 Research
 Open Access
 Published:
Synchronization stability of delayed discretetime complex dynamical networks with randomly changing coupling strength
Advances in Difference Equations volume 2012, Article number: 208 (2012)
Abstract
This paper addresses a delaydependent synchronization stability problem for discretetime complex dynamical networks with interval timevarying delays and randomly changing coupling strength. The randomly changing coupling strength is considered with the concept of binomial distribution. By constructing a suitable LyapunovKrasovskii functional and utilizing reciprocally convex approach and Finsler’s lemma, the proposed synchronization stability criteria for the networks are established in terms of linear matrix inequalities which can be easily solved by various effective optimization algorithms. The networks are represented by use of the Kronecker product technique. The effectiveness of the proposed methods will be verified via numerical examples.
1 Introduction
During the last few years, complex dynamic networks (CDNs), which is a set of interconnected nodes with specific dynamics, have received increasing attention from the real world such as the Internet, the World Wide Web, social networks, electrical power grids, global economic markets, and so on. Also, many models were proposed to describe various complex networks, smallworld network and scalefree network, etc. For more details, see the literature [1–4]. In the implementation of many practical CDNs, there exists timedelay because of the finite speed of information processing or amplifiers. It is well known that timedelay often causes undesirable dynamic behaviors such as oscillation and instability of the network. Therefore, various approaches to synchronization analysis for CDNs with timedelay have been investigated in the literature [5–12]. By using network modeling with coupling delays, Li et al. [5] proposed, for the first time, the synchronization criteria for the CDNs with timedelay which were expressed in the form of linear matrix inequalities (LMIs). Koo et al. [9] presented a synchronization criterion for singular CDNs with timevarying delays. In [10–12], various synchronization problems are addressed for discretetime CDNs with timedelay. In this regard, discretetime modeling with timedelay plays an important role in many fields of CDNs. Moreover, to implement the network, the network uses digital computers (usually a microprocessor or microcontrollers) with the necessary input/output hardware. The fundamental character of the digital computer is that it computes answers at discrete steps.
On the other hand, in [13–15], the problems for various systems with randomly occurring delay, uncertainties and nonlinearities were considered. The randomly occurring considerations in these literature works are described by the Bernoulli distribution. Here, the Bernoulli distribution is recognized as the experiment for the combination of U identical subexperiments. For more details, let A be the elementary event having one of the two possible outcomes as its element. \overline{A} is the only other possible elementary event. At this time, we shall repeat the basic experiments U times and determine the probability that A is observed exactly v times out of U trials. Such repeated experiments are called Bernoulli trials [16]. Moveover, in [17], the stability of stochastic difference equations was addressed based on the Lyapunov functionals.
Regarding the CDNs, the coupling strength is the information of coupling strength between agents and a leader. Since an environmental change exists in practical networks, the change of coupling strength is a considerable factor affecting dynamics for the worse of the networks. At this point, the randomly changing coupling strength is being put to use in the problem of synchronization stability for CDNs. Moreover, to the best of authors’ knowledge, the synchronization analysis of CDNs with changing coupling strength has not been formulated yet.
Motivated by the results mentioned above, in this paper, a synchronization stability problem for discretetime CDNs with interval timevarying delays and randomly changing coupling strength will be studied. This information is one of randomly occurring coupling strength with binomial distribution. Put simply, the first and simplest random variable is the Bernoulli random variable. Let X be a random variable that takes on only two possible numerical values, X(\mathrm{\Omega})=\{0,1\}, where Ω represents the universal set consisting of the collection of all objects of interest in a particular context. Multiple independent Bernoulli random variables can be combined to construct more sophisticated random variables. Suppose X is the sum of w independent and identically distributed Bernoulli random variables. Then X is called a binomial random variable with parameters w, the number of trials, and p, the probability of success for each trial. Thus, the binomial distribution is a generalization of the Bernoulli distribution. Also, since delaydependent analysis makes use of the information on the size of time delay, more attention has been paid to the delaydependent analysis than to the delayindependent one [18]. That is, the former is generally less conservative than the latter. Therefore, a great number of results on a delaydependent stability condition for timedelay systems have been reported in the literature [19–24]. So, by construction of a suitable LyapunovKrasovskii functional and utilization of a reciprocally convex approach [24], a synchronization stability problem for discretetime CDNs with interval timevarying delays and randomly changing coupling strength is derived in terms of LMIs which can be solved efficiently by use of standard convex optimization algorithms such as interiorpoint methods [25]. Moreover, the discretetime CDNs are represented by use of the Kronecker product technique. Two numerical examples are included to show the effectiveness of the proposed methods.
