In this paper, exponential synchronization problem of complex dynamical networks with unknown periodically coupling strengths was investigated. An aperiodically intermittent control synchronization strategy is proposed. Based on Lyapunov exponential stability theory, inequality techniques, and adaptive learning laws design, some sufficient exponential synchronization criteria for complex dynamical network with unknown periodical coupling weights are obtained. The numerical simulation example is presented to illustrate the feasibility of theoretical results.
National Natural Science Foundation of China6157301311705122Natural Science Funds of Science and Technology Department of Shaanxi Province2014JM2-1002nature scientific research project of Shaanxi Education Department18JK0829Natural Science Foundation of Xianyang Normal University14XSYK0041. Introduction
Complex dynamical networks, such as biological network, urban transportation network and human relationship network, are ubiquitous in natural world and social society generally. Researchers were mainly concerned with the issues of modelling, properties analysis, dynamical evolution, and synchronization control of complex dynamical networks. Among these issues involved, synchronization is one of the most interesting topics and has been extensively investigated [1–12]. There are mainly two kinds of strategies for synchronization of complex dynamical networks: one is to improve the network synchronization capability by changing the properties of the network itself, such as topology structure and coupling strengths. The other is to act on the network with external control injection, which is a representative of control theory, mainly including variable feedback control method, pinning control method [2, 4], adaptive control method [5], event-triggered control [8], impulse control method, intermittent pinning control [9], and slide mode control method [10, 11].
We can see that most of the above control methods are continuous, which requires continuous information exchange and increases the cost of control. However, the discontinuous control method such as intermittent control, exerting intermittent control over the controlled objects, can reduce the amount of the transmitted information and be much more economic. Zochowski firstly introduced intermittent control into dynamic control systems [12]. After that, many researchers pay attention to this control strategy and many large-scale dynamical systems are controlled successfully with the help of intermittent control. So, it can also be applied to realize synchronization of complex dynamical networks [13–23].
Over the first few years, the intermittent controller which can help realizing synchronization for complex dynamical networks applied is usually periodical. In papers [13–16], the authors investigated finite-time synchronization of complex dynamical networks with multilink, time delay, and multiswitch periods using periodically intermittent control scheme. Zhao and Cai [17] investigated the exponential synchronization of complex delayed dynamical networks with uncertain parameters adopting the intermittent control scheme, basing on the Lyapunov stability theory combined with the method of the adaptive control. Then semiperiodically intermittent control method was proposed, by applying multiple Lyapunov function method and the mode-dependent average dwell time approach. Qiu [18] derived less conservative synchronization criteria and the synchronization problem for switched complex networks with delayed coupling via semiperiodically intermittent control technique and a mode-dependent average dwell time method was studied. Liu [19] addressed synchronization problems of time-delay coupled network by aperiodically intermittent control. In 2017, with the help of aperiodically intermittent control method, Liu [20] discussed finite-time synchronization of delayed dynamical networks and showed the convergence time does not depend on control widths or rest widths. Combining intermittent control with pinning control scheme, which can greatly reduce control cost, researchers presented some conclusions about intermittent pinning control. Wang [21] proposed a new differential inequality, dealing with the synchronization problem of complex dynamical network with constant coupling, discrete-delay coupling, and distributed-delay coupling. The synchronization problem of switched complex networks with unstable modes and stochastic complex-valued dynamical networks by aperiodically intermittent adaptive control are discussed in [22, 23], respectively.
It is noteworthy that the above aperiodically intermittent adaptive control results can ensure all nodes of complex dynamical network realizing exponential synchronization, which can make the system tend to equilibrium state much more quickly compared with the asymptotic synchronization of networks [24]. In [25], the exponential synchronization problem of a class of hybrid coupled complex dynamical network with time-varying delay is studied. By designing appropriate intermittent feedback controller, a new synchronization criterion is presented. In paper [26], a kind of complex network model with directional topology and time-delay is studied, and sufficient conditions for global exponential synchronization of the system are obtained. Zhang et al. [27] investigated exponential stability of stochastic differential equations with impulse effects. As seen from the above literatures, exponential synchronization of complex dynamical network via aperiodically intermittent control could save the control cost, which is significant in practical applications.
