In this paper, control of fractional-order financial chaotic systems with saturated control input is
investigated by means of state-feedback control method. The saturation problem is tackled by using
Gronwall-Bellman lemma and a memoryless nonlinearity function. Based on Gronwall inequality and
Laplace transform technique, two sufficient conditions are achieved for the asymptotical stability of the
fractional-order financial chaotic systems with fractional orders 0 < α ≤ 1 and 1 < α < 2, respectively.
Finally, simulation studies are carried out to show the effectiveness of the proposed linear control
method.
1. Introduction
In the past two decades, studies of chaotic systems have received more and more attention in various fields of natural sciences. This is because chaotic systems are rich in dynamics and possess great sensitivity to initial conditions. Up to now, econophysics has been raised to an alternative scientific methodology to comprehend the highly complex dynamics in economic and financial systems. Many economists are working hard to explain the central features of economic data, including erratic macroeconomic fluctuations (business cycles), irregular microeconomic fluctuations, irregular growth, structural changes, and overlapping waves of economic development [1, 2]. Representative effects, that is, treated as random shocks, are political events, weather variables, and other human factors [3–7]. Compared with the opinion discussed above, chaos supports an endogenous explanation of the complexity appeared in economic series.
Since chaos in financial systems was firstly studied in 1985, great impact has been put on the prominent economics recently, because the occurrence of the chaotic phenomenon in the economic system indicates that the macroeconomic operation has in itself the inherent indefiniteness. Studies on the complicated financial systems by using nonlinear method are fruitful [2, 8, 9]. Controlling chaos in fractional-order financial systems is also studied in recent years [10–18]. In [15], an active sliding mode controller is constructed to synchronize fractional-order financial chaotic systems in master-slave structure. In [16], a necessary condition is introduced to confirm the existence of 1-scroll, 2-scroll, or multiscroll chaotic attractors in a fractional-order financial system and a sliding mode controller is proposed. Active control method is also used in [17], and the variable-order fractional derivative is defined in Caputo type. Wang et al. investigate impulsive synchronization and adaptive-impulsive synchronization of a novel financial hyperchaotic system [18]. In above literatures, the stability analysis is carried out based on fractional-order linear system stability theorem and only the situation where fractional order 0<α≤1 is concerned.
Most of real world technical systems are subjected to input constraints, especially in financial systems. In financial systems, input saturation does exist due to a limited size of weather variables, political events, and other human factors. The existence of input saturation may decrease the control performance or cause oscillations and even lead to instability of the system [19–21]. It is advisable for us to consider the control of financial systems with input saturation. For the integer-order linear and nonlinear systems, input saturation has received much attention from researchers in the past decade. The sector bounded condition associated with input nonlinearities is useful for analysis and synthesis of control systems subject to input saturation. Then the stability of the system can be formulated using Lyapunov stability theory and invariant theory.
Though many research efforts have been put to the fractional-order financial chaotic systems, the financial systems with saturated control input have rarely been investigated in literatures. Here, with the help of Laplace transform, Mittag-Leffler function, and Gronwall inequality, a linear controller will be derived for fractional-order financial chaotic systems in this paper. There are some main contributions that are worth to be emphasized as follows.
Two sufficient conditions are derived for the asymptotical stability of fractional-order financial chaotic systems with fractional orders 0<α≤1 and 1<α≤2, respectively.
A linear controller is given to control the fractional-order financial chaotic system.
A memoryless nonlinearity function is employed to handle the input saturation problem in fractional-order chaotic systems.
2. Preliminaries and System Description2.1. Preliminaries
The Caputo definition of fractional-order derivatives can be expressed as [21–24]
(1)0CDtαx(t)=1Γ(n-α)∫0tx(n)(τ)(t-τ)α-n+1dτ,n-1<α<n,
where α represents the fractional order and the Euler function Γ(·) is defined as Γ(τ)=∫0∞tτ-1e-tdτ.
The Laplace transform of Caputo fractional derivative can be given as
(2)∫0∞e-stDtα0Cx(t)dt=sα-∑k=0n-1sα-k-1x(k)(0).
The following definition and lemmas will be used.
Definition 1.
The Mittag-Leffler function with two parameters can be written as
(3)Eα,β(z)=∑k=0∞zkΓ(αk+β),
where α,β>0 and z∈C, and its Laplace transform can be given as
(4)L{tβ-1Eα,β(-λtα)}=sα-βsα+λ.
Lemma 2 (see [22]).
If A∈Rn×n,0<α≤1, β is an arbitrary real number, and C>0 is a real constant, then
(5)Eα,β(A)≤C1+∥A∥,
where μ≤|arg(eig
(A))|≤π with μ∈R satisfying πα/2<μ<min{π,πα}.