Notation {\mathbb{R}}^{n} is the ndimensional Euclidean space, and {\mathbb{R}}^{m\times n} denotes the set of all m\times n real matrices. For real symmetric matrices X and Y, X>Y (respectively, X\ge Y) means that the matrix XY is positive definite (respectively, nonnegative). {X}^{\perp} denotes the basis for the nullspace of X. {I}_{n}, {0}_{n}, and {0}_{m\times n} denote n\times n identity matrix, n\times n and m\times n zero matrices, respectively. \mathbb{E}\{\cdot \} stands for the mathematical expectation operator. \parallel \cdot \parallel refers to the Euclidean vector norm or the induced matrix norm. diag\{\cdots \} denotes the block diagonal matrix. ⋆ represents the elements below the main diagonal of a symmetric matrix. {X}_{[f(t)]}\in {\mathbb{R}}^{m\times n} means that the elements of matrix {X}_{[f(t)]} include the scalar value of f(t).
2 Problem statements
Consider the following discretetime CDNs with interval timevarying delays in the coupling term
Here, N is the number of coupled nodes, n is the number of state of each node, {y}_{i}(k)={[{y}_{i1}(k),{y}_{i2}(k),\dots ,{y}_{in}(k)]}^{T}\in {\mathbb{R}}^{n} is the state vector of the i th node. f({y}_{i}(k))=[f({y}_{i1}(k)),f({y}_{i2}(k)),\dots ,f({y}_{in}(k))]\in {\mathbb{R}}^{n} is a continuous differentiable vector function. The constant c>0 is the coupling strength. h(k) is an interval timevarying delay satisfying 0\le {h}_{m}\le h(k)\le {h}_{M}, where {h}_{m} and {h}_{M} are known positive integers.
\mathrm{\Gamma}={[{\gamma}_{ij}]}_{n\times n} is the innercoupling matrix of nodes, in which {\gamma}_{ij}\ne 0 means two coupled nodes are linked through their i th and j th state variables; otherwise, {\gamma}_{ij}=0. G={[{g}_{ij}]}_{N\times N} is the outercoupling matrix of the network, in which {g}_{ij} is defined as follows. If there is a connection between node i and node j (j\ne i), then {g}_{ij}={g}_{ji}=1; otherwise, {g}_{ij}={g}_{ji}=0 (j\ne i), and the diagonal elements of the matrix G are defined by
In order to investigate the synchronization stability analysis for discretetime CDNs with interval timevarying delays in the coupling term (1), we introduce the following definition and lemmas.
Definition 1 ([5])
The discretetime delayed dynamical network (1) is said to achieve asymptotic synchronization if
where s(k)\in {\mathbb{R}}^{n} is a solution of an isolated node, satisfying s(k+1)=f(s(k),s(kh(k))).
Lemma 1 ([11])
Consider the network (1). Let 0={\lambda}_{1}>{\lambda}_{2}\ge \cdots \ge {\lambda}_{N} be the eigenvalues of the outercoupling matrix G. If the following N1 linear delayed difference equations are asymptotically stable about their zero solution
where J and {J}_{d} are the Jacobian of f(x(k),x(kh(k))) at s(k) and s(kh(k)), respectively. Then the synchronized states (3) are asymptotically stable.
For the convenience of synchronization analysis for the system (4), the following Kronecker product and its properties are used.
Lemma 2 (Kronecker product [26])
Let ⊗ denote the notation of the Kronecker product. Then the following properties of the Kronecker product are easily established:

(i)
(\alpha A)\otimes B=A\otimes (\alpha B),

(ii)
(A+B)\otimes C=A\otimes C+B\otimes C,

(iii)
(A\otimes B)(C\otimes D)=(AC)\otimes (BD).
Let us define
where N is the number of agents.
Then the system (4) can be rewritten in the matrix form
Moreover, it is assumed that the coupling strength has changed by the following assumption.
