previous period of its neighbor nodes [7–10]. Obviously, the former networks cannot characterize
the phenomena. Hence, the dynamical networks with distributed time delays are introduced to model systems [7–10].
In order to serve the practical life better, the dynamical networks with distributed time delays are worthy of serious
investigation.
Synchronization, as an interesting and important collective behavior of dynamical networks, has attracted wide
attentions from various areas [11–21]. There are huge number of advantages in practical applications such as regulation of
power grid, parallel image processing, the operation of no-man air vehicle, the realization of chain detonation, etc [18,19].
However, achieving synchronization through inner adjustment of networks is difficult and even impossible, i.e., without
external control. Large number of obtained results show that dynamical networks combined with effective controllers are

∗ Corresponding author.
E-mail address: zhywu@jxnu.edu.cn (Z. Wu).

https://doi.org/10.1016/j.physa.2019.122602
0378-4371/© 2019 Elsevier B.V. All rights reserved.
2 H. Leng and Z. Wu / Physica A 536 (2019) 122602

easy to achieve synchronization. Many kinds of effective controllers are designed by adopting different control schemes,
including feedback control [22–24], impulsive control [25–34], intermittent control [35–40], and so on. Impulsive control is
a representative discontinuous control scheme, in which the controllers acted on the nodes are only at a series of discrete
instants. In regard to designing impulsive controllers, how to derive the sufficient conditions about the impulsive instants,
impulsive gains and system parameters for achieving synchronization is the key issue. In Ref. [28], synchronization of
nonlinear network is investigated via impulsive control and sufficient conditions are derived. From the derived conditions,
for any given network, one can choose proper impulsive instants and gains satisfying the conditions through calculations.
But for different networks with nonidentical parameters, there is no doubt that the repetitive calculations are required
and workload is increased. In Ref. [34], adaptive strategy is adopted and combined with impulsive control to design
unified controllers and avoid the repetitive calculations. In the proposed adaptive impulsive scheme, some constants with
respect to system parameters need not be calculated beforehand, and the impulsive gains and intervals can be adjusted
or estimated according to the adaptive strategy.
As we know, complex-variable systems have been widely used to analyze physical systems processes. For example,
the complex Lorenz system is introduced in Ref. [41], which is used to describe and simulate rotating fluids and detuned
laser [42,43]. Further, synchronization of complex-variable chaotic systems and networks have also been studied and some
valuable results have been obtained [44–50]. In Ref. [44], synchronization of complex-variable network is investigated via
pinning impulsive control. In Ref. [47], complex projective synchronization in drive-response stochastic coupled networks
with complex-variable systems and coupling time delays is studied. Compared with real-variable networks, complex-
variable networks have more interaction manners, e.g., the coupling matrix can be complex. Also, complex-variable
networks have higher dimensional dynamics. In other words, the obtained results about real-variable networks cannot
be directly used for complex-variable networks. Up to now, synchronization of complex-variable dynamical networks
with distributed time delays via impulsive control has seldom been investigated. Thus, how to design effective impulsive
controllers for this kind of network deserves further studies.
Motivated by the above discussions, the synchronization of complex-variable dynamical networks with distributed
time delays is investigated in this paper. The main contributions of this paper are: Firstly, we derive the sufficient
conditions without adaptive strategy for achieving synchronization according to the Lyapunov function method and
mathematical analysis technique. Secondly, we combine adaptive strategy with impulsive control to design unified
impulsive controllers and derive the sufficient conditions as well. Finally, we provide the updating law of the impulsive
gain when the impulsive intervals are fixed, and the method for estimating the impulsive instants when the impulsive
gains are fixed.
The rest of this paper is organized as follows. In Section 2, the network model is introduced and some preliminaries
are given. In Section 3, the impulsive controllers for achieving synchronization are designed and the sufficient conditions
are provided. In Section 4, several numerical simulations are performed to verify the results. In Section 5, the conclusion
for this paper is given.

2. Model description and preliminaries

Consider a dynamical network consisting of N nodes with distributed time delays, described by
N
∑ ∫ τ
ẋk (t) = f (xk (t)) + akl H q(v )xl (t − v )dv, (1)
l=1 0

where k = 1, 2, . . . , N, and xk (t) = (xk1 (t), xk2 (t), . . . , xkn (t))T ∈ C n is the state variable of node k, f : C n → C n is
a vector function, H = diag(h1 , h2 , . . . , hn ) is the ∫inner coupling matrix, τ > 0 and the continuous density function
τ
q : [0, τ ] → [0, +∞] is prescribed which satisfies 0 q(v )dv = 1. A = (akl ) ∈ RN ×N is the zero-row-sum outer coupling
matrix denoting the network topology and coupling strength, defined as: if node l has influence on node k (k ̸ = l), then
akl ̸ = 0; otherwise, akl = 0.
The objective here is to synchronize network (1) with desired orbit ω(t), i.e.,
lim ∥xk (t) − ω(t)∥ = 0,
t →∞

where ∥ · ∥ denotes Euclidean norm, ω̇(t) = f (ω(t)).

