This paper investigates the finite-time control problem for discrete-time Markov jump systems subject to saturating actuators. A finite-state Markovian process is given to govern the transition of the jumping parameters. The finite-time H∞ controller via state feedback is designed to guarantee that the resulting system is mean-square locally asymptotically finite-time stabilizable. Based on stochastic finite-time stability analysis, sufficient conditions that ensure stochastic control performance of discrete-time Markov jump systems are derived in the form of linear matrix inequalities. Finally, a numerical example is provided to illustrate the effectiveness of the proposed approach.
1. Introduction
During the past several decades, the issue of finite-time control has drawn increasing attention of academic researchers in the area of control field, and various results have been reported. To this end, a considerable amount of research has been carried out; see Hong et al. [1]; He and Liu [2, 3]; Li et al. [4]; Song et al. [5]; Lan et al. [6]. Among the proposed solutions, state feedback control is an important approach to improve finite-time control performance. For instance, by using both state feedback and dynamic output feedback control, finite-time control of the robot system is studied in Hong et al. [7]. Furthermore, based on dynamic observer-based state feedback and Lyapunov-Krasovskii functional approach, the finite-time H∞ control problem for time-delay nonlinear jump systems was addressed in the work of He and Liu [2, 3, 8].
On the other hand, more and more attention has been paid to the study of actuator saturation due to its practical and theoretical importance. Therefore, various approaches were investigated to handle systems with actuator saturation, such as in the work of Cao and Lin [9]; the stability of discrete-time systems with actuator saturation was analyzed by a saturation-dependent Lyapunov function. By introducing a time-varying sliding surface, the robust stabilization problem of linear unstable plants with saturating actuators was studied in Corradini and Orlando [10]. Furthermore, the controller design method of Markov jumping systems subject to actuator saturation was presented in Liu et al. [11]. Via dynamic anti-windup fuzzy design, the robust stabilization problem of state delayed T-S fuzzy systems with input saturation was proposed in Song et al. [12]. Other results can refer to [13–17] and references therein.
It is well known that the control problem of Markov jump systems has also been extensively studied and a large variety of control problems have been widely investigated, for instance, the stabilization of Markov jump systems with time delays [18–24], robust control [25], control of singular Markov systems [26], control of discrete-time stochastic Markov jump systems [27, 28], and fuzzy dissipative control for nonlinear Markovian jump systems [29]. Furthermore, robust stability for uncertain delayed neural networks with Markov jumping parameters was analyzed in Li et al. [30]. Robust H∞ filter was designed for uncertain discrete Markov jump singular systems with mode-dependent time delay in Ma and Boukas [31]. Delay-dependent robust stabilization problem for uncertain stochastic switching systems with distributed delays was studied in Shen et al. [32]. Via retarded output feedback, passivity-based control problem for Markov jump systems was addressed in Shen et al. [33]. Observer based finite-time H∞ control problem of discrete-time Markov jump systems was studied in Zhang and Liu [34]. However, to the best of our knowledge, the problem of finite-time stabilization of discrete-time stochastic systems has not been fully investigated and it is the main purpose of our study.
In this paper, the attention is focused on the finite-time H∞ control problem of discrete-time Markov jump systems with actuator saturation. A state feedback controller is designed to ensure the stochastic finite-time boundedness and stochastic finite-time stabilization of the resulting closed-loop system for all admissible disturbances. The desired controller can be designed via solving a convex optimization problem. Finally, a numerical example is employed to show the effectiveness of the proposed method.
Notation. Throughout the paper, for symmetric matrices X and Y, the notation X≥Y (resp., X>Y) means that the matrix X-Y is positive semidefinite (resp., positive definite). I is the identity matrix with appropriate dimension. The notation NT represents the transpose of the matrix N;λmax(M) (resp., λmin(M)) means the largest (resp., smallest) eigenvalue of the matrix M; (Ω,ℱ,𝒫) is a probability space; Ω is the sample space, ℱ is the σ-algebra of subsets of the sample space, and 𝒫 is the probability measure on ℱ;ℰ{·} denotes the expectation operator with respect to some probability measure 𝒫. Matrices, if not explicitly stated, are assumed to have compatible dimensions. The symbol * is used to denote a matrix which can be inferred by symmetry, He{A}=AT+A.
2. Preliminaries and Problem Description2.1. Preliminaries
Throughout this paper, we will use the following definitions and lemmas.