What is more, with the research of complex dynamic network control, researchers found that the differences of nodes and coupling modes play an important role for the evolution of complex dynamical networks. In addition, many practical complex dynamical networks are time-varying generally, in which, the changes of link strengths usually lead to variations of the network topology and coupling configuration [16, 17]. Therefore, it is meaningful to investigate the synchronization problem of complex dynamical networks with unknown time-varying coupling weights. In order to find the time-varying law of the complex dynamical networks, researchers usually use adaptive methods to estimate the coupling strengths of the network [21]. In paper [5], the author analyzed mean square synchronization of time-delay coupled network, and the unknown periodical coupling weight was estimated successfully with the adaptive leaning update laws and proper controller designed. In 2015, Hao and Li [28] combined adaptive control with learning control to estimate the network unknown periodic coupling structure under the stochastic disturbance successfully. However, the controllers designed in [5, 28] are all continuous, the discontinuous method such as intermittent control was not taken into account. Not only that, but also the intrinsic time delay is ubiquitous in neural networks [29], chaotic attractors [30], and even complex dynamical networks [31], which will affect the synchronization performance. So the synchronization of complex dynamical network with time delay is needed to be investigated.
Motivated by the above discussions, we will consider the exponential synchronization problem for time-delay complex dynamical networks with unknown time-varying coupling weights by using an adaptive aperiodically intermittent controller in this paper. The main contribution of our paper is that, combining the advantages of aperiodic intermittent control, adaptive control, and learning control, a new type of aperiodically intermittent synchronization of complex dynamical network is accomplished when the time varying coupling weights among nodes are unknown primitively. The rest of this paper is organized as follows. A new complex dynamical network model is presented, and the problem formulation and preliminaries are presented in Section 2. Section 3 proposes the aperiodically intermittent synchronization approach for complex dynamical networks. In Section 4, a numerical example is given to illustrate the effectiveness of the designed method. Finally, conclusions are presented in Section 5.
2. Problem Statement and Preliminaries
In the paper, the time-delay complex dynamical network with unknown periodical outer coupling strengths is given as follows:(1)x˙it=ft,xit,xit-τ0t+∑j=1NcijtaijΓxjt+ui,i=1,2,…,N.where xi(t)=(xi1(t),xi2(t),…,xin(t))T∈Rn is the state variable of node i and f:R+×Rn×Rn→Rn,i=1,2,…,N is a smooth nonlinear function, describing the local dynamics of each node for network (1). τ0(t) is the unknown bounded time-varying delay, and there exists a positive constant τ0 such that 0≤τ0(t)≤τ0. The positive definite Γ∈Rn×n is the inner coupling matrix. ui(t) is the outer controller to be designed later. The matrix A=(aij)N×N∈RN×N is the outer coupling matrix, which describes the topology structure of the whole network. And it is defined as follows: if there is a directional connection from node j to node i(i≠j), then aij≠0; otherwise aij=0. cij(t) is the unknown periodical coupling strengths between nodes i and j. These elements satisfy the following condition: (2)ciitaii=-∑j=1j≠iNcijtaij,i=1,2,…,N.
The initial conditions of network (1) are assumed to be xi(t)=ψi(t)∈C([-τ0,0],Rn), where C([-τ0,0],Rn) represents the set of all n-dimensional continuous functions defined on interval [-τ0,0].
Without loss of generality, let the solution s(t)∈Rn of (3) be the global exponential synchronization goal orbit:(3)s˙t=ft,st,st-τ0t.
Let ei(t)=xi(t)-s(t); then the error system is(4)e˙it=x˙it-s˙t,i=1,2,…,N.
In the paper, we will add proper designed controllers ui(t),i=1,2,…,N intermittently to the dynamical network (1) to realize globally exponential synchronization.
Definition 1.
The complex dynamical network (1) is said to be globally exponentially synchronized, if there exist two positive constants p and q such that, for any initial state xi(t)=ψi(t)∈C([-τ0,0],Rn), the synchronous error satisfies(5)et≤pe-qt,∀t∈-τ0,+∞.where e(t)=(e1(t),e2(t),…,eN(t))T.
For more discussion, the following assumptions and lemmas are needed to be introduced firstly.
Assumption 2 (see [21]).
