A novel fixed-time controller is proposed to perform fixed-time complex modified function projective lag synchronization of chaotic (hyperchaotic) complex systems. The synchronization is obtained after a finite time that could be preestablished without the knowledge of the initial states of both synchronized chaotic (hyperchaotic) complex systems. The global fixed-time stability of the closed-loop systems is rigorously proven based on Lyapunov analysis. Some simulation examples are provided to demonstrate the effectiveness and feasibility of the proposed method and verify the theoretical results.
National Research Foundation of KoreaNRF-2016R1D1A3B039304961. Introduction
In the past two decades, much work has been devoted to investigate dynamic behavior and synchronization of chaotic (hyperchaotic) complex systems because of their potential applications. For instance, Gibbon and McGuinness [1] presented a complex set of Lorenz equations derived in laser optics and baroclinic instability, Zhang and Liu [2] applied synchronization to communication problem under considering time delay, and Wu et al. [3] proposed a method to improve the secure communications via passive synchronization of hyperchaotic complex systems. In the literature, researchers have introduced several types of synchronization of chaotic (hyperchaotic) complex systems such as complete synchronization (CoS) [4], antisynchronization (AS) [5], compound synchronization [6], projective synchronization (PS) [7], modified projective and modified function projective synchronization (MPS-MFPS) [8], combination-combination synchronization [9], complete lag synchronization (CoLS) [10], general hybrid projective complete dislocated synchronization [11], modified projective lag synchronization (MPLS) [12], and modified function projective lag synchronization (MFPLS) [13]. Recently, some new types of synchronization that utilized complex scaling factors have been introduced to achieve the synchronization of both module and phase, the concept first proposed by Nian et al. [14]. Complex complete synchronization (CCoS) [15, 16], complex projective synchronization (CPS) [17], combination complex synchronization [18], complex modified projective synchronization (CMPS) [19], and complex function projective synchronization (CFPS) [20, 21] are some examples. In these synchronizations, the scaling factors are complex numbers (or functions) that increase the unpredictability and complexity; they have thus significant meaning for enhancing the security of communication [20]. Observation from [8, 20, 21] shows that mentioned kinds of synchronizations only are special cases of complex modified function projective synchronization (CMFPS). Furthermore, there might be time delay between drive system and response system from practical point of view [13, 20, 21]; complex modified function projective lag synchronization (CMFPLS), which is a more general case, is rarely studied or not mentioned up to date.
In order to obtain such complex synchronization of chaotic (hyperchaotic) complex systems, some control methods such as Lyapunov-based control, feedback control and/or adaptive control [2, 5, 6, 8, 11–15, 17–20], passive control [3], finite-time stability theory-based control [4, 9, 16], nonlinear observer-based control [7], backstepping-based control [10], and adaptive fuzzy logic control [21] have been proposed. In the mentioned methods, while the ones in [4, 9, 16] ensure obtaining finite-time synchronization in which the convergence time still depends on both control design parameters and initial conditions, these remainders guarantee that the synchronization errors might converge to origin asymptotically in the best case. It is certain that synchronization errors could reach to zero in an amount of time which is more desirable object. Finite-time stability theory-based control methods have been known as useful and efficient techniques with some advanced properties such as fast and finite-time convergence, robustness, and high tracking precision [22]. In some last years, some researchers have introduced fixed-time stability concept, which is a further extension of finite-time stability one, in the literature [23–26]. The fixed-time stabilization problem of linear systems has solved in [23] based on nonlinear feedback control. Inspired by well-known super-twisting (STW) algorithm and high-order differentiator [27, 28], Basin et al. have proposed an adaptive STW-like controller with uniform finite/fixed convergence time in [24] and a continuous finite-time and fixed-time control method for an n-dimensional chain of integrators [25]. In [26], Polyakov et al. have introduced finite-time and fixed-time control methods for stabilizing linear multi-input-multioutput systems. The fixed-time control methods ensure that the origin of the system is stabilized in some finite time that is not dependent on the initial conditions; in other words, the convergent time is completely predefined via designed controller parameters.
Motivated by the above discussion, the main goal of this article is to propose a novel fixed-time control law to, for the first time, obtain fixed-time complex modified function projective lag synchronization of chaotic (hyperchaotic) complex systems; that is, the modified function projective lag synchronization (CMFPLS) error states converge to zero in a predefined amount of time despite initial conditions. The global stability and fixed-time convergence are rigorously proven, and an upper bound of the convergence time is estimated.
The rest of this paper is organized into four sections. Section 2 gives some preliminaries and problem formulation. In Section 3, a fixed-time controller is introduced to solve the CMFPLS problem. Two illustrative examples are provided to demonstrate the effectiveness and the feasibility of the suggested control method in Section 4. Finally, some conclusions are drawn in Section 5.