Lemma 3 (see [24]).
If t∈[0,T] and
(6)x(t)≤h(t)+∫0tk(τ)x(τ)dτ,
where k(t)≥0 and all the functions involved are continuous on the interval [0,T], then we can obtain
(7)x(t)≤h(t)+∫0tk(τ)h(τ)exp[∫τtk(u)du]dτ.
Definition 4 (see [18]).
A memoryless nonlinearity φ(t,x):[0,∞)×Rp→Rp is said to satisfy a sector condition if the following inequality holds:
(8)(φ(t,x)-K1x)T(φ(t,x)-K2x)≤0,∀x∈S
for constant matrices K1 and K2, where K2-K1 is a symmetric positive matrix and S contains the origin.
Based on the Definition 4, the following lemma holds.
Lemma 5 (see [18]).
Let
(9)S(LK,u0)={x(t)∈Rn∣-u0≤LKx(t)≤u0},
where L=
diag
[l1,l2,…,ln] with 0<li≤1,∀i=1,2,…,n, and
(10)φ(t,x(t))=sat(Kx(t))-LKx(t);
then the following inequalities are equivalent:
(sat(Kx(t))-LKx(t))T(sat(Kx(t))-Kx(t))≤0;
φ(t,x(t))T(φ(t,x(t))-(K-LK)x(t))≤0;
∥φ(t,x(t))∥≤∥K-LK∥∥x(t)∥.
Lemma 6.
The autonomous dynamic system
(11)Dαx(t)=Ax(t),x(0)=x0
is asymptotically stable if the following condition holds:
(12)|arg(
eig
(A))|>πα2.
The stability region for 0<α<1 is depicted in Figure 1.
Stability region of linear system (11) with fractional order 0<α<1.
2.2. Description of Fractional-Order Financial Chaotic Systems
The fractional-order financial chaotic systems are proposed by [1]. The mathematical model describes a fractional-order financial system including three nonlinear differential equations. The states, x1(t), x2(t), and x3(t), represent the interest rate, the investment demand, and the price index, respectively. The fractional-order model of the system can be described as
(13)0CDtαx1(t)=x3(t)+(x2(t)-a)x1(t),0CDtαx2(t)=1-bx2(t)-x12(t),0CDtαx3(t)=-x1(t)-cx3(t),
where a denotes the saving amount, b is the cost per investment, and c is the elasticity of demand of commercial market. 0<α<2 is the fractional-order derivative.
3. State-Feedback Controller Design and Stability Analysis3.1. Fractional Order α: 0<α≤1
Let us rewrite the controlled system (13) as the following compact form:
(14)0CDtαx(t)=Px(t)+f(x(t))+[0,1,0]T+sat(u(t)),x(0)=x0,
where x(t),u(t)∈R3, represent the state variables and the control input, respectively. Consider that(15)P=[-a010-b0-10-c],f(x)=[x1(t)x2(t)-x12(t)0]sat(u)=[sat(u1),sat(u2),sat(u3)]T
is the vector-valued saturation function with
(16)sat(ui)=sign(ui)min(u0i,|ui|),i=1,2,3,
where u0i represents the symmetric saturation level of the ith control input.
Noting that in chaotic systems the states are bounded, the nonlinear function f(x) satisfies
(17)∥f(x(t))∥=|x1(t)|x12(t)+x22(t)≤M∥x(t)∥,
where M>|x1(t)| is a constant.
Let us define the state-feedback control input as
(18)u(t)=Kx(t)+[0,-1,0]T,
where K∈R3×3 is the control gain matrix. Then we have
(19)0CDtαx(t)=Px(t)+[0,1,0]T+f(x(t))+sat(Kx(t)+[0,-1,0]T)=Ax(t)+f(x(t))+φ(t,x(t)),
where A=P+LK.
Theorem 7.
Consider system (14). If we choose matrices L and K such that |arg(
eig
(A))|>απ/2 and α∥A∥>d, then system (14) is asymptotically stable. d is a positive constant and will be defined later.
Proof.
Taking Laplace transform on (19), we can obtain
(20)sαX(s)=sα-1x0+AX(s)+L{f(x)}+L{φ(t,x)},
where X(s) represents the Laplace transform of x(t). Let I denote the 3×3 identity matrix; we have
(21)X(s)=(Isα-A)-1(sα-1x0+L{f(x)}+L{φ(t,x)}).