For the vector-valued function f(t,x(t),x(t-τ0(t))), suppose the uniform semi-Lipschitz condition with respect to the time t holds; i.e., for any x∈Rn,y∈Rn, there exists two positive constants l¯1 and l¯2, such that(6)xt-ytTft,xt,xt-τ0t-ft,yt,yt-τ0t≤l¯1xt-ytTxt-yt+l¯2xt-τ0t-yt-τ0tTxt-τ0t-yt-τ0t.
Assumption 3.
In network (1), the unknown time-varying coupling strengths cij(t) are periodical parameters; that is, cij(t+T)=cij(t) for t∈[0,+∞), in which T is the known common period of cij(t).
For further discussion, the aperiodically intermittent control strategy can be expressed as follows. Each controlling cycle [wi,wi+1) usually contains two types of time zones, one is working time [wi,si) and the other is rest time [si,wi+1). The controller is activated at each working time and closed at rest time.
As shown in Figure 1, the solid line represents the working time, while the dotted line represents the rest time.
Aperiodical intermittent control sketch.
Assumption 4 (see [32]).
For the aperiodically intermittent control strategy, there exist two positive scalars 0<θ<ω<+∞, such that, for i=0,1,2,⋯,(7)infisi-wi=θ,supiwi+1-wi=ω.
Define the maximum proportion of rest width wi+1-si in the time span wi+1-wi as(8)ϕ=limsupi→+∞wi+1-siwi+1-wi.
Lemma 5 (see [32]).
For any i=0,1,2,…, if we denote(9)ϕt=t-sit-wi,t∈si,wi+1.Then ϕ(t) is an strictly increasing function, so ϕ(t)≤ϕ(wi+1)=wi+1-si/wi+1-wi.
Lemma 6 (see [32, 33]).
Suppose that function y(t) is continuous and nonnegative for t→[-τ,+∞) and satisfies the following condition:(10)y˙t≤-γ1yt+γ2supt-τ≤s≤tys,wi≤t≤si,y˙t≤γ3yt+γ4supt-τ≤s≤tys,si≤t≤wi+1.where γ1,γ2,γ3,γ4 are positive constants and i=0,1,2,⋯. Suppose that, for the aperiodically intermittent control, there exists a constant ϕ defined in (8). If(11)γ1>γ∗=maxγ2,γ4>0,ρ=γ1+γ3>0,ω=λ-ρϕ>0,then(12)yt≤sup-τ≤s≤0ysexp-ωt,t≥0,where λ>0 is the unique positive solution of the equation λ-γ1+γ∗expλτ=0.
3. Aperiodically Intermittent Synchronization for Complex Dynamical Networks
In this section, we consider the synchronization problem of coupled complex dynamical network (1) via aperiodically intermittent control. In order to achieve the synchronization objective (3), we choose the adaptive controller ui(t) for the i-th node as follows:(13)uit=-kitΓeit-∑j=1Nc^ijtaijΓejt,t∈wi,si,0,t∈si,wi+1,1≤i≤N.The update law of the adaptive parameter ki(t),i=1,2,…,N is(14)k˙it=αiexpβ1teiTtΓeit,t∈wi,si,0,t∈si,wi+1.where αi and β1 are positive constants.
For the unknown time-varying periodical coupling strengths cij(t),i,j=1,2,…,N, the parameters estimation are designed by(15)c^ijt=c^ijt-T+ϱij∗aijeiTtΓejt,t∈kT,k+1T,k=1,2,⋯ϱijtaijeiTtΓejt,t∈0,T0,t∈-T,0.where c^ij(t) is the estimation of coupling strength cij(t). Denote c~ij(t)=cij(t)-c^ij(t) is the estimation error. ϱij∗ are positive constants, ϱij(t) is a continuous and strictly increasing function for t∈[0,T] and satisfies ϱij(0)=0,ϱij(T)=ϱij∗.
Basing on the designed controllers, the following synchronization error system can be obtained:(16)e˙it=f¯t,eit,eit-τ0t+∑j=1Nc~ijtaijΓejt-kitΓeit,t∈wi,si,f¯t,eit,eit-τ0t+∑j=1NcijtaijΓejt,t∈si,wi+1,1≤i≤N.where f¯(t,ei(t),ei(t-τ0(t)))=f(t,xi(t),xi(t-τ0(t)))-f(t,s(t),s(t-τ0(t))).
Obviously, we can see that exponential synchronization of complex dynamical network (1) with controller ui equals the exponential stability of error system (16). So, next, a sufficient condition for the controlled complex network realizing global exponential synchronization will be presented as follows.