Lemma 1 (see [12]).
For the matrix Ki and the system Σ, the appropriate matrix Li∈ℝm×n is given if x(k) is in the set D(uo), where D(uo) is defined as follows:
(1)D(uo)={x(k)∈ℝn;-u0(k)≤(Ki(k)-Li(k))x(t)≤u0(k),oo1u0(k)>0,k=1,…,m};
then for any diagonal positive matrix T∈ℝm×m, we derive
(2)ψ(u(k))TT(ψ(u(k))-Lix(k))≤0.
Lemma 2 (see [32]).
For the given symmetric matrix S∈ℝ(n+m)×(n+m):
(3)S=[S11S12S12TS22],
where S11∈ℝn×n, S12∈ℝn×m, and S22∈ℝm×m; the following conditions are equivalent:
S<0,
S11<0,S22-S12TS11-1S12<0,
S22<0,S11-S12S22-1S12T<0.
Definition 3 (see [34]).
The resulting closed-loop system (12) is stochastic finite-time stable (SFTB) with respect to (δx,ϵ,Ri,N,d) with 0<δx<ϵ, Ri>0, and N∈Zk≥0 if there exists state feedback controller such that
(4)E{xT(0)Rix(0)}≤δx2⟹E{xT(k)Rix(k)}<ϵ2,∀k∈{1,2,…,N}.
Definition 4 (see [34]).
The resulting closed-loop system (12) is said to be stochastic H∞ finite-time stable via state feedback with respect to (δx,ϵ,Ri,N,γ,d) with 0<δx<ϵ, Ri>0, γ>0, and N∈Zk≥0 if the system (11)-(12) is SFTB with respect to (δx,ϵ,Ri,N,γ,d) and under the zero-initial condition the output z(k) satisfies
(5)E{Σj=0NzT(j)z(j)}≤γ2E{Σj=0NwT(j)w(j)},
for any nonzero w(k) which satisfies (10), where γ is a prescribed positive scalar. Moreover, the state feedback controller (11) is called H∞ controller of MJS (12).
2.2. Problem Description
Consider the following discrete-time Markov jump system (Σ) in the probability space (Ω,ℱ,𝒫):
(6)x(k+1)=A(r(k))x(k)+B(r(k))sat(u(k))+G(r(k))w(k),z(k)=C(r(k))x(k)+D1(r(k))sat(u(k))+D2(r(k))w(k),
where x(k)∈ℝn is the state vector, z(k)∈ℝl is the controlled output, and sat(u(k))∈ℝm is the saturated control input. w(k)∈L2p[0+∞) is the external disturbances. {r(k)} is a discrete-time Markov process and takes values from a finite set S={1,2,…,𝒩} with transition probabilities given by
(7)Pr(rk+1=j∣rk=i)=πij,
where πij≥0, for ∀j,i∈S, and Σj∈Sπij=1. Moreover, the transition rates matrix of the system (Σ) is defined by
(8)[π11π12⋯π1𝒩π21π22⋯π2𝒩⋮⋮⋱⋮π𝒩1π𝒩2⋯π𝒩𝒩].
The inputs of the plant are supposed to be bounded as follows:
(9)-u0(k)≤u(k)≤u0(k),u0(k)>0,k=1,…,m.
For the system (Σ), to simplify the notation, we denote Ai=A(r(k)) for each r(k)=i∈S, and the other symbols are similarly denoted. Ai, Bi, Gi, Ci, D1i, and D2i are known mode-dependent constant matrices with appropriate dimensions.
Assumption 5 (see [34]).
The external disturbance w(k) is varying and satisfies the following constraint condition:
(10)Σk=0Tw(k)Tw(k)≤d,d≥0.
For the system (Σ), we construct the following state feedback controller:
(11)u(k)=K(r(k))x(k).
Then, the resulting closed-loop discrete-time Markov jump system (MJS) is as follows:
(12)x(k+1)=(Ai+BiKi)x(k)+Biψ(u(k))+Giw(k),z(k)=(Ci+D1iKi)x(k)+D1iψ(u(k))+D2iw(k),
where ψ(u(k))=sat(u(k))-u(k).
3. Main Results
In this section, we investigate the design of a state feedback controller which guarantees the locally finite-time stabilizable of the resulting closed-loop system. Some sufficient conditions and the method of designing state feedback controller are given.
Theorem 6.