Assumption 1 The coupling strength is randomly changing. This means that, {\rho}_{m} is a stochastic process representing the information changing process of coupling strength; that is, the transition of coupling strength is described by the following binomial probability:
where m is the number of changes, {\rho}_{0} is the probability of change in one term and {\rho}_{m} satisfies
With Assumption 1, a model of discretetime CDNs with interval timevarying delays in the coupling term and the above assumptions is considered as
Remark 1 As mentioned in Section 1, the Bernoulli random variable takes on only two possible numerical values, X(\mathrm{\Omega})=\{0,1\}, where Ω represents the universal set consisting of the collection of all objects of interest in a particular context. However, with parameters w, the number of trials, and p, the probability of success for each trial, the binomial random variable X is the sum of w independent and identically distributed Bernoulli random variables. Thus, the former is more general than the latter. So, in this paper, the randomly changing coupling strength is considered with the concept of binomial distribution. In addition, the Bernoulli random variable has been used in the concept of randomly occurring which has various types such as randomly occurring delay, randomly occurring uncertainties, randomly occurring nonlinearities, and so on [13–15].
Remark 2 From Assumption 1, the coupling strength per discrete step is changed with the multiple of given strength, c, and the number of changes, m. Also, the probability of change is given as {\rho}_{0}. Based on the results mentioned above, for the concerned discretetime CDNs (1), the coupling strength with the binomial random variable can be represented in the model of discretetime CDNs with randomly changing coupling strength (6). It should be noted that this problem for the change of coupling strength has not been investigated yet.
The aim of this paper is to investigate the delaydependent synchronization stability analysis for the system (6). In order to do this, we introduce the following definition and lemmas.
Lemma 3 For any constant matrix 0<M={M}^{T}\in {\mathbb{R}}^{n\times n}, integers {h}_{m} and {h}_{M} satisfying {h}_{m}\le {h}_{M}, and vector function x:\{{h}_{m},{h}_{m}+1,\dots ,{h}_{M}\}\to {\mathbb{R}}^{n}, the following inequality holds:
Proof From Lemma 1 in [27], the following inequality holds for {h}_{m}\le s\le {h}_{M}:
Sum of the inequality (8) from {h}_{m} to {h}_{M} yields
Therefore, the inequality (9) is equivalent to the inequality (7) according to the Schur complement [25]. □
Lemma 4 (Finsler’s lemma [28])
Let \zeta \in {\mathbb{R}}^{n}, \mathrm{\Phi}={\mathrm{\Phi}}^{T}\in {\mathbb{R}}^{n\times n}, and \mathrm{\Upsilon}\in {\mathbb{R}}^{m\times n} such that rank(\mathrm{\Upsilon})<n. The following statements are equivalent:

(i)
{\zeta}^{T}\mathrm{\Phi}\zeta <0, \mathrm{\forall}\mathrm{\Upsilon}\zeta =0, \zeta \ne 0,

(ii)
{{\mathrm{\Upsilon}}^{\perp}}^{T}\mathrm{\Phi}{\mathrm{\Upsilon}}^{\perp}<0,

(iii)
\mathrm{\exists}F\in {\mathbb{R}}^{n\times m}:\mathrm{\Phi}+F\mathrm{\Upsilon}+{(F\mathrm{\Upsilon})}^{T}<0.
3 Main results
In this section, we propose new synchronization criteria for the system (6). For the sake of simplicity on matrix representation, {e}_{i}\in {\mathbb{R}}^{5\kappa \times \kappa}, where \kappa =(N1)n, are defined as block entry matrices, e.g., {e}_{2}={[{0}_{\kappa},{I}_{\kappa},{0}_{\kappa},{0}_{\kappa},{0}_{\kappa}]}^{T}. The notations of several matrices are defined as:
Now, the following theorem is given for synchronization stability of the model of discretetime CDNs with interval timevarying delays in the coupling term (6).