Theorem 7.
Suppose that Assumptions 2–4 hold. If there exist positive constants β1 and β2(β2>β1), such that the following conditions hold,(17)η2+λmaxA^Cs<0,(18)β1-2l2>0,(19)ξ-β2ϕ>0.where ξ>0 is the unique positive solution of equation ξ-β1+lexpξτ0=0,l=maxl2,l3, then the controlled network (1) with adaptive periodical outer coupling update laws (15) is globally exponentially synchronized under the adaptive aperiodically intermittent controllers (13) and (14).
Proof.
Construct the Lyapunov-Krasovskii-like function candidate as follows:(20)Vt=12∑i=1NeiTteit+12∑i=1N∑j=1N1ϱij∗∫t-Ttc~ij2τdτ+12∑i=1Nexp-β1tkit-ki∗2αi,where t≥T, ki∗ are some sufficiently large positive constants which will be determined later.
When t∈[wi,si), taking the derivative along the trajectories of the error system, we get(21)V˙t=∑i=1NeiTte˙it+12∑i=1N∑j=1N1ϱij∗c~ij2t-c~ij2t-T-β12∑i=1Nexp-β1tkit-ki∗2αi+∑i=1Nexp-β1tkit-ki∗αi·k˙it=∑i=1NeiTtf¯t,eit,eit-τ0t+∑j=1Nc~ijtaijΓejt-β12∑i=1Nexp-β1tkit-ki∗2αi+∑i=1Nexp-β1tkit-ki∗αi·k˙it+12∑i=1N∑j=1N1ϱij∗c~ij2t-c~ij2t-T-∑i=1NeiTtkitΓeit.
Next we will calculate the items separately. Combining with Assumption 2, we can have(22)eiTtf¯t,eit,eit-τ0t≤l¯1eTtet+l¯2eTt-τ0tet-τ0t.
Considering the adaptive law (14), one can have(23)∑i=1Nexp-β1tkit-ki∗αi·k˙i=∑i=1Nkit-ki∗eiTtΓeit.
Using the matrix equality (a-b)TH(a-b)-(a-c)TH(a-c)=(c-b)TH[2(a-b)+(b-c)] and the estimation of unknown coupling item (15), we can obtain(24)∑i=1N∑j=1N1ϱij∗c~ij2t-c~ij2t-T=-∑i=1N∑j=1N1ϱij∗c^ijt-c^ijt-T2-2∑i=1N∑j=1N1ϱij∗cijt-c^ijt-T·c^ijt-c^ijt-T=-2∑i=1N∑j=1Nc~ijtaijeiTtΓejt-∑i=1N∑j=1N1ϱij∗c^ijt-c^ijt-T2≤-2∑i=1N∑j=1Nc~ijtaijeiTtΓejt.
With the above results and properties of Kronecker production, substitute (22), (23), and (24) into (21), one gets(25)V˙t≤l¯1eTtet+l¯2eTt-τ0tet-τ0t-∑i=1NeiTtki∗Γeit-β12∑i=1Nexp-β1tkit-ki∗2αi≤η1eTtIN⊗Γet+l¯2eTt-τ0tet-τ0t-eTtK∗⊗Γet-β12∑i=1NeiTteit+∑i=1Nexp-β1tkit-ki∗2αi≤eTtR-K∗⊗Γet+l¯2eTt-τ0tet-τ0t+β12∑i=1N∑j=1N1ϱij∗∫t-Ttc~ij2τdτ-β12∑i=1NeiTteit+∑i=1Nexp-β1tkit-ki∗2αi+∑i=1N∑j=1N1ϱij∗∫t-Ttc~ij2τdτ,where the vectors e(t)=(e1(t),e2(t),…,eN(t))T, e(t-τ0(t))=(e1(t-τ0(t)),e2(t-τ0(t)),…,eN(t-τ0(t)))T, η1=l¯1/λmin(Γ)+β1/2,R=η1IN,K∗=diagk1∗,⋯,kN∗.
So we can see that if we choose ki∗>η1, the matrix R-K∗<0. Let l2=max2l¯2,β1; then we can acquire(26)V˙t≤-β1Vt+l2supt-τ0≤ς≤tVς.