By taking Laplace inverse transform on (21), we get the solution of system (14):
(22)x(t)=Eq,1(A(t)α)x0+∫0t(t-τ)α-1gggggggggggghhggggg×Eα,α(A(t-τ)α)f(x)dτ+∫0t(t-τ)α-1Eα,α(A(t-τ)α)φ(t,x)dτ.
According to Lemma 2, we know there exist some constants ci>0,i=1,2, such that
(23)∥x(t)∥≤c1∥x0∥1+∥A∥tα+∫0t(t-τ)α-1c21+∥A∥(t-τ)α∥f(x)∥dτ+∫0t(t-τ)α-1c21+∥A∥(t-τ)α∥φ(t,x)∥dτ.
Then (23) can be rewritten as
(24)∥x(t)∥≤c1∥x0∥1+∥A∥tα+M∫0t(t-τ)α-1c21+∥A∥(t-τ)α∥x(t)∥dτ+∫0t(t-τ)α-1c21+∥A∥(t-τ)α∥φ(t,x)∥dτ.
From Definition 4 and Lemma 5, we know that φ(t,x(t)) satisfies
(25)∥φ(t,x(t))∥≤∥K-LK∥∥x(t)∥.
Let d=Mc2+c2∥K-LK∥; we can obtain
(26)∥x(t)∥≤c1∥x0∥1+∥A∥tα+d∫0t(t-τ)α-11+∥A∥(t-τ)α∥x(t)∥dτ.
By using Lemma 3, we have
(27)∥x(t)∥≤c1∥x0∥1+∥A∥tα+∫0tdc1(t-τ)α-1∥x0∥(1+∥A∥(t-τ)α)(1+∥A∥τα)×exp(∫τtd(t-y)α-11+∥A∥(t-y)αdy)dτ=c1∥x0∥1+∥A∥tα+∫0tdc1(t-τ)α-1∥x0∥(1+∥A∥τα)(1+∥A∥(t-τ)α)1-c/α∥A∥dτ≤c1∥x0∥1+∥A∥tα+dc1∥x0∥∥A∥d/α∥A∥-2∫0t(t-τ)d/∥A∥-1τ-αdτ=c1∥x0∥1+∥A∥tα+dc1∥x0∥∥A∥c/α∥A∥-2×Γ(d/∥A∥)Γ(1-α)Γ(1+d/∥A∥-α)td/∥A∥-α.
Since α∥A∥>d, from (27) we can conclude that
(28)limt→∞∥x(t)∥=0,
and this ends the proof.
3.2. Fractional Order α: 1<α<2
Let the initial conditions be x(i)(0)=xi, i=1,2. Then we have the following results.
Theorem 8.
Consider system (14). If we choose matrices L and K such that |
arg
(
eig
(A))|>απ/2 and (α-1)∥A∥>d, then system (14) is asymptotically stable. d is a positive constant.
Proof.
Similar to the proof of Theorem 8, taking the Laplace transform and Laplace inverse transform on (14) gives
(29)x(t)=Eq,1(A(t)α)x1+∫0t(t-τ)α-1Eα,α(A(t-τ)α)f(x)dτ+∫0t(t-τ)α-1Eα,α(A(t-τ)α)φ(t,x)dτ+tEα,2(Atα)x2.
According to Lemma 2, there exist positive constants ci,i=1,2,3, such that
(30)∥x(t)∥≤c1∥x1∥1+∥A∥tα+∫0t(t-τ)α-1c21+∥A∥(t-τ)α∥f(x)∥dτ+∫0t(t-τ)α-1c21+∥A∥(t-τ)α∥φ(t,x)∥dτ+c3t∥x2∥1+∥A∥tα.
From Definition 4 and Lemma 5, we know that φ(t,x(t)) satisfies
(31)∥φ(t,x(t))∥≤∥K-LK∥∥x(t)∥.
Let d=Mc2+c2∥K-LK∥; we can obtain
(32)∥x(t)∥≤c1∥x1∥+c3t∥x1∥1+∥A∥tα+d∫0t(t-τ)α-11+∥A∥(t-τ)α∥x(t)∥dτ.