For achieving the synchronization, proper impulsive controllers are designed and applied onto network (1). The
controlled network can be described as follows
N
∑ ∫ τ
ẋk (t) =f (xk (t)) + akl H q(v )xl (t − v )dv, t ̸ = tφ ,
l=1 0 (2)
xk (tφ+ ) =xk (tφ− ) + η(tφ ) xk (tφ− ) − ω(tφ ) , t = tφ ,
( )

where k = 1, 2, . . . , N, φ = 1, 2, . . ., the impulsive time instants tφ satisfies 0 = t0 < t1 < t2 < · · · < tφ < · · ·, and
tφ → ∞ as φ → ∞. xk (tφ+ ) = limt →t + xk (t), xk (tφ− ) = limt →t − xk (t), Any solution of (2) is assumed to be left continuous
φ φ
at each tφ , i.e., xk (tφ− ) = xk (tφ ). η(tφ ) is the impulsive gain at t = tφ and η(t) = 0 for t ̸ = tφ .
 H. Leng and Z. Wu / Physica A 536 (2019) 122602 3

∑N ∫τ
The synchronization errors are denoted by Ek (t) = xk (t) − ω(t). It is clear that l=1 akl H 0
q(v )ω(t − v )dv = 0. Then,
the error systems are
N
∑ ∫ τ
Ėk (t) =f (xk (t)) − f (ω(t)) + akl H q(v )El (t − v )dv, t ̸ = tφ ,
l=1 0 (3)
Ek (tφ+ ) = 1 + η(tφ ) Ek (tφ− ), t = tφ ,
( )

Assumption 1. Suppose that there exists a positive constant L such that

(y(t) − x(t))T (f (y(t)) − f (x(t))) + (f (y(t)) − f (x(t)))T (y(t) − x(t)) ≤ L(y(t) − x(t))T (y(t) − x(t)),
holds for any x(t), y(t) ∈ C n and t > 0.

3. Main results

Let E(t) = (E1T (t), E2T (t), . . . , ENT (t))T , τφ = tφ − tφ−1 be the impulsive intervals, λ be the largest eigenvalue of matrix
(A ⊗ H)(A ⊗ H)T , Ω = L + 1 + λ and δ (t) = (1 + η(t))2 . From the definition of η(t), one has δ (t) = 1 for t ̸ = tφ . Clearly,
the network (2) is discontinuous at each impulsive instant tφ . Denote K (t , v ) as the discontinuous points at time interval
(t − v, t), i.e., K (t , v ) = {φ| t − v < tφ < t , t > 0, v ∈ [0, τ ], φ = 1, 2, . . .}. Moreover, let Kmin (t , v ) and Kmax (t , v ) be
the minimum and maximum values in K (t , v ).

Theorem 1. Suppose that Assumption 1 holds. If there exists a constant ξ > 0 such that
ln δ (tφ ) + ξ + Ω τφ < 0, φ = 1, 2, . . . , (4)
hold, then network (2) can achieve synchronization.

Proof. Consider the following Lyapunov function:

N N ∫ τ (∫ tK−
min (t ,v )
∑ ∑
V (t) = EkT (t)Ek (t) + δ (tφ ) q(v ) EkT (s)Ek (s)ds
k=1 k=1 0 t −v

Kmax (t ,v )−1 ∫ −
t(p ∫ t
∑ +1) )
+ EkT (s)Ek (s)ds + EkT (s)Ek (s)ds dv,
p=Kmin (t ,v ) tp+ tK+
max (t ,v )

for t ∈ (tφ−1 , tφ ], φ = 1, 2, . . ..
When t ∈ (tφ−1 , tφ ),
N N ∫ τ (∫ tK−
min (t ,v )
∑ ∑
V (t) = EkT (t)Ek (t) + q(v ) EkT (s)Ek (s)ds
k=1 k=1 0 t −v