When t∈[si,wi+1), taking the derivative along the trajectories of the error system (16), we have(27)V˙t≤l¯1eTtet+l¯2eTt-τ0tet-τ0t+∑i=1N∑j=1Nc^ijteiTtaijΓejt-β12∑i=1Nexp-β1tkit-ki∗2αi≤η2eTtIN⊗Γet+l¯2eTt-τ0tet-τ0t+eTtA^Cs⊗Γet+β2-β12∑i=1NeiTteit+∑i=1Nexp-β1tkit-ki∗2αi≤eTtR~⊗Γet+l¯2eTt-τ0tet-τ0t+β2-β12∑i=1NeiTteit+∑i=1lexp-β1tkit-ki∗2αi+∑i=1N∑j=1N1ϱij∗∫t-Ttc~ij2τdτ,where R~=η2IN+A^Cs, A^Cs=A^C+A^CT/2,A^C=cijt·aijN×N, η2=l¯1/λmin(Γ)-β2-β1/2. According to Lemma 6, it follows that λmax(R~)≤η2+λmax(A^Cs). If condition (17) is satisfied, it is easy to have R~<0. Let l3=2l¯2. Then we can have(28)V˙t≤β2-β1Vt+l3supt-τ0≤ς≤tVς.Thus, we obtain(29)V˙t≤-β1Vt+l2supt-τ0≤ς≤tVς,t∈wi,si,V˙t≤β2-β1Vt+l3supt-τ0≤ς≤tVς,t∈si,wi+1.
Combining Lemma 6 with conditions (18) and (19), we can get V(t)≤supt-τ0≤ς≤tVςexp-ηt for t≥0 immediately, where η=ξ-(l+β2-β1)ϕ and the parameter ξ satisfies the equation ξ-β1+lexpξτ0=0. So the zero solution of the error system (15) is globally exponentially stable. According to the calculation of V(t), we can obtain et2≤2V(t); i.e., e(t)≤2Vt1/2. And hence e(t)≤2supt-τ0≤ς≤tVςexp(-η/2t) holds, which implies the global exponential synchronization is achieved. Thus the proof is completed.
If the complex dynamical network (1) has no any internal time delay, i.e., τ0(t)=0, the system (1) can be rewritten as(30)x˙it=ft,xit+∑j=1NcijtaijΓxjt+ui,i=1,2,…,N.and the synchronization target orbit is described by s˙(t)=f(t,s(t)). In this case, we can see that l¯2=0 holds. By using a similar line of arguments as that in Theorem 7, the following result is easily achieved.
Corollary 8.
Suppose that Assumptions 2–4 hold. If there exist positive constants β1 and β2(β2>β1), such that the following conditions hold,(31)η2+λmaxA^Cs<0,(32)ξ-β2ϕ>0,then the controlled network (1) with adaptive periodical outer coupling update laws (15) is globally exponential synchronized under the adaptive aperiodically intermittent controllers (13) and (14).
Remark 9.
When l¯2=0, the equation is ξ-β1+l3expξτ0=0, so inequality (19) can be expressed as (32).
4. Numerical Simulations
In this section, we will illustrate the effectiveness of the proposed approach to achieve globally exponential synchronization of complex networks (1) with unknown time-varying coupling strengths via aperiodically intermittent control. Without loss of generality, we choose the delayed Chua’s chaotic models as the nodes’ dynamics. The chaotic Chua’s system dynamical function is as follows:(33)x˙1tx˙2tx˙3t=-α1+m2α01-110-β-ωx1tx2tx3t+00-βζsinυx1t-τ1t+-12αm1-m2x1t+1-x1t-100.where the parameters are α=10,β=18.53,ω=0.1636,m1=-1.4325,m2=-0.7831,ζ=0.2,υ=0.5, respectively, and the time delay τ1(t)=0.02. By calculation [9], we have l¯1=11.6435,l¯2=0.3088.
The outer coupling matrix which depicts the structure of network (1) is a 10×10 matrix, which is chosen as follows:(34)A=-301.500.500.5000.50-200.30.5000.700.501-40010101100-3000.50.5010.200.50-30.500.81001001-40002011020-5001100.51.5000.5-4.5100.50101100.5-40010.5000100-2.5.