For each r(k)=i∈S, there exists a feedback controller u(k)=Kix(k), Ki=YiXi-1, such that the resulting closed-loop system (12) is SFTB with respect to (δx,ϵ,Ri,N,d) with 0<δx<ϵ, if there exist scalars μ≥0, σ1≥0, and σ2≥0, three sets of mode-dependent symmetric matrices Xi>0, Ji>0, and Qi>0, and two sets of mode-dependent matrices Yi and L-i=LiXi, such that the following conditions hold:
(13)[-μXi0L-iTL-1iT*-Qi0L2iT**-2JiL3iT***-W]<0,(14)[σ2d2-μ-Nϵ2*δx-σ1]<0,(15)[Xi*Yi+L-iu0(k)2]>0,k=1,…,m,(16)σ1Ri-1<Xi<Ri-1,(17)0<Qi<σ2I,
where
(18)W=diag{X1,X2,…,Xn},L-1iT=[πi1(XiAiT+YiTBT)πi2(XiAiT+YiTBT)⋯πin(XiAiT+YiTBT)],L2iT=[πi1GiTπi2GiT⋯πinGiT],L3iT=[πi1JiBiTπi2JiBiT⋯πinJiBiT].
Proof.
Define the following Lyapunov function for each δ(t)=i∈S:
(19)V(k)=x(k)TPix(k).
It is readily obtained that
(20)E{V(k+1)}=E{Σj=1nπijx(k+1)TPjx(k+1)}=ξ(k)T[L1iL2iL3i]TW-[L1iL2iL3i]ξ(k),
where
(21)ξ(k)=[x(k)Tw(k)Tψ(k)T],W-=diag{P1,P2,…,Ph},L1iT=[πi1(Ai+BiKi)Tπi2(Ai+BiKi)T⋯πin(Ai+BiKi)T],L2iT=[πi1GiTπi2GiT⋯πinGiT],L3iT=[πi1BiTπi2BiT⋯πinBiT].
Then by pre- and postmultiplying (13) by diag{Pi,I,Ti,I} with Pi=Xi-1, Ti=Ji-1, we have
(22)[-μPi0LiTTiL1iT*-Qi0L2iT**-2TiL3iT***-W]<0.
By using Schur complement lemma, we derive
(23)ξ(k)T[-μPi0LiTTi*-Qi0**-2Ti]×ξ(k)+ξ(k)T[L1iL2iL3i]TW-[L1iL2iL3i]ξ(k)<0.
It follows that
(24)E{V(k+1)}<μx(k)TPix(k)+w(k)TQiw(k)+2ψ(k)TTiψ(k)-2ψ(k)TTiLix(k).
Since 2ψ(k)TTiψ(k)-2ψ(k)TTiLix(k)≤0, we get
(25)E{V(k+1)}<μx(k)TPix(k)+w(k)TQiw(k).
It is shown that
(26)E{V(k+1)}<μV(k)+sup(i∈S){λmax(Qi)}w(k)Tw(k).
Then we have
(27)E{V(k+1)}<μE{V(k)}+sup(i∈S){λmax(Qi)}E{w(k)Tw(k)}.
Since μ≥1, it is easily found that
(28)E{V(k+1)}<μE{V(0)}+sup(i∈S){λmax(Qi)}E×{∑j=0k-1μk-j-1w(j)Tw(j)}≤μkE{V(0)}+sup(i∈S){λmax(Qi)}μkd2.
Let
(29)P-i=Ri-1/2PiRi-1/2,
and noting that
(30)E{xT(0)Rix(0)}≤δx2,
it can be verified that
(31)E{V(0)}=E{xT(0)Pix(0)}=E{xT(0)Ri1/2P-iRi1/2x(0)}≤supi∈S{λmax(P-i)}E{xT(0)Rix(0)}≤supi∈S{λmax(P-i)}δx2.
Similarly, for all i∈S, we can obtain
(32)E{V(k)}=E{xT(k)Pix(k)}=E{xT(k)Ri1/2P-iRi1/2x(k)}≥infi∈S{λmin}(P-i)E{xT(k)Rix(k)}.
Then it is not difficult to find that
(33)E{xT(k)Rix(k)}<supi∈S{λmax(P-i)}μkδx2+sup(i∈S){λmax(Qi)}μkd2infi∈S{λmin}(P-i),
which implies
(34)supi∈S{λmax(P-i)}μkδx2+sup(i∈S){λmax(Qi)}μkd2infi∈S{λmin}(P-i)<ϵ2.