Theorem 1 For given positive integers {h}_{m}, {h}_{M}, l and positive scalars c, {\rho}_{0}<1, the system (6) is asymptotically synchronous for {h}_{m}\le h(k)\le {h}_{M}, if there exist positive matrices P\in {\mathbb{R}}^{n\times n}, {Q}_{i}\in {\mathbb{R}}^{n\times n}, {R}_{j}\in {\mathbb{R}}^{n\times n}, any symmetric matrices {S}_{i}\in {\mathbb{R}}^{n\times n}, where i=1,2 and j=1,\dots ,4, and any matrix M\in {\mathbb{R}}^{n\times n} satisfying the following LMIs:
where Φ and {\mathrm{\Upsilon}}_{[{\rho}_{m}]} are defined in (10).
Proof Let us consider the following LyapunovKrasovskii functional candidate as
where
The mathematical expectation of the \mathrm{\Delta}{V}_{1}(k) and \mathrm{\Delta}{V}_{2}(k) are calculated as
By calculating the \mathbb{E}\{\mathrm{\Delta}{V}_{3}(k)\}, we get
Inspired by the work of [29], the following two zero equalities hold with any symmetric matrices {S}_{1} and {S}_{2}:
and
Here, Eqs. (18) and (19) still hold even when we multiply both sides by ({h}_{M}{h}_{m}). So, by adding the results into Eq. (17), we get
where
By Lemma 3, the term Σ in (20) can be estimated as
where \alpha (k)=({h}_{M}h(k))/({h}_{M}{h}_{m}).
Here, when {h}_{m}<h(t)<{h}_{M}, since \alpha (k) satisfies 0<\alpha (t)<1, by a reciprocally convex approach [24], the following inequality for any matrix M holds:
which implies
Also, when h(k)={h}_{m} or h(k)={h}_{M}, we get
or
respectively.
Thus, if the inequality (12) holds, then from Eqs. (22) and (23), the following inequality still holds:
which means
Lastly, the \mathbb{E}\{{V}_{4}(k)\} is calculated as
where
Furthermore, if the inequalities (13) hold, then the \mathbb{E}\{\mathrm{\Delta}{V}_{3}(k)\}+\mathbb{E}\{\mathrm{\Delta}{V}_{4}(k)\} has an upper bound as follows:
Therefore, from Eqs. (15)(26) and by application of the Sprocedure [25], the mathematical expectation on \mathrm{\Delta}V(k) has a new upper bound as
Also, the system (6) with the augmented vector \zeta (k) can be rewritten as
where {\mathrm{\Upsilon}}_{[{\rho}_{m}]} is defined in (10).
Then a delaydependent stability condition for the system (6) is
subject to
From Lemma 4(iii) and Assumption 1, the inequality (29) is equivalent to the following condition:
where F is any matrix with appropriate dimension.
Here, by utilizing Lemma 4(ii), the condition (30) is equivalent to the following inequality:
From the inequality (31), if the LMIs (11)(13) are satisfied, then the synchronization stability condition (29) holds by Definition 1. This completes our proof. □
As a special case, consider the following discretetime CDNs with only interval timevarying delays in nodes and randomly changing coupling strength:
By use of the similar method in the driven procedure of the model (6), a model of discretetime CDNs (32) can be obtained as
The following is given for synchronization stability of the model of discretetime CDNs with only interval timevarying delays in nodes (33).
Theorem 2 For given positive integers {h}_{m}, {h}_{M}, l, and positive scalars c, {\rho}_{0}<1, the system (33) is asymptotically synchronous for {h}_{m}\le h(k)\le {h}_{M}, if there exist positive matrices P\in {\mathbb{R}}^{n\times n}, {Q}_{i}\in {\mathbb{R}}^{n\times n}, {R}_{j}\in {\mathbb{R}}^{n\times n}, any symmetric matrices {S}_{i}\in {\mathbb{R}}^{n\times n}, i=1,2, j=1,\dots ,4, and any matrix M\in {\mathbb{R}}^{n\times n} satisfying the following LMIs with (12) and (13):
Proof The above criterion is derived in the similar method as the proof of Theorem 1, instead of the matrix {\mathrm{\Upsilon}}_{[{\rho}_{m}]}, using the following matrix:
The other procedure is straightforward from the proof of Theorem 1, so it is omitted. □
Remark 3 The systems (6) and (33) with randomly changing coupling strength and the switched systems [30–38] are similar in the concept of changing parameters. In [30–38], the various problems for the switched neural networks with timeinvariant delay were addressed. However, since time delay has not only a fixed value in a practical system [39], the concerned systems with interval timevarying delays were considered in this paper. Moreover, the changing information of a parameter was considered with the probabilistic rule; that is, the Bernoulli sequence.