Kmax (t ,v )−1 ∫ −
t(p ∫ t
∑ +1) )
+ EkT (s)Ek (s)ds + EkT (s)Ek (s)ds dv,
p=Kmin (t ,v ) tp+ tK+
max (t ,v )

and the derivative of V (t) is

N
∑
EkT (t)(f (xk (t)) − f (ω(t))) + (f (xk (t)) − f (ω(t)))T Ek (t)
( )
V̇ (t) =
k=1
N
∑ N
∑ ∫ τ ∫ τ
akl EkT (t)H q(v )El (t − v )dv + q(v )ElT (t − v )dv HEk (t)
( )
+
k=1 l=1 o o

N ∫
∑ τ N ∫
∑ τ
+ q(v )EkT (t)Ek (t)dv − q(v )EkT (t − v )Ek (t − v )dv.
k=1 0 k=1 0

According to Assumption 1, one has

N
∑ N
∑ N
∑ ∫ τ ∫ τ
EkT (t)Ek (t) + + akl EkT (t)H q(v )El (t − v )dv + q(v )ElT (t − v )dv HEk (t)
( )
V̇ (t) ≤L
k=1 k=1 l=1 o o

∑ N N
∑∫ τ
+ EkT (t)Ek (t) − q(v )EkT (t − v )Ek (t − v )dv
k=1 k=1 0
4 H. Leng and Z. Wu / Physica A 536 (2019) 122602
∫ τ
=LE T (t)E(t) + q(v ) E T (t)(A ⊗ H)E(t − v ) + E T (t − v )(A ⊗ H)T E(t) dv
( )
0
∫ τ
T
+ E (t)E(t) − q(v )E T (t − v )E(t − v )dv.
0

As we know, for any complex vectors x and y, the inequality xT ȳ + yT x̄ ≤ xT x̄ + yT ȳ holds. Therefore, one has
E T (t)(A ⊗ H)E(t − v ) + E T (t − v )(A ⊗ H)T E(t) ≤E T (t)(A ⊗ H)(A ⊗ H)T E(t) + E T (t − v )E(t − v )
≤λE T (t)E(t) + E T (t − v )E(t − v ),
and
∫ τ ∫ τ
q(v ) E T (t)(A ⊗ H)E(t − v ) + E T (t − v )(A ⊗ H)T E(t) dv ≤ λE T (t)E(t) + q(v )E T (t − v )E(t − v )dv,
( )
0 0

and
V̇ (t) ≤(L + 1 + λ)E T (t)E(t)
≤Ω V (t),
which leads to

1 ) exp Ω (t − tφ−1 ) , t ∈ (tφ−1 , tφ ).

+
( )
V (t) ≤ V (tφ− (5)
When t = tφ ,
N N ∫ τ (∫ tK−
min (t ,v )
∑ ∑
V (tφ+ ) = EkT (tφ+ )Ek (tφ+ ) + δ (tφ ) q(v ) EkT (s)Ek (s)ds
k=1 k=1 0 tφ+ −v

Kmax (t ,v )−1 ∫ −
t(p ∫ tφ+
∑ +1) )
+ EkT (s)Ek (s)ds + EkT (s)Ek (s)ds dv
p=Kmin (t ,v ) tp+ tK+
max (t ,v )

N N ∫ τ (∫ tK− (6)
)2 ∑ min (t ,v )
∑
= 1 + η(tφ ) EkT (tφ− )Ek (tφ− ) + δ (tφ ) q(v ) EkT (s)Ek (s)ds
(
k=1 k=1 0 tφ− −v

Kmax (t ,v )−1 ∫ −
t(p ∫ tφ−
∑ +1) )
+ EkT (s)Ek (s)ds + EkT (s)Ek (s)ds dv
p=Kmin (t ,v ) tp+ tK+
max (t ,v )

=δ (tφ )V (tφ− ).
When φ = 1, from (5) and (6),
V (t1− ) ≤ V (t0 ) exp Ω τ1
( )

V (t1+ ) ≤ δ (t1 )V (t1− ) ≤ δ (t1 )V (t0 ) exp(Ω τ1 ).

When φ = 2,
V (t2− ) ≤ V (t1+ ) exp(Ω τ2 ) ≤ δ (t1 )V (t0 ) exp Ω (τ1 + τ1 ) ,
( )

V (t2+ ) ≤ δ (t2 )V (t2− ) ≤ δ (t2 )δ (t1 )V (t0 ) exp Ω (τ1 + τ2 )

( )

2
∏
= V (t0 ) δ (tϵ ) exp(Ω τϵ ).
ϵ=1

By induction,
φ
∏
V (tφ+ ) ≤ V (t0 ) δ (tϵ ) exp(Ω τϵ ), φ = 1, 2, 3, . . . .
ϵ=1

From inequalities (4),

δ (tϵ ) exp(Ω τϵ ) ≤ exp(−ξ ), ϵ = 1, 2, . . . ,
and
V (tφ+ ) ≤ V (t0 ) exp(−φξ ).