Then we have
(33)∥x(t)∥≤c1∥x1∥+c3t∥x1∥1+∥A∥tα+∫0td(t-τ)α-1(c1∥x1∥+c3t∥x2∥)(1+∥A∥(t-τ)α)(1+∥A∥τα)hhhhhhhh×exp(∫τtd(t-y)α-11+∥A∥(t-y)αdy)dτ=c1∥x1∥+c3t∥x1∥1+∥A∥tα+∫0td(t-τ)α-1(c1∥x1∥+c3∥x2∥)(1+∥A∥τα)(1+∥A∥(t-τ)α)1-c/α∥A∥dτ≤c1∥x1∥+c3t∥x1∥1+∥A∥tα+dc1∥x1∥∥A∥d/α∥A∥-2∫0t(t-τ)1/∥A∥-1τ-αdτ+dc3∥x2∥∥A∥d/α∥A∥-2∫0t(t-τ)1/∥A∥-1τ1-αdτ=c1∥x1∥+c3t∥x1∥1+∥A∥tα+dc1∥x1∥∥A∥d/α∥A∥-2×Γ(d/∥A∥)Γ(1-α)Γ(1+d/∥A∥-α)td/∥A∥-α+dc3∥x2∥∥A∥d/α∥A∥-2×Γ(d/∥A∥)Γ(2-α)Γ(2+d/∥A∥-α)td/∥A∥+1-α.
Since (α-1)∥A∥>d, from (33) we know that
(34)limt→∞∥x(t)∥=0,
and this ends the proof.
4. Simulation Studies
The system (14) has 3 equilibriums:
(35)E1=(0;1b;0),E2=(c-b-abcc;1+acc;-1cc-b-abcc),E3=(-c-b-abcc;1+acc;1cc-b-abcc).
The Jacobian matrix of the fractional-order chaotic system (14), at the equilibrium E*=[x*,y*,z*]T, can be given as
(36)JE=[-a+y*x*1-2x*-b0-10-c].
Let a=1,b=0.1,andc=1. The eigenvalues for the system equilibrium E1=(0;10;0) are λ1=8.8990,λ2=-0.8990,andλ3=-0.1. And it is a saddle point. For equilibrium points E2=(0.8944;2;-0.8944) and E3=(-0.8944;2;0.8944) they are λ1=-0.7609 and λ2,3=0.3304±1.4112i. It is a saddle-focus point. Since it is an unstable equilibrium, the condition for chaos is satisfied and the system (14) can show chaotic behavior. We can easily get the minimal commensurate order of the system which is α>0.8537.
Case 1 (0<α≤1).
Assume the fractional order is α=0.9. The characteristic equation of the linearized system for the equilibrium E1 is
(37)λ27-7.9λ18-8.8λ9-0.8=0.
The characteristic equation of the linearized system for the equilibriums E2 and E3 is
(38)λ27+0.1λ18+1.6λ9+1.6=0
and the unstable eigenvalues are λ1,2=1.0306±0.1547i.
Let the initial condition be x(0)=[2,-1,1]T. The chaotic behavior of uncontrolled fractional-order financial system (13) is shown in Figure 2.
Chaotic attractor of fractional-order financial system with fractional order α=0.90: (a) x1-x2 plane; (b) x1-x3 plane.
The control gain matrices are chosen as L=diag[0.9,0.9,0.9].
Let
(39)A=[010001-2-2-3];
then we have ∥A∥=4.2160. From simulation results (see Figure 2), we know |x1(t)|≤3. From A=P+LK, we have
(40)K=[1.11111.1111-1.111100.11111.1111-1.1111-2.2222-2.2222].
From above discussion, we know ∥A∥=4.2160 and |arg(eig(A))|=[3.1416,1.8428,1.8428]T. Then we can easily test that the conditions |arg(eig(A))|>απ/2 and α∥A∥>d are satisfied.
Let u01=u02=u03=2. The simulation results can be seen in Figures 3 and 4. From the results, we can see that the states variables converge rapidly. The involved system is asymptotic stable. Figure 4 shows the boundedness and smoothness of the saturated control inputs. It can be concluded that good control performance has been achieved.
Time response of the state variables with fractional order α=0.90.
Time response of the control inputs with fractional order α=0.90.
Case 2 (1<α<2).
Let the fractional order be α=1.04. The chaotic behavior is depicted in Figure 5. In the simulation, the control gain matrices are chosen as L=diag[0.9,0.9,0.9],
(41)K=[1.11111.1111-1.111100.11111.1111-4.4444-6.6667-6.6667];
then we have
(42)A=P+LK=[010001-5-6-7].
From above discussion, we know ∥A∥=10.5249 and |arg(eig(A))|=[3.1416,2.0573,2.0573]T. Then we have |arg(eig(A))|>απ/2 and α∥A∥>d.
Chaotic attractor of fractional-order financial system with fractional order α=1.04: (a) x1-x2 plane; (b) x1-x3 plane.
Let u01=u02=u03=2. The simulation results can be seen in Figures 6 and 7. From the simulation results we can conclude that good control performance has been achieved.
Time response of the state variables with fractional order α=1.04.
Time response of the control inputs with fractional order α=1.04.