Then, one can obtain that
(35)supi∈S{λmax(P-i)}δx2+sup(i∈S){λmax(Qi)}d2<infi∈S{λmin}(P-i)μ-Nϵ2.
Set
(36)Xi=Pi-1,σ1Ri-1<Xi<Ri-1,0<Qi<σ2I;
it is easy to see that
(37)σ1-1δx2+σ2d2<μ-Nϵ2.
It is obvious that (37) is equivalent to (14). Based on Lemma 1, it is easy to obtain condition (15). This completes the proof.
Theorem 7.
For each r(k)=i∈S, there exists a feedback controller u(k)=Kix(k), Ki=YiXi-1, such that the resulting closed-loop system (12) is said to be stochastic H∞ finite-time stable via state feedback with respect to (δx,ϵ,Ri,N,γ,d), if there exist three scalars μ≥0, σ1≥0, and γ≥0, two sets of mode-dependent symmetric matrices Xi>0 and Ji>0, and two sets of mode-dependent matrices Yi and L-i=LiXi, such that the following conditions hold:
(38)[-μXi0L-iL-1iYiTDiT+XiCiT*-γ2μ-NI0L2iTD2iT**-2JiL3iTD1iT***-W0****-I]<0,(39)[μ-N(d2γ2-ϵ2)*δx-σ1]<0,(40)[Xi*Yi+L-iu0(k)2]>0,k=1,…,m,(41)σ1Ri-1<Xi<Ri-1.
Proof.
Choose the similar Lyapunov function as Theorem 6 and denote
(42)Π(x(k),w(k),r(k)=i)=E{V(k+1)}-μV(k)+z(k)Tz(k)-γ2μ-Nw(k)Tw(k).
Thus, in the light of Theorem 6, we have
(43)Π(x(k),w(k),rk=i)≤ξ(k)T[L1iL2iL3i]TW-[L1iL2iL3i]ξ(k)+ξ(k)T[Ci+D1iKiD2iD3i]T×[Ci+D1iKiD2iD3i]ξ(k)+ξT(k)[-μXi0LiTTi*-γ2μ-NI0**-2Ti]ξ(k).
Then by pre- and postmultiplying (38) by diag{Pi,I,Ti,I} and considering Schur complement Lemma and (43), we derive
(44)Π(x(k),w(k),r(k)=i)<0,
holds for all rk=i∈S. According to (44), one can obtain that
(45)E{V(k+1)}<μE{V(k)}-E{z(k)Tz(k)}+γ2μ-NE{w(k)Tw(k)}.
Then, we have
(46)E{V(k)}<μkE{V(0)}-Σj=0k-1μk-j-1E{z(j)Tz(j)}+γ2μ-NE{Σj=0k-1μk-j-1w(j)Tw(j)}.
Under the zero-value initial condition and noting that V(k)≥0, for all K∈Zk≥0, it is shown that
(47)Σj=0k-1μk-j-1E{z(j)Tz(j)}<γ2μ-NE{Σj=0k-1μk-j-1w(j)Tw(j)}.
Since μ≥1 and from (47), we have
(48)E{Σj=0Nz(j)Tz(j)}=Σj=0NE{z(j)Tz(j)}≤Σj=0NE{μN-jz(j)Tz(j)}≤γ2μ-NE{Σj=0NμN-jw(j)Tw(j)}≤γ2E{Σj=0Nw(j)Tw(j)}.
The following proof is similar to the process of Zhang and Liu [34].
Since ϵ(Pi,1)⊂D(u0), it follows that
(49)[Pi*Ki+Liu0(k)2]>0,k=1,…,m,
and then by pre- and postmultiplying (49) by diag(Xi,I,) and its transpose, respectively, we derive condition (40). This completes the proof.
Remark 8.
For the given scalars (δx,ϵ,Ri,N,γ,d), we can take γ2 as the optimized variable to obtain an optimized finite-time stabilized controller. The attenuation lever γ2 can be reduced to the minimum possible value such that LMIs (38)–(41) hold. The optimization problem can be described as follows:
(50)minρkkkki(XiYiKiδx)s.t.LMIs(38)–(41)withρ=γ2.
4. Illustrative Examples
In this section, a numerical example is provided to demonstrate the effectiveness of the proposed method. Consider the following systems with four operation modes.
The transition rate matrix is given as follows:
(55)[0.40.30.20.10.30.40.10.20.10.20.40.30.20.10.30.4].