Notations. R and C denote, respectively, the space of real number and complex number, and diag(·) represents a block-diagonal matrix. If x∈R, the following notation is introduced for simplicity of expression [29]:(1)xϕ=xϕsignx,where ϕ>0, and it can be verified that as ϕ≥1,(2)ddtxϕ=ϕxϕ-1x˙.If x∈C, that is, x=xr+jxi (superscripts r and i stand for the real and imaginary parts of the complex number x, and j=-1, we then adopt the notation that x¯=xr-jxi,xϕ=xrϕ+jxiϕ,x=x1,x2,…,xnT∈Rn,x=x1,x2,…,xnT∈Cn.
2. Preliminaries and Problem Formulation
First at all, some concepts related to finite-time stability and fixed-time stability given in [23] and a lemma that will be used in the analysis later are introduced as follows.
Consider the system(3)x˙=ψt,x,x0=x0,where x∈Rn and ψ:R+×Rn→Rn is a nonlinear function which can be discontinuous. Assume the origin is an equilibrium point of (3).
Definition 1.
The origin of (3) is said to be globally finite-time stable if it is globally asymptotically stable and any solution x(t,x0) of (3) reaches the equilibria at some finite-time moment; that is, xt,x0=0,∀t≥Tx0, where T:Rn→R+∪{0} is the settling-time function.
Definition 2.
The origin of (3) is said to be globally fixed-time stable if it is globally finite-time stable and the settling-time function T(x0) is bounded; that is, ∃Tmax>0:T(x0)≤Tmax,∀x0∈Rn.
Lemma 3 (see [<xref ref-type="bibr" rid="B30">30</xref>], Jensen’s inequality).
One has(4)∑i=1mziθ21/θ2≤∑i=1mziθ11/θ1,0<θ1<θ2with zi≥0 and 0≤i≤m.
Two different n-dimensional complex chaotic (hyperchaotic) systems under this study are described as follows.
The drive system is (5)x˙t=x˙r+jx˙i=fxt,x¯t,t,where xt and fxt,x¯t,t are the state complex vector and the vector of nonlinear complex functions of the drive system, respectively.
And response system is (6)y˙t=y˙r+jy˙i=gyt,y¯t,t+ut,where yt,gyt,y¯t,t, and ut are the state complex vector, the vector of nonlinear complex functions, and the control inputs of the controlled response system, respectively.
Define complex modified function projective lag synchronization (CMFPLS) error between the drive system and the controlled response system as (7)et=er+jei=yt-Ωtxt-τ,where Ωt=Ωr(t)+jΩi(t)=diag(ω1t,ω2t,…,ωnt) is a function scaling matrix, ωkt=ωkr(t)+jωki(t)∈C are different and continuously differentiable bounded functions, ωkt≠0 for all t≥0,τ is positive time lag, and et=e1t,e2t,…,entT; here ekt=ekrt+jekit,k=1,2,…,n. For convenience of notation, we utilize ek,ekr, and eki instead of ek,ekrt, and ekit, respectively.
Then, the CMFPLS error dynamics is determined by the following equation:(8)e˙k=e˙kr+je˙ki=-ωktfkxt-τ,x¯t-τ,t-τ-ω˙ktxkt-τ+gkyt,y¯t,t+ukt.
Definition 4.
The fixed-time complex modified function projective lag synchronization between the drive system (3) and the response system (4) is achieved if there exists a time moment T>0, which is independent of initial error states, e(t0), such that e(t)=0,∀t≥t0+T.
The main goal of this article is to design control laws ukt,k=1,2,…,n such that the resulting closed-loop system is global fixed-time stability; that is, the fixed-time CMFPLS of the error system (5) is obtained in the sense of Definition 4.
Remark 5.
The differential equation of form (3) is the expression of most of the well-known chaotic (hyperchaotic) complex systems such as chaotic (hyperchaotic) complex Chen, Lorenz, and Lü systems.
Remark 6.
Besides several kinds of synchronization mentioned in the Introduction, CMFPLS considered in this study also consists of some other special cases such as complex function projective lag synchronization (CFPLS), complex modified projective lag synchronization (CMPLS), complex projective lag synchronization (CPLS), complex complete lag synchronization (CCoLS), complex lag antisynchronization (CLAS), complex antisynchronization (CAS), function projective lag synchronization (FPLS), function projective synchronization (PLS), projective lag synchronization (PLS), and lag antisynchronization (LAS) summarized in Table 1.
Some types of synchronization.