That is, V (tφ+ ) → 0 when φ → ∞.

 H. Leng and Z. Wu / Physica A 536 (2019) 122602 5

Therefore, when t ∈ (tφ , tφ+1 ],

V (t) ≤ V (tφ+ ) exp(Ω (t − tφ )),
i.e., V (t) → 0 and ∥Ei (t)∥ → 0 as t → ∞. This completes the proof.

Remark 1. Specially, when K (t , v ) = ∅, the Lyapunov function is

N
∑ N ∫
∑ τ ∫ t
V (t) = EkT (t)Ek (t) + δ (tφ ) q(v ) EkT (s)Ek (s)dsdv;
k=1 k=1 0 t −v

when K (t , v ) = {a}, i.e., Kmin (t , v ) = Kmax (t , v ) = a,

N
∑ N ∫
∑ τ (∫ ta− ∫ t )
V (t) = EkT (t)Ek (t) + δ (tφ ) q(v ) EkT (s)Ek (s)ds + EkT (s)Ek (s)ds dv.
k=1 k=1 0 t −v ta+

Theorem 2. Suppose that Assumption 1 holds. If there exists a constant ξ > 0 such that
ln δ (tφ ) + ξ + L̂(tφ )τφ < 0, φ = 1, 2, . . . , (7)
˙ ∑N
hold, where L̂(t) = ρ k=1 EkT (t)Ek (t) and ρ > 0 is the adaptive gain, then network (2) can achieve synchronization.

Proof. Consider the following Lyapunov function:

N N ∫ τ (∫ tK−
min (t ,v )
∑ ∑
V (t) = EkT (t)Ek (t) + δ (tφ ) q(v ) EkT (s)Ek (s)ds
k=1 k=1 0 t −v

Kmax (t ,v )−1 ∫ −
t(p ∫ t
∑ +1) )
+ EkT (s)Ek (s)ds + EkT (s)Ek (s)ds dv
p=Kmin (t ,v ) tp+ tK+
max (t ,v )

δ (tφ ) ( )2
+ L̂(t) − L − 1 − λ ,
ρ
for t ∈ (tφ−1 , tφ ], φ = 1, 2, . . ..
When t ∈ (tφ−1 , tφ ),
1( )
˙
V̇ (t) ≤ (L + 1 + λ)E T (t)E(t) + L̂(t) − L − 1 − λ L̂(t)
ρ
≤ L̂(t)V (t)
≤ L̂(tφ )V (t),
which gives

1 ) exp L̂(tφ )(t − tφ−1 ) , t ∈ (tφ−1 , tφ ),

+
( )
V (t) ≤ V (tφ−
When t = tφ ,
V (tφ+ ) = δ (tφ )V (tφ− ).
Thus, similar to the proof of Theorem 1, the proof can be completed.

Remark 2. In general, for any given τφ and ξ , the conditions (7) can be satisfied by choosing
( ξ + L̂(t )τ ) ( ξ + L̂(t )τ )
φ φ φ φ
− exp − − 1 + ε ≤ η(tφ ) ≤ exp − − 1 − ε,
2 2
where ε > 0 is an arbitrary constant. In practical applications, one concrete network model with adaptive impulsive
control can be written as follows:
N
∑ ∫ τ
ẋk (t) =f (xk (t)) + akl H q(v )xl (t − v )dv, t ̸ = tφ ,
l=1 0

xk (tφ+ ) =xk (tφ− ) + η(tφ ) xk (tφ− ) − ω(tφ ) , t = tφ ,

( )

N
˙ ∑
L̂(t) =ρ EkT (t)Ek (t),
k=1
6 H. Leng and Z. Wu / Physica A 536 (2019) 122602

Fig. 1. The orbits of the real and imaginary parts of xkm (t) and ωm (t), k = 1, 2, . . . , 8, m = 1, 2, 3.

( ξ + L̂(t )τ )
φ φ
η(tφ ) = exp − − 1 − ε,
2

where ρ > 0 is the adaptive gain, ξ and ε are small positive constants, and L̂(0) > 0. According to the definition of L̂(t),
L̂(t) is monotone increasing and converges to a positive constant when the synchronization is achieved. Then, η(tφ ) is
monotone decreasing and converges to the needed value as time goes on.