4 Numerical examples
In this section, we provide two numerical examples to show the effectiveness of the presented stability criteria in this paper.
Example 1 Consider the following 2order system with the structure in Figure 1 and the innercoupling matrix \mathrm{\Gamma}=0.01{I}_{n}:
which is asymptotically stable at the equilibrium point s(k)=0 and s(kh(k))=0. To analyse the synchronization stability for randomly changing coupling strength, the N1 linear delayed difference equations (6) are
with the Jacobian matrices
From Figure 1, the outercoupling matrices are considered as two cases.

Case 1:
G=\left[\begin{array}{ccccc}2& 1& 0& 0& 1\\ 1& 3& 1& 1& 0\\ 0& 1& 2& 1& 0\\ 0& 1& 1& 3& 1\\ 1& 0& 0& 1& 2\end{array}\right], 
Case 2:
G=\left[\begin{array}{cccccccccc}4& 1& 0& 0& 1& 1& 1& 0& 0& 0\\ 1& 3& 0& 0& 0& 0& 0& 1& 1& 0\\ 0& 0& 1& 0& 0& 0& 0& 1& 0& 0\\ 0& 0& 0& 2& 0& 1& 0& 1& 0& 0\\ 1& 0& 0& 0& 2& 0& 0& 0& 1& 0\\ 1& 0& 0& 1& 0& 2& 0& 0& 0& 0\\ 1& 0& 0& 0& 0& 0& 3& 0& 1& 1\\ 0& 1& 1& 1& 0& 0& 0& 3& 0& 0\\ 0& 1& 0& 0& 1& 0& 1& 0& 4& 1\\ 0& 0& 0& 0& 0& 0& 1& 0& 1& 2\end{array}\right].
Moreover, from Case 1,
and from Case 2,
The result of the maximum bound of timedelay with fixed c=0.5, l=10, {\rho}_{0}=0.7, {h}_{m}=1 and G in Case 1 provided by Theorem 1 is 5. For Case 1, Figure 2 shows the simulation results for the state trajectories of the network (36) with h(k)=4sin(k\pi /2)+1. When the coupling strength is changeless, the size of overshoot shown in Figure 2(b) is smaller than the one shown in Figure 2(a) at time 10\sim 30\text{[sec]}. Here, if we consider the setting time and rise time on the basis of Figure 2(b), which are the results of an ideal model for CDNs without the changing coupling strength, we know the effect of the change of coupling strength. Next, the result of maximum bound of timedelay with fixed c=1, {\rho}_{0}=0.2, {h}_{m}=5, and G in Case 2 by Theorem 1 is 7. With the condition of timevarying delay as h(k)=2sin(k\pi /2)+5, the simulation results for the state trajectories of the network (36) are shown in Figure 3. Also, this figure shows that the trajectories between the synchronized states converge to zero under the timedelay h(k). Lastly, in Figure 4, the distribution of a binomial random variable is drawn with the values of probability, {\rho}_{0}, 0.7 and 0.2.
Example 2 Recall the system (35) in Example 1 and the structure shown in Figure 1. Thus, consider the following coupled networks with only interval timevarying delay in nodes
where the associated parameters are defined in Example 1.
The simulation results for the state trajectories of the network (37) and the curve of mode with the probability {\rho}_{0}=0.7 are shown in Figure 5 with the condition of timevarying delay as h(k)=13sin(k\pi /2)+5 (c=1, l=10, {\rho}_{0}=0.7 and {h}_{m}=5). At this time, by Theorem 2, the maximum bound of timedelay is 18.
5 Conclusions
In this paper, new delaydependent synchronization criteria for the discretetime CDNs with interval timevarying delays and randomly changing coupling strength are proposed. The randomly changing coupling strength is considered with the concept of binomial distribution, which is a generalization of the Bernoulli distribution. To drive these results, the suitable LyapunovKrasovskii functional and reciprocally convex approach are used to obtain the feasible region of synchronization stability criteria. Two numerical examples have been given to show the effectiveness and usefulness of the presented criteria.
References
 1.
Watts DJ, Strogatz SH: Collective dynamics of ‘smallworld’ networks. Nature 1998, 393: 440–442. 10.1038/30918
 2.