5. Conclusions
We investigate the control problem for fractional-order financial chaotic systems subject to input saturation by means of linear control. Two sufficient conditions are given for the stabilization of such systems with fractional orders 0<α≤1 and 1<α<2, respectively. A state-feedback controller is designed and the asymptotical stability of the involved system is guaranteed. It is shown that state-feedback controller can be designed to control the fractional-order financial chaotic systems. Simulation studies confirm the results of this paper.
Conflict of Interests
The authors do not have a direct financial relation with any commercial identity mentioned in their paper that might lead to a conflict of interests for any of the authors.
Acknowledgments
This work is supported by the National Natural Science Foundation of China (Grant no. 61001086) and the Fundamental Research Funds for the Central Universities (Grant no. ZYGX2011X004).
ChenW. C.Nonlinear dynamics and chaos in a fractional-order financial system20083651305131410.1016/j.chaos.2006.07.0512-s2.0-38649117916CaiG.HuangJ.A new finance chaotic attractor200733213220MR2314282WenF.YangX.Skewness of return distribution and coefficient of risk premium200922336037110.1007/s11424-009-9170-xMR25388682-s2.0-71449108473HuangC.PengC.ChenX.WenF.Dynamics analysis of a class of delayed economic model2013201312962738MR303917310.1155/2013/962738HuangC.GongX.ChenX.WenF.Measuring and forecasting volatility in Chinese stock market using HAR-CJ-M model201320131314319410.1155/2013/1431942-s2.0-84876515221WenF.LiuZ.A copula-based correlation measure and its application in chinese stock market20098478780110.1142/S0219622009003612ZBL1186.912362-s2.0-77149164766WenF.GongX.ChaoY.ChenX.The effects of prior outcomes on risky choice: evidence from the stock market20142014827251810.1155/2014/272518MR3193662RiveroM.TrujilloJ. J.VázquezL.VelascoM. P.Fractional dynamics of populations201121831089109510.1016/j.amc.2011.03.017MR28313622-s2.0-80052273583WangZ.HuangX.ShenH.Control of an uncertain fractional order economic system via adaptive sliding mode201283838810.1016/j.neucom.2011.11.0182-s2.0-84856604260CaoJ.HoD. W. C.YangY.Projective synchronization of a class of delayed chaotic systems via impulsive control2009373353128313310.1016/j.physleta.2009.06.056MR2562277ZBL1233.340172-s2.0-67651211662CaoJ. D.LiL. L.Cluster synchronization in an array of hybrid coupled neural networks with delay200922433534210.1016/j.neunet.2009.03.0062-s2.0-67349144397CaoJ.AlofiA.Al-MazrooeiA.ElaiwA.Synchronization of switched interval networks and applications to chaotic neural networks2013201311940573MR314357310.1155/2013/940573ZhangH.CaoJ.JiangW.Controllability criteria for linear fractional differential systems with state delay and impulses201320139146010MR306492310.1155/2013/146010CaoJ. D.WanY.Matrix measure strategies for stability and synchronization of inertial BAM neural network with time delays201453165172TavazoeiM. S.HaeriM.Synchronization of chaotic fractional-order systems via active sliding mode controller20083871577010.1016/j.physa.2007.08.0392-s2.0-35748971445TavazoeiM. S.HaeriM.Chaotic attractors in incommensurate fractional order systems2008237202628263710.1016/j.physd.2008.03.037MR25141222-s2.0-51249088596XuY. F.HeZ. M.Synchronization of variable-order fractonal financial systems via active control method2013116824835WangZ.HuangX.ShiG. D.Analysis of nonlinear dynamics and chaos in a fractional order financial system with time delay20116231531153910.1016/j.camwa.2011.04.057MR28247402-s2.0-79960978682LimY.-H.OhK.-K.AhnH.-S.Stability and stabilization of fractional-order linear systems subject to input saturation20135841062106710.1109/TAC.2012.2218064MR30388122-s2.0-84875633742GußnerT.JostM.AdamyJ.Controller design for a class of nonlinear systems with input saturation using convex optimization201261125826510.1016/j.sysconle.2011.11.0032-s2.0-82955189187LuoJ.State-feedback control for fractional-order nonlinear systems subject to input saturation20142014889163910.1155/2014/891639MR3166928PodlubnyI.1999198San Diego, Calif, USAMathematics in Science and Engineering, Academic PressAhnH. S.ChenY. Q.Necessary and sufficient stability condition of fractional-order interval linear systems200844112985298810.1016/j.automatica.2008.07.003MR25272302-s2.0-55249087154ChenL.ChaiY.WuR.YangJ.Stability and stabilization of a class of nonlinear fractional-order systems with caputo derivative201259960260610.1109/TCSII.2012.22069362-s2.0-84866493921