Remark 3. In view of practical applications, when impulsive gains are fixed, the larger impulsive intervals, the lower
control cost. Therefore, one can choose the impulsive intervals as large as possible through finding the maximum value
of tφ subject to

tφ = tφ−1 − ln δ (tφ ) + ξ L̂−1 (tφ ) − ι, φ = 1, 2, . . . ,

( )

where ι is a small positive constant.

4. Numerical illustrations

In this section, three numerical examples are provided to demonstrate the effectiveness of the theoretical results. The
inner coupling matrix and the density function are chosen as H = I3 and q(v ) = 1/τ with τ = 0.2. The values of xi (t) are
chosen as (cos(t), sin(t), cos(t) + sin(t))T for t ∈ [−τ , 0) and i = 1, 2, . . . , N.

Example 1. We consider complex-variable Chen system [49] as node dynamics of an eight-node network,

ẋk1 =27(xk2 − xk1 ),

ẋk2 = − 4xk1 − xk1 xk3 + 23xk2 ,
ẋk3 =(xk1 xk2 + xk1 xk2 )/2 − xk3 ,

where xk1 and xk2 are complex variable, xk3 is real variable. According to Ref. [50], there exists a constant L = 79 such
that Assumption 1 holds.
 H. Leng and Z. Wu / Physica A 536 (2019) 122602 7

Fig. 2. The orbits of the real and imaginary parts of xkm (t) and ωm (t), k = 1, 2, . . . , 8, m = 1, 2, 3.

Choose the outer coupling matrix A as

−5 1 1 1 0 1 1 0
⎡ ⎤
⎢ 1 −8 3 4 0 0 0 0 ⎥
⎢
⎢ 1 3 −8 2 0 2 0 0 ⎥
⎥
1 4 2 −9 0 0 2 0
A = 0.1 × ⎢ ⎥,
⎢ ⎥
⎢ 0 0 0 0 −2 1 0 1 ⎥
⎢ ⎥
⎢ 1 0 2 0 1 −6 2 0 ⎥
1 0 0 2 0 1 −4 0
⎣ ⎦
0 0 0 1 1 0 0 −2
then the maximum eigenvalue of (A ⊗ H)(A√⊗ H)T is λ = 1.6535 and Ω = 81.6535. Choose ω(0) = (5 − 4j, 2 + j, 3)T ,
xk (0) = (0.5k + 0.2kj, k + 0.3kj, k)T with j = −1, τφ = 0.15, η(tφ ) = −0.999 and ξ = 0.001, one has ln δ (tφ ) +ξ + Ω τφ =
−1.5665 < 0, and thus conditions (4) hold. Fig. 1 shows the orbits of the real and imaginary parts of xkm (t) and ωm (t),
k = 1, 2, . . . , 8, m = 1, 2, 3. The superscripts r and i denote the real and imaginary parts, respectively.

Example 2. Consider the same network in Example 1 via adaptive impulsive control.
First, choose the impulsive intervals τφ as a constant τφ = 0.15, ρ = 0.001, ξ = 0.001 and L̂(0) = 0.05. According to
Remark 2, choose

ξ + L̂(tφ )τφ
η(tφ ) = exp(− )−1−ε
2
with ε = 0.001. Fig. 2 shows the orbits of the real and imaginary parts of xij (t) and ωj (t). Fig. 3 shows the impulsive gain
η(tφ ).
Second, choose the impulsive gain η(tφ ) = −0.999, L̂(0) = 0.05, ξ = 0.001 and ρ = 0.0012. The impulsive instants
can be estimated in terms of Remark 3. Fig. 4 shows the orbits of the real and imaginary parts of xij (t) and ωj (t). Fig. 4
shows the impulsive interval τφ .

From Fig. 5, it is clear that the network achieve synchronization when the impulsive interval converge to 1.463, which
is much larger than the estimated value 0.15. That is, adaptive impulsive control scheme can make the impulsive interval
as large as possible and heavily reduce the control cost.
8 H. Leng and Z. Wu / Physica A 536 (2019) 122602

Fig. 3. The impulsive gain η(tφ ) versus φ .

Fig. 4. The orbits of the real and imaginary parts of xkm (t) and ωm (t), k = 1, 2, . . . , 8, m = 1, 2, 3.

Example 3. The network topology is generated as the BA scale-free network [51] with N = 20, m = m0 = 3. Choose the
node dynamics as complex-variable Lorenz system [52].