In this case, we choose the initial values for Ri=I2, i=1,2,3,4, δx=1, N=5, α=10-10, μ=2.5, and d=1; Theorem 6 yields to ϵ=36.2671, σ1=0.4906, σ2=13.7421, and the bounds of the input saturation u0=0.05.
Based on Theorem 7, we derive
(56)K1=[-0.77230.5862],K2=[-0.1021-0.0506],K3=[-0.87060.5591],K4=[-0.3019-0.3635].
Remark 9.
The figures are given on the last page. In this part Figure 1 is rk of the jump rates; Figure 2 shows the states of the open-loop Markovian jump system; Figure 3 shows the states of the closed-loop Markovian jump system. By applying the controller studied in this paper to the closed-loop plant, it is obviously noticed that x1 and x2 converge to zero quickly. Based on the figures provided, the controller we designed guarantees that the resulting closed-loop systems are mean-square locally asymptotically finite-time stabilizable.
rk of jump rates.
x(k) of open-loop system.
x(k) of closed-loop system.
5. Conclusions
This paper considers the finite-time H∞ stabilization problem for a class of discrete-time Markov jump systems with input saturation. The finite-time H∞ controller via state feedback is designed to guarantee the stochastic finite-time boundedness and stochastic finite-time stabilization of the considered closed-loop system for all admissible disturbances. Based on stochastic finite-time stability analysis, sufficient conditions are derived in the form of linear matrix inequalities. Finally, simulation results are given to illustrate the effectiveness of the proposed approach. In the future, we will study the finite-time stabilization problem for a class of Markov jump systems with constrained input and time delay.
Conflict of Interests
The authors declare that there is no conflict of interests regarding the publication of this paper.
Acknowledgments
This work was supported by the National Natural Science Foundation of China under Grant 61203047, the Natural Science Foundation of Anhui Province under Grant 1308085QF119, and the Key Foundation of Natural Science for Colleges and Universities in Anhui province under Grant KJ2012A049.
HongY.WangJ.ChengD.Adaptive finite-time control of nonlinear systems with parametric uncertainty20065158588622-s2.0-3364691568610.1109/TAC.2006.875006HeS.LiuF.Finite-time H∞ filtering of time-delay stochastic jump systems with unbiased estimation201022489479592-s2.0-7865008047410.1243/09596518JSCE1006HeS.LiuF.Observer-based finite-time control of time-delayed jump systems20102176232723382-s2.0-7795800221210.1016/j.amc.2010.07.031MR2733675ZBL1207.93113LiS.LiuH.DingS.A speed control for a PMSM using finite-time feedback control and disturbance compensation20103221701872-s2.0-7795195508310.1177/0142331209339860SongH.YuL.ZhangD.ZhangW.-A.Finite-time H∞ control for a class of discrete-time switched time-delay systems with quantized feedback201217124802481410.1016/j.cnsns.2012.05.002MR2960277ZBL1263.93072LanQ.LiS.YangJ.GuoL.Finite-time soft landing on asteroids using nonsingular terminal sliding mode control201310.1177/0142331213495040HongY.XuY.HuangJ.Finite-time control for robot manipulators20024642432532-s2.0-003716241710.1016/S0167-6911(02)00130-5MR2010242ZBL0994.93041HeS.LiuF.Finite-time H∞ fuzzy control of nonlinear jump system with time delays via dynamic observer-based state feedback201220460561410.1109/TFUZZ.2011.2177842CaoY.-Y.LinZ.Stability analysis of discrete-time systems with actuator saturation by a saturation-dependent Lyapunov function2003397123512412-s2.0-003784041210.1016/S0005-1098(03)00072-4MR2140854ZBL1139.93342CorradiniM. L.OrlandoG.Linear unstable plants with saturating actuators: robust stabilization by a time varying sliding surface200743188942-s2.0-3375120149910.1016/j.automatica.2006.07.018MR2266773ZBL1140.93467LiuH.BoukasE.