Strogatz SH: Exploring complex networks. Nature 2001, 410: 268–276. 10.1038/35065725
 3.
Li C, Xu C, Sun W, Xu J, Kurths J: Outer synchronization of coupled discretetime networks. Chaos 2009., 19: Article ID 013106
 4.
Boccaletti S, Latora V, Moreno Y, Chavez M, Hwang DU: Complex networks: structure and dynamics. Phys. Rep. 2006, 424: 175–308. 10.1016/j.physrep.2005.10.009
 5.
Li C, Chen G: Synchronization in general complex dynamical networks with coupling delays. Physica A 2004, 343: 263–278.
 6.
Li K, Gong X, Guan S, Lai CH: Synchronization stability of general complex dynamical networks with timevarying delays. Phys. Lett. A 2008, 372: 7133–7139. 10.1016/j.physleta.2008.10.054
 7.
Wang B, Guan ZH: Chaos synchronization in general complex dynamical networks with coupling delays. Nonlinear Anal., Real World Appl. 2010, 11: 1925–1932. 10.1016/j.nonrwa.2009.04.020
 8.
Park MJ, Kwon OM, Park JH, Lee SM, Cha EJ: Synchronization criteria for coupled stochastic neural networks with timevarying delays and leakage delay. J. Franklin Inst. 2012, 349: 1699–1720. 10.1016/j.jfranklin.2012.02.002
 9.
Koo JH, Ji DH, Won SC: Synchronization of singular complex dynamical networks with timevarying delays. Appl. Math. Comput. 2010, 217: 3916–3923. 10.1016/j.amc.2010.09.055
 10.
Gao H, Lam J, Chen G: New criteria for synchronization stability of general complex dynamical networks with coupling delays. Phys. Lett. A 2006, 360: 263–273. 10.1016/j.physleta.2006.08.033
 11.
Yue D, Li H: Synchronization stability of continuous/discrete complex dynamical networks with interval timevarying delays. Neurocomputing 2010, 73: 809–819. 10.1016/j.neucom.2009.10.008
 12.
Liu Y, Wang Z, Liu X, Liang J: Synchronization and state estimation for discretetime complex networks with distributed delays. IEEE Trans. Syst. Man Cybern., Part B, Cybern. 2008, 38: 1314–1325.
 13.
Bao H, Cao J: Synchronization of discretetime stochastic neural networks with random delay. Discrete Dyn. Nat. Soc. 2011. doi:10.1155/2011/713502
 14.
Wu ZG, Park JH, Su H, Chu J: Robust dissipativity analysis of neural networks with timevarying delay and randomly occurring uncertainties. Nonlinear Dyn. 2012, 69: 1323–1332. 10.1007/s1107101203501
 15.
Hu J, Wang Z, Gao H, Stergioulas LK: Robust sliding mode control for discrete stochastic systems with mixed time delays, randomly occurring uncertainties, and randomly occurring nonlinearities. IEEE Trans. Ind. Electron. 2012, 59: 3008–3015.
 16.
Peebels PZ: Probability, Random, Variables, and Random Signal Principles. McGrawHill, New York; 2001.
 17.
Shaikhet L: Lyapunov Functionals and Stability of Stochastic Difference Equations. Springer, New York; 2011.
 18.
Xu S, Lam J: A survey of linear matrix inequality techniques in stability analysis of delay systems. Int. J. Syst. Sci. 2008, 39: 1095–1113. 10.1080/00207720802300370
 19.
Niculescu SI: Delay Effects on Stability: A Robust Approach. Springer, New York; 2002.
 20.
Richard JP: Timedelay systems: an overview of some recent advances and open problems. Automatica 2003, 39: 1667–1694. 10.1016/S00051098(03)001675
 21.
Gao HJ, Chen TW: New results on stability of discretetime systems with timevarying state delay. IEEE Trans. Autom. Control 2007, 52: 328–334.
 22.
Ahn CK: Fuzzy delayed output feedback synchronization for timedelayed chaotic systems. Nonlinear Anal. Hybrid Syst. 2010, 4: 16–24. 10.1016/j.nahs.2009.07.002
 23.