ẋk1 =2(xk2 − xk1 ),

ẋk2 =(60 + 0.02j)xk1 − xk1 xk3 − (1 − 0.06j)xk2 ,
ẋk3 =(xk1 xk2 + xk1 xk2 )/2 − 0.8xk3 ,

Choose the initial values ω(0) = (5 − 4j, 2 + j, 3)T and xk (0) = (0.1k + 0.2kj, 0.1k + 0.3kj, 0.2k)T .
First, choose the impulsive intervals τφ as a constant τφ = 0.5, ρ = 0.00005, ξ = 0.001, and L̂(0) = 0.05. According
to Remark 2, choose

ξ + L̂(tφ )τφ
η(tφ ) = exp(− )−1−ε
2
 H. Leng and Z. Wu / Physica A 536 (2019) 122602 9

Fig. 5. The impulsive interval τφ versus φ .

Fig. 6. The orbits of the real and imaginary parts of xkm (t) and ωm (t), k = 1, 2, . . . , 8, m = 1, 2, 3.

with ε = 0.001. Fig. 6 shows the orbits of the real and imaginary parts of xij (t) and ωj (t), i = 1, 2, . . . , 50, j = 1, 2, 3.
Fig. 7 shows the impulsive gain η(tφ ).
Second, choose the impulsive gain η(tφ ) = −0.8, L̂(0) = 0.05, ξ = 0.001, ρ = 0.002. The impulsive instants can be
estimated in terms of Remark 3. Fig. 8 shows the orbits of the real and imaginary parts of xij (t) and ωj (t). Fig. 9 shows
the impulsive interval τφ .
Examples 2 and 3 show that networks with totally different node dynamics and topological structures can achieve
synchronization without calculating constants about the system parameters in advance. That is, the adaptive impulsive
controllers are unified for different networks.

5. Conclusions

In this paper, the synchronization of complex-variable dynamical network with distributed time delays via impulsive
control is studied. Based on the Lyapunov function method and mathematical analysis technique, several synchronization
10 H. Leng and Z. Wu / Physica A 536 (2019) 122602

Fig. 7. The impulsive gain η(tφ ) versus φ .

Fig. 8. The orbits of the real and imaginary parts of xkm (t) and ωm (t), k = 1, 2, . . . , 8, m = 1, 2, 3.

conditions without adaptive strategy are obtained and generalized to adaptive cases. Meanwhile, for achieving syn-
chronization of the dynamical network, proper impulsive controllers are designed and improved for designing unified
controllers by adopting adaptive strategy. What is more, the updating law of the impulsive gain and the method for
estimating the impulsive instants are provided from the synchronization conditions. That is, any given networks via
adaptive impulsive control can achieve synchronization without calculating constants about the system parameters in
advance. Finally, the effectiveness of the theoretical results are demonstrated by several numerical examples.

Acknowledgments

This work is jointly supported by the NSFC under Grant No. 61463022, the NSF for Distinguished Young Scholar of
Jiangxi Province of China under Grant 20171BCB23031, and the program of China Scholarships Council under Grant No.
201708360078.
 H. Leng and Z. Wu / Physica A 536 (2019) 122602 11

Fig. 9. The impulsive interval τφ versus φ .