-K.SunF.HoD. W. C.Controller design for Markov jumping systems subject to actuator saturation20064234594652-s2.0-3114446681110.1016/j.automatica.2005.10.017MR2195249ZBL1121.93027SongX.LuJ.XuS.ShenH.LuJ.Robust stabilization of state delayed T-S fuzzy systems with input saturation via dynamic anti-windup fuzzy design2011712666566762-s2.0-80455137246ChenB. M.LeeT. H.PengK.VenkataramananV.Composite nonlinear feedback control for linear systems with input saturation: theory and an application20034834274392-s2.0-003735106110.1109/TAC.2003.809148MR1962250LanW.HuangJ.Semiglobal stabilization and output regulation of singular linear systems with input saturation2003487127412802-s2.0-004176493510.1109/TAC.2003.814276MR1988104HuT.TeelA. R.ZaccarianL.Anti-windup synthesis for linear control systems with input saturation: achieving regional, nonlinear performance20084425125192-s2.0-3784902855910.1016/j.automatica.2007.06.003MR2530800AraujoJ. M.BarrosP. R.DóreaC. E. T.Dynamic output feedback control of constrained descriptor systems20133581129113810.1177/0142331213487921ShenH.SongX.WangZ.Robust fault-tolerant control of uncertain fractional-order systems against actuator faults2013791233124110.1049/iet-cta.2012.0822MR3100550GaoH.ChenT.New results on stability of discrete-time systems with time-varying state delay20075223283342-s2.0-3394738061710.1109/TAC.2006.890320MR2295017ParkP.Wan KoJ.Stability and robust stability for systems with a time-varying delay20074310185518582-s2.0-3454853256210.1016/j.automatica.2007.02.022MR2572899ZBL1120.93043ShenH.XuS.LuJ.ZhouJ.Passivity-based control for uncertain stochastic jumping systems with mode-dependent round-trip time delays20123495166516802-s2.0-8486027546710.1016/j.jfranklin.2011.11.011MR2917548ZBL1254.93148WuM.HeY.SheJ.-H.LiuG.-P.Delay-dependent criteria for robust stability of time-varying delay systems2004408143514392-s2.0-294260460810.1016/j.automatica.2004.03.004MR2153058ZBL1059.93108WuZ.ShiP.SuH.ChuJ.Passivity analysis for discrete-time stochastic Markovian jump neural networks with mixed time delays20112210156615752-s2.0-8005365509710.1109/TNN.2011.2163203WuZ.ShiP.SuH.ChuJ.Stability analysis for discrete-time Markovian jump neural networks with mixed time-delays2012396617461812-s2.0-8485651714610.1016/j.eswa.2011.11.111WuZ.ShiP.SuH.ChuJ.Stochastic synchronization of Markovian jump neural networks with time-varying delay using sampled-data20134361796180610.1109/TSMCB.2012.2230441LiuH.SunF.BoukasE. K.Robust control of uncertain discrete-time Markovian jump systems with actuator saturation20067978058122-s2.0-3364677988010.1080/00207170600688842MR2231230ZBL05062605RaoufJ.BoukasE. K.Stabilisation of singular Markovian jump systems with discontinuities and saturating inputs2009379719822-s2.0-6765032154310.1049/iet-cta.2007.0451WuZ.ShiP.SuH.ChuJ.Asynchronous l2−l∞ filtering for discrete-time stochastic Markov jump systems with randomly occurred sensor nonlinearities201450118018610.1016/j.automatica.2013.09.041ShenH.ParkJ. H.ZhangL.WuZ.-G.Robust extended dissipative control for sampled-data Markov jump systems201410.1080/00207179.2013.878478ShenH.WangZ.HuangX.WangJ.Fuzzy dissipative control for nonlinear Markovian jump systems via retarded feedback201310.1016/j.jfranklin.2013.02.031LiH.ChenB.ZhouQ.QianW.Robust stability for uncertain delayed fuzzy hopfield neural networks with Markovian jumping parameters20093919410210.1109/TSMCB.2008.2002812MaS.BoukasE. K.Robust H∞ filtering for uncertain discrete Markov jump singular systems with mode-dependent time delay2009333513612-s2.0-6184915331710.1049/iet-cta:20080091MR2502457ShenH.XuS.SongX.LuoJ.Delay-dependent robust stabilization for uncertain stochastic switching systems with distributed delays20091155275352-s2.0-7244912077810.1002/asjc.133MR2567084ShenH.XuS.SongX.ShiG.Passivity-based control for Markovian jump systems via retarded output feedback20123111892022-s2.0-8485668969010.1007/s00034-011-9328-3MR2881981ZBL1242.93040ZhangY.LiuC.Observer-based finite-time H∞ control of discrete-time Markovian jump systems20133763748376010.1016/j.apm.2012.07.060MR3020529ZBL1270.93046