Kwon OM, Park JH, Lee SM: An improved delaydependent criterion for asymptotic stability of uncertain dynamic systems with timevarying delays. J. Optim. Theory Appl. 2010, 145: 343–353. 10.1007/s109570099637x
 24.
Park P, Ko JW, Jeong CK: Reciprocally convex approach to stability of systems with timevarying delays. Automatica 2011, 47: 235–238. 10.1016/j.automatica.2010.10.014
 25.
Boyd S, Ghaoui LE, Feron E, Balakrishnan V: Linear Matrix Inequalities in System and Control Theory. SIAM, Philadelphia; 1994.
 26.
Graham A: Kronecker Products and Matrix Calculus: With Applications. Wiley, New York; 1982.
 27.
Gu K: An integral inequality in the stability problem of timedelay systems. The 39th IEEE Conf. Decision Control 2000, 2805–2810.
 28.
de Oliveira MC, Skelton RE: Stability tests for constrained linear systems. In Perspectives in Robust Control. Springer, Berlin; 2001:241–257.
 29.
Kim SH:Improved approach to robust {H}_{\mathrm{\infty}} stabilization of discretetime TS fuzzy systems with timevarying delays. IEEE Trans. Fuzzy Syst. 2010, 18: 1008–1015.
 30.
Ahn CK: Sets of generalized H2 exponential stability criteria for switched multilayer dynamic neural networks. Adv. Differ. Equ. 2012. doi:10.1186/1687–1847–2012–150
 31.
Ahn CK: Linear matrix inequality optimization approach to exponential robust filtering for switched Hopfield neural networks. J. Optim. Theory Appl. 2012, 154: 573–587. 10.1007/s1095701200087
 32.
Ahn CK: An error passivation approach to filtering for switched neural networks with noise disturbance. Neural Comput. Appl. 2012, 21: 853–861. 10.1007/s0052101004745
 33.
Ahn CK: Switched exponential state estimation of neural networks based on passivity theory. Nonlinear Dyn. 2012, 67: 573–586. 10.1007/s110710110010x
 34.
Ahn CK: Exponentially convergent state estimation for delayed switched recurrent neural networks. Eur. Phys. J. E 2011, 34: 122.
 35.
Ahn CK:{L}_{2}{L}_{\mathrm{\infty}} filtering for timedelayed switched Hopfield neural networks. Int. J. Innov. Comput. Inf. Control 2011, 7: 1831–1844.
 36.
Ahn CK:Switched synchronization with a guaranteed{H}_{\mathrm{\infty}} performance. Chin. Phys. Lett. 2011., 28: Article ID 010501
 37.
Ahn CK: Passive learning and inputtostate stability of switched Hopfield neural networks with timedelay. Inf. Sci. 2010, 180: 4582–4594. 10.1016/j.ins.2010.08.014
 38.
Ahn CK:An {H}_{\mathrm{\infty}} approach to stability analysis of switched Hopfield neural networks with timedelay. Nonlinear Dyn. 2010, 60: 703–711. 10.1007/s1107100996256
 39.
Jiang L, Yao W, Wu QH, Wen JY, Cheng SJ: Delaydependent stability for load frequency control with constant and timevarying delays. IEEE Trans. Power Syst. 2012, 27: 932–941.
Acknowledgements
This research was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology (20120000479), and by a grant of the Korea Healthcare Technology R & D Project, Ministry of Health & Welfare, Republic of Korea (A100054).
Author information
Affiliations
Corresponding author
Additional information
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
All authors contributed equally and significantly in writing this paper. All authors read and approved the final manuscript.
Authors’ original submitted files for images
Below are the links to the authors’ original submitted files for images.
Rights and permissions
Open Access This article is distributed under the terms of the Creative Commons Attribution 2.0 International License ( https://creativecommons.org/licenses/by/2.0 ), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
About this article
Cite this article
Park, M., Kwon, O., Park, J.H. et al. Synchronization stability of delayed discretetime complex dynamical networks with randomly changing coupling strength. Adv Differ Equ 2012, 208 (2012). https://doi.org/10.1186/168718472012208
Received:
Accepted:
Published:
DOI: https://doi.org/10.1186/168718472012208
Keywords
 Coupling Strength
 Bernoulli Random Variable
 Synchronization Criterion
 Binomial Random Variable
 Synchronization Analysis