The coupling weight is (35)c11t=16+16sinπt-1360cosπ2t,c12t=2,c13t=0,c14t=2.8,c15t=1-1.3cosπ2t,c16t=-0.5cosπ2t,c17t=0,c18t=sinπt,c19t=0,c110t=sinπt;c21t=1,c22t=12,c23t=cosπt,c24t=0,c25t=0,c26t=0.5cosπ3t,c27t=cosπt,c28t=0,c29t=cosπ3t,c210t=1;c31t=1-sinπ3t,c32t=1+0.5cosπ3t,c33t=12+12cosπ3t+14sinπ3t,c34t=0,c35t=0.7,c36t=1+1.5cosπ3t,c37t=cosπ3t,c38t=sinπ3t,,c39t=-1,c310t=0;c41t=1+cosπ3t,c42t=sinπt,c43t=2sinπt,c44t=16+23cosπ3t+13sinπt,c45t=0,c46t=cosπ3t,c47t=sinπt,c48t=-1+sinπt,c49t=-cosπ3t,c410t=cosπ3t;c51t=0,c52t=0,c53t=1,c54t=0,c55t=13-1112cos2π3t,c56t=1+0.5cos2π3t,c57t=0,c58t=0,c59t=-3cos2π3t,c510t=2.3+cos2π3t;c61t=-2.5-1.4sinπ3t-1.3cos2π3t,c62t=0.1+0.1sinπ3t,c63t=0,c64t=0,c65t=1-cos2π3t,c66t=1140+940sinπ3t-14cos2π3t,c67t=0,c68t=1+0.5sinπ3t,c69t=0,c610t=0.5sinπ3t;c71t=1,c72t=-5-cosπt,c73t=1+cosπt,c74t=0,c75t=0,c76t=0.5cosπ3t,c77t=-45-110cosπ3t,c78t=2cosπt,c79t=0,c710t=-0.5cosπ3t;c81t=1-sinπt,c82t=1+1.5cosπt,c83t=0,c84t=0,c85t=0.7,c86t=1-0.5cosπt,c87t=0,c88t=29-29sinπt,c89t=0,c810t=-cosπt;c91t=1+cosπ3t,c92t=sinπt,c93t=-2sinπt,c94t=-1,c95t=0,c96t=3-0.5cosπ3t,c97t=sinπt,c98t=0,c99t=78-12sinπt,c910t=-cosπ3t;c101t=0,c102t=cosπ2t,c103t=1-0.5cosπ2t,c104t=-cosπ2t,c105t=-1-2cos2π3t,c106t=1+0.5cosπ2t,c107t=-sinπt,c108t=sinπt,c109t=-sinπt,c1010t=-15-25sinπt+310cosπ2t.
By calculation, we get the common period T=6 of cij(t) and the parameters Ω∗=ϱij∗10×10 are designed as(36)Ω∗=ϱij∗=1.50.30.721.10.11.70.20.10.40.10.20.40.50.60.30.60.20.92.10.90.80.50.71.10.50.50.40.21.10.51.50.92.11.80.60.80.60.81.61.21.41.822.20.20.21.00.80.10.51.31.71.20.10.40.10.30.30.90.10.80.40.50.91.10.70.60.50.40.90.21.50.71.40.60.60.90.50.70.81.10.10.10.80.30.50.71.11.30.20.40.81.01.20.80.11.20.11.7,and ϱij(t)=t/6ϱij∗ for i,j=1,2,…,10. For simplicity, the inner coupling matrices are designed as follows: Γ=diag{0.5,1,0.8}. And the initial values of target orbit are s(0)=(-0.2,0.2,0.5).
The dynamical network with ten nodes can be expressed as follows:(37)x˙it=ft,xit,xit-τ0t+∑j=110cijtaijΓxjt+ui,i=1,2,…,10.
Choosing initial values of complex dynamical network (37) randomly in [-5,5], for the time span [0,15], the work time is [0,2)∪[4,7)∪[10,12.5), and the rest time is [2,4)∪[7,10)∪[12.5,15). With aperiodically intermittent controller (13), (14), and adaptive update law (15), the synchronization simulations of the complex dynamical network (37) are shown in Figures 2–9.
The first component of error evolution for each node.
The second component of error evolution for each node.
The third component of error evolution for each node.
The first component of the controller for each node.
The second component of the controller for each node.
The third component of the controller for each node.
The error evolution of the first node.
The error evolution of the first node.