Function scaling matrixΩt=Ωr(t)+jΩi(t)=diag(ω1t,ω2t,…,ωnt)
Type of synchronization
ωkt=ωkr(t)+jωki(t),(k=1,2,…,n)
τ≠0
τ=0
ωkt,k=1,2,…,n are identical complex functions
CFPLS
CFPS
ωkt,k=1,2,…,n are different complex constants
CMPLS
CMPS
ωkt,k=1,2,…,n are identical complex constants
CPLS
CPS
ωkt=1+j,k=1,2,…,n
CCoLS
CCoS
ωkt=-1-j,k=1,2,…,n
CLAS
CAS
ωkrt,k=1,2,…,n are different real functions and Ωit=0
MFPLS
MFPS
ωkrt,k=1,2,…,n are identical real functions and Ωit=0
FPLS
FPS
ωkrt,k=1,2,…,n are different real constants and Ωit=0
MPLS
MPS
ωkrt,k=1,2,…,n are identical real constants and Ωit=0
PLS
PS
ωkt=1,k=1,2,…,n
CoLS
CoS
ωkt=-1,k=1,2,…,n
LAS
AS
3. Main Results
In this section, a novel control method is proposed to achieve the fixed-time CMFPLS between the drive system (3) and the controlled response system (4). The main result of the article is summarized in the following theorem.
Theorem 7.
Consider the CMFPLS error dynamics (8). If a feedback control law is designed as(9)ukt=ukr+juki=ωktfkxt-τ,x¯t-τ,t-τ+ω˙ktxkt-τ-gkyt,y¯t,t-h1kek2-βk-h2keλkekrekr+jeλkekieki-h3kekβk,where h1k,h2k, and h3k are positive constants and satisfy 4h1kh3k>h2k2,0<βk<1,λk>0,(k=1,2,…,n), and e is Euler’s number, then the origin of the resulting closed-loop system is globally fixed-time stable; that is, the fixed-time CMFPLS between the drive system (3) and the controlled response system (4) is obtained in the sense of Definition 4. In addition, the CMFPLS error states of (8) reach to zero at most after the fixed time (10)T=21-βk4h1kh3k-h2k2π2-tan-1h2k4h1kh3k-h2k2.
Proof.
The CMFPLS error dynamics (8) under the control law (9) results in the closed-loop system represented in the form(11)e˙k=e˙kr+je˙ki=-h1kek2-βk-h2keλkekrekr+jeλkekieki-h3kekβk.The resulting closed-loop system can be rewritten as(12)e˙kr=-h1kekr2-βk-h2keλkekrekr-h3kekrβk,e˙ki=-h1keki2-βk-h2keλkekieki-h3kekiβk.Consider the following Lyapunov function candidate(13)V=∑k=1nekr2+eki2.Its time derivatives along the trajectories of (12) lead to(14)V˙=2∑k=1nekr-h1kekr2-βk-h2keλkekrekr-h3kekrβk+2∑k=1neki-h1keki2-βk-h2keλkekieki-h3kekiβk=2∑k=1n-h1kekr3-βk-h2keλkekrekr2-h3kekr1+βk+2∑k=1n-h1keki3-βk-h2keλkekieki2-h3keki1+βk.Applying Lemma 3, we can have (15)V˙≤2∑k=1n-h1kekr2+eki23-βk/2-h2kekr2+eki2-h3kekr2+eki21+βk/2≤-2h1kV3-βk/2-2h2kV-2h3kV1+βk/2.Hence, the Lyapunov theorem [31] indicates that origin of (12) is globally asymptotically stable. Further, we will show that the convergence takes place in finite time.
We consider the differential equation (16)φ˙=-2h1kφ3-βk/2-2h2kφ-2h3kφ1+βk/2,φ0=φ0≥0.It follows that(17)dt=-dφ2h1kφ1-β+h2kφ1-βk/2+h3kφβk+1/2=-11-βkdφ1-βk/2h1kφ1-βk+h2kφ1-βk/2+h3k.The solution of (17) found under the condition, h2k2≤4h1kh3k, satisfies the following equation:(18)t=21-βk4h1kh3k-h2k2tan-12h1kφ0+h2k4h1kh3k-h2k2-tan-12h1kφt+h2k4h1kh3k-h2k2.Since expression (16) implies that φt is decreasing, (18) indicates that φ(t) will become zero in finite time.(19)Tφ=21-βk4h1kh3k-h2k2tan-12h1kφ0+h2k4h1kh3k-h2k2-tan-1h2k4h1kh3k-h2k2.In addition, on the basis of the comparison principle [32], we can obtain that Vt≤φt when Vt=0≤φ0, and it follows that Vt, and therefore the origin of (12), converge to zero in finite time TV≤Tφ, with(20)TV=21-βk4h1kh3k-h2k2tan-12h1kVt=0+h2k4h1kh3k-h2k2-tan-1h2k4h1kh3k-h2k2.Thus, the origin of (12) is globally finite-time stable on the basic of Definition 1.