References

[1] T. Heil, I. Fischer, W. Elsäser, J. Mulet, C.R. Mirasso, Chaos synchronization and spontaneous symmetry-breaking in symmetrically delay-coupled
semiconductor lasers, Phys. Rev. Lett. 86 (2001) 795–C798.
[2] M. Ciszak, F. Marino, R. Toral, S. Balle, Dynamical mechanism of anticipating synchronization in excitable systems, Phys. Rev. Lett. 93 (2004)
114102.
[3] Y. Zhao, F. Fu, J. Wang, J. Feng, H. Zhang, Synchronization of hybrid-coupled delayed dynamical networks with noises by partial mixed impulsive
control strategy, Physica A 492 (2018) 1181–1193.
[4] C. Zhang, X. Wang, C. Luo, J. Li, C. Wang, Robust outer synchronization between two nonlinear complex networks with parametric disturbances
and mixed time-varying delays, Physica A 494 (2018) 251–264.
[5] Y. Xu, D. Meng, C. Xie, G. You, W. Zhou, A class of fast fixed-time synchronization control for the delayed neural network, J. Franklin Inst. B
355 (2018) 164–176.
[6] W. Yang, W. Yu, J. Cao, F.E. Alsaadi, T. Hayat, Global exponential stability and lag synchronization for delayed memristive fuzzy CohenCGrossberg
BAM neural networks with impulses, Neural Netw. 98 (2018) 122–153.
[7] C. Wille, J. Lehnert, E. Schoell, Synchronization-desynchronization transitions in complex networks: An interplay of distributed time delays and
inhibitory nodes, Phys. Rev. E 90 (2014) 032908.
[8] P. Rao, Z. Wu, M. Liu, Adaptive projective synchronization of dynamical networks with distributed time delays, Nonlinear Dynam. 67 (2012)
1729–1736.
[9] Y. Zhang, S. Liu, R. Yang, Y. Tan, X. Li, Global synchronization of fractional coupled networks with discrete and distributed delays, Physica A
514 (2019) 830–837.
[10] A.D. Kachhvah, The effect of distributed time-delays on the synchronization of neuronal networks, Eur. Phys. J. B 90 (2017) 8.
[11] S. Jiang, X. Lu, G. Cai, S. Cai, Adaptive fixed-time control for cluster synchronisation of coupled complex networks with uncertain disturbances,
Internat. J. Systems Sci. 48 (2017) 3382–3390.
[12] Y. Xu, W. Zhou, J. Fang, C. Xie, D. Tong, Finite-time synchronization of the complex dynamical network with non-derivative and derivative
coupling, Neurocomputing 173 (2016) 1356–1361.
[13] K. Li, J. Zhou, W. Yu, M. Small, X. Fu, Adaptive cluster synchronization in networks with time-varying and distributed coupling delays, Appl.
Math. Model. 38 (2014) 1300–1314.
[14] K. Li, W. Sun, M. Small, X. Fu, Practical synchronization on complex dynamical networks via optimal pinning control, Phys. Rev. E 92 (2015)
010903.
[15] W. Sun, Y. Wu, J. Zhang, S. Qin, Inner and outer synchronization between two coupled networks with interactions, J. Franklin Inst. B 352 (2015)
3166–3177.
[16] J. Zhou, Y. Zhao, Z. Wu, Cluster synchronization of fractional-order directed networks via intermittent pinning control, Physica A 519 (2019)
22–33.
[17] F. Li, Z. Ma, Q. Duan, Partial component synchronization on chaotic networks, Physica A 515 (2019) 707–714.
[18] S. Boccaletti, V. Latora, Y. Moreno, M. Chavez, D. Hwang, Complex networks: Structure and dynamics, Phys. Rep. 424 (2006) 175–308.
[19] G. Chen, X. Wang, X. Li, Introduction To Complex Networks: Models, Structure and Dynamics, High Education Press, Beijing, 2012.
[20] L. Wang, J. Zhang, W. Sun, Adaptive outer synchronization and topology identification between two complex dynamical networks with
time-varying delay and disturbance, IMA J. Math. Control Inform. (2018) dny013.
[21] S. Cai, F. Zhou, Q. He, Fixed-time cluster lag synchronization in directed heterogeneous community networks, Physica A 525 (2019) 128–142.
[22] D. Tong, W. Zhou, X. Zhou, J. Yang, L. Zhang, Y. Xu, Exponential synchronization for stochastic neural networks with multi-delayed and Markovian
switching via adaptive feedback control, Commun. Nonlinear Sci. Numer. Simul. 29 (2015) 359–371.
[23] V. Novicenko, Delayed feedback control of synchronization in weakly coupled oscillator networks, Phys. Rev. E 92 (2015) 022919.
[24] Y. Liu, J. Lu, J. Liang, Feedback controller design for the synchronization of boolean control networks, IEEE Trans. Neural Netw. Learn. Syst. 27
(2016) 1991–1996.
[25] F. Wang, Y. Yang, A. Hu, X. Xu, Exponential synchronization of fractional-order complex networks via pinning impulsive control, Nonlinear
Dynam. 82 (2015) 1979–1987.
[26] X. Li, S. Song, Research on synchronization of chaotic delayed neural networks with stochastic perturbation using impulsive control method,
Commun. Nonlinear Sci. Numer. Simul. 19 (2014) 3892–3900.
[27] Z. Wu, G. Chen, X. Fu, Outer synchronization of drive-response dynamical networks via adaptive impulsive pinning control, J. Franklin Inst. B
352 (2015) 4297–4308.
[28] J. Lu, Z. Wang, J. Cao, D.W.C. Ho, J. Kurths, Pinning impulsive stabilization of nonlinear dynamical networks with time-varying delay, Int. J.
Bifurcation Chaos 22 (2012) 1250176.
12 H. Leng and Z. Wu / Physica A 536 (2019) 122602