Figures 2–4 display the dynamical evolution curves of the synchronization errors ei(t),i=1,2,…,10; as time goes on, the component of errors for each node tends to zero quickly. The aperiodically intermittent controllers are presented in Figures 5–7, from which we can see that all components of the controller are changed violently until they reach zero, which is consistent with the errors evolution results. Moreover, Figure 8 shows the feedback gain convergence to the fixed values: k1=53.54,k2=5.45,k3=29.01,k4=63.51,k5=8.48,k6=11.97,k7=11.50,k8=2.08,k9=55.79,k10=43.52. The estimation of unknown periodical coupling strengths is presented in Figure 9. By simulation, we can see that the controlled complex dynamical network (37) achieved exponential synchronization with the help of controller designed in (13) and (14), and the unknown periodical coupling strengths are all estimated successfully.
5. Conclusions
In this paper, we investigated aperiodically intermittent synchronization problem of complex dynamical network, which contains unknown periodically coupling strengths and bounded time varying delay and is correspondence with the practical complex network system in a great extent. Based on theories of intermittent control, adaptive control, and learning control, some useful aperiodically intermittent synchronization criteria for complex dynamical network with unknown periodical couplings have been obtained.
Also an illustrative example by the numerical simulation is provided to demonstrate the effectiveness and feasibility of the proposed synchronization method. From the simulation results, we can see that the complex dynamical network is exponentially synchronized when the aperiodically intermittent control is injected, and the coupling strengths is estimated successfully. Meanwhile, in the future, the author will take switching and noise disturbance into account, to study the systems’ finite-time aperiodically intermittent synchronization.
Data Availability
The Matlab based models used to support the findings of this study are available from the corresponding author upon request.
Conflicts of Interest
The authors declare that they have no conflicts of interest.
Acknowledgments
The work is supported by the National Natural Science Foundation of China under Grant nos. 61573013 and 11705122, the Natural Science Funds of Science and Technology Department of Shaanxi Province under Grant no. 2014JM2-1002, the nature scientific research project of Shaanxi Education Department under Grant no. 18JK0829, and Natural Science Foundation of Xianyang Normal University under Grant no. 14XSYK004.
WangJ.-L.WuH.-N.HuangT.2019Springer10.1007/978-981-13-1352-3MR3837511HeP.Pinning control and adaptive control for synchronization of linearly coupled reaction-diffusion neural networks with mixed delays20183281103112310.1002/acs.2890MR3844208Zbl1398.93157HeG.FangJ.-a.LiZ.Finite-time synchronization of cyclic switched complex networks under feedback control201735493780379610.1016/j.jfranklin.2016.10.016MR3649252Zbl1367.93245GhaffariA.ArebiS.Pinning control for synchronization of nonlinear complex dynamical network with suboptimal SDRE controllers2016831-21003101310.1007/s11071-015-2383-8MR3435922Zbl1349.34200GuoX.LiJ.A new synchronization algorithm for delayed complex dynamical networks via adaptive control approach201217114395440310.1016/j.cnsns.2012.03.022MR29303422-s2.0-84861682763Zbl1248.93061SelvarajP.KwonO.SakthivelR.Disturbance and uncertainty rejection performance for fractional-order complex dynamical networks2019112738410.1016/j.neunet.2019.01.009SelvarajP.SakthivelR.AhnC. K.Observer-based synchronization of complex dynamical networks under actuator saturation and probabilistic faults201899216822162-s2.0-85042873122DaiH.JiaJ.YanL.WangF.ChenW.Event-triggered exponential synchronization of complex dynamical networks with cooperatively directed spanning tree topology201933035536810.1016/j.neucom.2018.11.013MeiJ.JiangM.WuZ.WangX.Periodically intermittent controlling for finite-time synchronization of complex dynamical networks201579129530510.1007/s11071-014-1664-ySunJ.WuY.CuiG.WangY.Finite-time real combination synchronization of three complex-variable chaotic systems with unknown parameters via sliding mode control2017883167716902-s2.0-8501288821810.1007/s11071-017-3338-zZbl1380.34097JingT.ChenF.ZhangX.Finite-time