It can be easy to show that Tφ is bounded by(21)TV≤T=21-βk4h1kh3k-h2k2π2-tan-1h2k4h1kh3k-h2k2.Since the above expression of T is only dependent on the controller parameters, it follows from Definition 2 that the origin of (12) is globally fixed-time stable. Consequently, the fixed-time CMFPLS between the drive system (3) and the controlled response system (4) is achieved in the sense of Definition 4. Theorem 7 is proven.
Remark 8.
Regarding differential equation (11), -h1kek2-βk-h2keλkekrekr+jeλkekieki holds a dominant function that ensures fast convergence speed while ek is much larger than 1, and h3kekβk takes the dominant one that determines finite-time convergence while ek is much smaller than 1. Moreover, the “soft gain” h2κeλkekr (or h2κeλkeki) is small if ekr (or eki) is close to zero. It implies that the influence of h3kekβk is strong in the finite-time convergence period.
4. Numerical Simulations
Two illustrative examples are provided to demonstrate the effectiveness and applicability of the proposed control method in this section. In the numerical simulations, the fourth-order Runge-Kutta method is used to solve differential equations with a time step size of 0.001.
4.1. Example 1
This example demonstrates the efficacy of the proposed approach by comparing to the existing method [19] through the CMPS between two chaotic systems. It should be noted that chaotic complex systems and the controller parameters of the existing method were directly taken from [19] to make a fair comparison.
The chaotic complex Lorenz system is assumed as the drive system given by(22)x˙1=14x2-x1,x˙2=35x1-x2-x1x3,x˙3=12x¯1x2+x1x¯2-3.7x3and the chaotic complex Lu system is used as the controlled response system as(23)y˙1=40y2-y1,y˙2=22y2-y1y3,y˙3=12y¯1y2+y1y¯2-5y3.The initial states of systems (22) and (23) are chosen as x0=-1+j2,-3+j4,5T and y0=3+j4,-3+j,3T, respectively. Based on Remark 6 and Table 1, the CMPS is obtained by choosing the time delay as τ=0 and the factor scaling functions as Ω=diag1,-1,0.
The proposed control law is designed as (9) where the parameter values are set as h1k=2,h2k=1,h3k=2,λk=0.1,βk=0.8,k=1,2,3, and then the upper bound of the convergence time is T1≈3.4 s based on (10), which is not dependent on the initial conditions of the chaotic systems.
The numerical simulation results are depicted in Figures 1 and 2. It is observed that all the CMPS errors reach origin in a finite amount of time that is less than T1=3.4 s, which implies that the fixed-time CMPS between the drive system (22) and controlled response system (23) is achieved as in Figure 2. In contrast, we can see that the CMPS errors under the method in [19] is to converge to zero asymptotically; that is, limt→+∞ei=0. Furthermore, we could conclude from Figure 1 that the proposed method exhibits much better performance than the method in [19] does.
The time histories of the CMPS error states.
The time histories of the state variables of the drive system (22) and the controlled response system (23).
4.2. Example 2
The CMFPLS between two hyperchaotic complex systems is investigated in this example to completely verify the theoretical result introduced. The hyperchaotic complex Lu system [12] and hyperchaotic complex Lorenz system [12] are adopted as the drive system and the controlled response system, respectively.
The drive system is described as follows:(24)x1=42x2-x1+x4,x2=25x2-x1x3+x4,x3=12x¯1x2+x1x¯2-6x3,x4=12x¯1x2+x1x¯2-5x4and the controlled response system is expressed as(25)y1=14y2-y1+jy4,y2=40y1-y2-y1y3+jy4,y3=12y¯1y2+y1y¯2-5y3,y4=12y¯1y2+y1y¯2-13y4.To show that the upper bound of the convergence time, T2, is independent of the initial conditions of synchronized systems, we will consider five cases in this example. The initial condition of the drive system (24) is set as x0=-1+j2,-2+j3,4,-3T, and the initial condition of the controlled response system (25) is selected in the form of y0=σ∗2+j6,7-j5,5,-1T, and here σ=-2,-1,0.5,1,3. The time delay is chosen as τ=0.2, and the scaling factor functions are chosen as ω1t=1+jsint,ω2t=2-jcost,ω3t=1.5,ω4t=cos2t.