[29] J. Lu, C. Ding, J. Lou, J. Cao, Outer synchronization of partially coupled dynamical networks via pinning impulsive controllers, J. Franklin Inst.
352 (2015) 5024–5041.
[30] X. Yang, J. Cao, J. Lu, Stochastic synchronization of complex networks with nonidentical nodes via hybrid adaptive and impulsive control, IEEE
Trans. Circuits Syst. I. Regul. Pap. 59 (2012) 371–384.
[31] S. Cai, P. Zhou, Z. Liu, Synchronization analysis of hybrid-coupled delayed dynamical networks with impulsive effects: A unified synchronization
criterion, J. Franklin Inst. B 352 (2015) 2065–2089.
[32] S. Cai, X. Li, Q. Jia, Z. Liu, Exponential cluster synchronization of hybrid-coupled impulsive delayed dynamical networks: average impulsive
interval approach, Nonlinear Dynam. 85 (2016) 2405–2423.
[33] J. Wang, J. Feng, C. Xu, Y. Zhao, M.Z.Q. Chen, Synchronization of linear dynamical networks under stochastic impulsive coupling protocols, J.
Franklin Inst. 354 (2017) 4882–4895.
[34] Z. Wu, Adaptive impulsive outer synchronization between drive-response dynamical networks, Commun. Theor. Phys. 61 (2014) 590–594.
[35] X. Liu, T. Chen, Synchronization of complex networks via aperiodically intermittent pinning control, IEEE Trans. Automat. Control 60 (2015)
3316–3321.
[36] S. Cai, X. Li, F. Zhou, Adaptive intermittent control for exponential synchronization of directed networks with complex-variable systems and
distributed delays coupling, Internat. J. Modern Phys. C 29 (2018) 1850110.
[37] P. Zhou, S. Cai, S. Jiang, Z. Liu, Exponential cluster synchronization in directed community networks via adaptive nonperiodically intermittent
pinning control, Physica A 492 (2018) 1267–1280.
[38] S. Cai, X. Lei, Z. Liu, Outer synchronization between two hybrid-coupled delayed dynamical networks via aperiodically adaptive intermittent
pinning control, Complexity 21 (2016) 593–605.
[39] P. Zhou, S. Cai, J. Shen, Z. Liu, Adaptive exponential cluster synchronization in colored community networks via aperiodically intermittent
pinning control, Nonlinear Dynam. 92 (2018) 905–921.
[40] J. Wang, J. Feng, C. Xu, Y. Zhao, Almost sure exponential synchronisation of networked harmonic oscillators via intermittent coupling subject
to Markovian jumping, IET Control Theory Appl. 12 (2018) 1658–1664.
[41] A.C. Fowler, J.D. Gibbon, M.J. McGuinness, The complex lorenz equations, Physica D 4 (1982) 139–163.
[42] C.Z. Ning, H. Haken, Detuned lasers and the complex lorenz equations: Subcritical and supercritical hopf bifurcations, Phys. Rev. A 41 (1990)
3826.
[43] J.D. Gibbon, M.J. McGuinnes, The real and complex lorenz equations in rotating fluids and lasers, Physica D 5 (1982) 108–122.
[44] Z. Wu, D. Liu, Q. Ye, Pinning impulsive synchronization of complex-variable dynamical network, Commun. Nonlinear Sci. Numer. Simul. 20
(2015) 273–280.
[45] Z. Wu, X. Fu, Complex projective synchronization in drive-response networks coupled with complex-variable chaotic systems, Nonlinear Dynam.
72 (2013) 9–15.
[46] D. Ding, J. Yan, N. Wang, D. Liang, Adaptive synchronization of fractional order complex-variable dynamical networks via pinning control,
Commun. Theor. Phys. 68 (2017) 366–374.
[47] X. Wu, C. Xu, J. Feng, Complex projective synchronization in drive-response stochastic coupled networks with complex-variable systems and
coupling time delays, Commun. Nonlinear Sci. Numer. Simul. 20 (2015) 1004–1014.
[48] D. Ding, J. Yan, N. Wang, D. Liang, Pinning synchronization of fractional order complex-variable dynamical networks with time-varying coupling,
Chaos Solitons Fractals 104 (2017) 41–50.
[49] G.M. Mahmoud, T. Bountis, E.E. Mahmoud, Active control and global synchronization of the complex chen and lü systems, Int. J. Bifurcation
Chaos 17 (2007) (2007) 4295–4308.
[50] Z. Wu, G. Chen, X. Fu, Synchronization of a network coupled with complex-variable chaotic systems, Chaos 22 (2012) 023127.
[51] A.L. Barabási, R. Albert, Emergence of scaling in random networks, Science 286 (1999) 509–512.
[52] J.D. Gibbon, M.J. McGuinness, The real and complex lorenz equations in rotating fluids and lasers, Physica D 5 (1982) 108–122.