lag synchronization of time-varying delayed complex networks via periodically intermittent control and sliding mode control20161991781842-s2.0-8497946798410.1016/j.neucom.2016.03.018ZochowskiM.Intermittent dynamical control20001453-41811902-s2.0-034668699510.1016/S0167-2789(00)00112-3MeiJ.JiangM.WangX.HanJ.WangS.Finite-time synchronization of drive-response systems via periodically intermittent adaptive control201435152691271010.1016/j.jfranklin.2014.01.008MR31919152-s2.0-84898801293ZhengM.LiL.PengH.XiaoJ.YangY.ZhaoH.RenJ.Finite-time synchronization of complex dynamical networks with multi-links via intermittent controls20168911210.1140/epjb/e2016-60935-7FanY.LiuH.ZhuY.MeiJ.Fast synchronization of complex dynamical networks with time-varying delay via periodically intermittent control20162051821942-s2.0-8496716321210.1016/j.neucom.2016.03.049LiL.TuZ.MeiJ.JianJ.Finite-time synchronization of complex delayed networks via intermittent control with multiple switched periods201685137538810.1007/s11071-016-2692-6MR3510620Zbl06653673ZhaoH.CaiG.Exponential synchronization of complex delayed dynamical networks with uncertain parameters via intermittent control201593779198QiuJ.ChengL.ChenX.LuJ.HeH.Semi-periodically intermittent control for synchronization of switched complex networks: a mode-dependent average dwell time approach20168331757177110.1007/s11071-015-2445-yMR3449506LiuM.YuZ.JiangH.HuC.Synchronization of complex networks with coupled and self-feedback delays via aperiodically intermittent strategy20171962062207510.1002/asjc.1577MR37301952-s2.0-85028523880LiuM.JiangH.HuC.Finite-time synchronization of delayed dynamical networks via aperiodically intermittent control2017354135374539710.1016/j.jfranklin.2017.05.030MR3679202Zbl1395.93348WangJ.Synchronization of delayed complex dynamical network with hybrid-coupling via aperiodically intermittent pinning control201735441833185510.1016/j.jfranklin.2016.11.034MR3607407ChengL.ChenX.QiuJ.LuJ.CaoJ.Aperiodically intermittent control for synchronization of switched complex networks with unstable modes via matrix ω -measure approach2018923109111022-s2.0-8504219490210.1007/s11071-018-4110-8Zbl1398.34039JiangW.LiL.TuZ.FengY.Semiglobal finite‐time synchronization of complex networks with stochastic disturbance via intermittent control20192982351236310.1002/rnc.4496ZhangL.YangX.XuC.FengJ.Exponential synchronization of complex-valued complex networks with time-varying delays and stochastic perturbations via time-delayed impulsive control2017306223010.1016/j.amc.2017.02.004BotmartT.NiamsupP.Exponential synchronization of complex dynamical network with mixed time-varying and hybrid coupling delays via intermittent control201435441833185510.1186/1687-1847-2014-116MR3357318AhmedM. A. A.LiuY.ZhangW.AlsaediA.HayatT.Exponential synchronization for a class of complex networks of networks with directed topology and time delay20172662742832-s2.0-8502001486910.1016/j.neucom.2017.05.039ZhangL.YangX.XuC.FengJ.Exponential stability of stochastic differential equations with impulse effects at random times: stochastic differential equations with random impulses201810.1002/asjc.1937HaoX.LiJ.Stochastic synchronization for complex dynamical networks with time-varying couplings2015803135713632-s2.0-8493995659010.1007/s11071-015-1947-yZbl1351.34058SelvarajP.SakthivelR.KwonO. M.Finite-time synchronization of stochastic coupled neural networks subject to Markovian switching and input saturation20181051541652-s2.0-8504818050410.1016/j.neunet.2018.05.004SunJ.WangY.WangY.ShenY.Finite-time synchronization between two complex-variable chaotic systems with unknown parameters via nonsingular terminal sliding mode control20168521105111710.1007/s11071-016-2747-8MR3511427Zbl1355.340972-s2.0-84961639091YanL.LiJ.Adaptive finite-time synchronization for complex dynamical network with different dimensions of nodes and time-varying outer coupling structures2018201812247415010.1155/2018/2474150MR3894386LiuX.ChenT.Synchronization of linearly coupled networks with delays via aperiodically intermittent pinning control201526102396240710.1109/TNNLS.2014.2383174MR34531842-s2.0-85027941991LiuX.ChenT.Synchronization of complex networks via aperiodically intermittent pinning control201560123316332110.1109/TAC.2015.2416912MR3432701Zbl1360.93359