For this example, the controller parameters are h1k=2,h2k=1,h3k=2,λk=0.1, and βk=0.65,k=1,2,3,4; then this yields T2=1.94 s from (10). It is assumed that the controller is activated at t0=3 s, and the upper bound of the whole convergence time, thus, is t0+T2=5.94 s.
The time responses of the CMFPLS error states are displayed in Figure 3. All the synchronization errors reach zero within the period t0+T2=5.94 s despite the different initial conditions; evidently, the CMFPLS could be obtained in fixed time.
The time histories of the CMFPLS error states.
5. Conclusions
In this work, a novel fixed-time control method has been proposed to achieve fixed-time complex modified function projective lag synchronization of chaotic (hyperchaotic) complex systems. The numerical simulations have been given to verify the theoretical results and demonstrate the effectiveness and feasibility of the suggested scheme. It should be noted that the proposed method could be extended to either chaos control problems or synchronization types of chaotic (hyperchaotic) systems with different orders and non-square scaling factor matrices.
Conflicts of Interest
The authors declare that there are no conflicts of interest regarding the publication of this paper.
Acknowledgments
This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (NRF-2016R1D1A3B03930496).
GibbonJ. D.McGuinnessM. J.The real and complex Lorenz equations in rotating fluids and lasersZhangF.LiuS.Self-time-delay synchronization of time-delay coupled complex chaotic system and its applications to communicationWuX.ZhuC.KanH.An improved secure communication scheme based passive synchronization of hyperchaotic complex nonlinear systemZhouX. B.JiangM. R.CaiX. M.Synchronization of a novel hyperchaotic complex-variable system based on finite-time stability theorySunJ.ShenY.Adaptive anti-synchronization of chaotic complex systems and chaotic real systems with unknown parametersSunJ.ShenY.YinQ.XuC.Compound synchronization of four memristor chaotic oscillator systems and secure communicationLiuP.SongH. J.LiX.Observe-based projective synchronization of chaotic complex modified van der pol-duffing oscillator with application to secure communicationSunJ.ShenY.ZhangX.Modified projective and modified function projective synchronization of a class of real nonlinear systems and a class of complex nonlinear systemsSunJ.ShenY.WangX.ChenJ.Finite-time combination-combination synchronization of four different chaotic systems with unknown parameters via sliding mode controlWangX.ZhangH.Backstepping-based lag synchronization of a complex permanent magnet synchronous motor systemSunJ.WangY.YaoL.ShenY.CuiG.General hybrid projective complete dislocated synchronization between a class of chaotic real nonlinear systems and a class of chaotic complex nonlinear systemsMahmoudG. M.MahmoudE. E.Modified projective lag synchronization of two nonidentical hyperchaotic complex nonlinear systemsWangX.WeiN.Modified function projective lag synchronization of hyperchaotic complex systems with parameter perturbations and external perturbationsNianF.WangX.NiuY.LinD.Module-phase synchronization in complex dynamic systemMahmoudE. E.Complex complete synchronization of two nonidentical hyperchaotic complex nonlinear systemsSunJ.WangY.WangY.ShenY.Finite-time synchronization between two complex-variable chaotic systems with unknown parameters via nonsingular terminal sliding mode controlWuZ.DuanJ.FuX.Complex projective synchronization in coupled chaotic complex dynamical systemsSunJ.CuiG.WangY.ShenY.Combination complex synchronization of three chaotic complex systemsMahmoudG. M.MahmoudE. E.Complex modified projective synchronization of two chaotic complex nonlinear systemsZhangF.LiuS.Adaptive complex function projective synchronization of uncertain complex chaotic systemsLeeJ. W.LeeS. M.WonS. C.Complex function projective synchronization of general networked chaotic systems by using complex adaptive fuzzy logicBhatS. P.BernsteinD. S.Finite-time stability of continuous autonomous systemsPolyakovA.Nonlinear feedback design for fixed-time stabilization of linear control systemsBasinM.PanathulaC. B.ShtesselY.Adaptive uniform finite-/fixed-time convergent second-order sliding-mode controlBasinM.ShtesselY.AldukaliF.Continuous finite- and fixed-time high-order regulatorsPolyakovA.EfimovD.PerruquettiW.Robust stabilization of MIMO systems in finite/fixed timeLevantA.Higher-order sliding modes, differentiation and output-feedback controlAnguloM. T.MorenoJ. A.FridmanL.Robust exact uniformly convergent arbitrary order differentiatorPolyakovA.FridmanL.Stability notions and Lyapunov functions for sliding mode control systemsBechkenbackE.Bell ManR.SlotineJ. J. E.LiW.KhalilH. K.