INT J COMPUT COMMUN, ISSN 1841-9836 Vol.7 (2012), No. 4 (November), pp. 776-784 Stability of Discrete-Time Systems with Time-Varying Delay: Delay Decomposition Approach S.B. Stojanovic, D.LJ. Debeljkovic, N. Dimitrijevic Sreten B. Stojanovic University of Nis, Faculty of Technology Serbia, 16000 Leskovac, Bulevar oslobodjenja 124 E-mail: ssreten@ptt.rs Dragutin LJ. Debeljkovic, Nebojsa Dimitrijevic University of Belgrade, Faculty of Mechanical Engineering Serbia, 11120 Beograd, Kraljice Marije 16 E-mail: ddebeljkovic@mas.bg.ac.rs, ndimitri@verat.net Abstract: This article deals with the problem of obtaining delay-dependent stability conditions for a class of discrete-time systems with interval time-varying delay. Using the de- composition the delay interval into two unequal subintervals by tuning parameter α, a new interval delay-dependent Lyapunov-Krasovskii functional is constructed to de- rive novel delay-dependent stability conditions which are expressed in terms of linear matrix inequalities. This leads to reduction of conservatism in terms of the upper bounds of the maximum time-delay. The numerical examples show that the obtained result is less conservative than some existing ones in the literature. Keywords: time-delay systems, interval time-varying delay, asymptotic stability, delay-dependent stability, Lyapunov-Krasovskii methods. 1 Introduction Time-delay frequently occurs in many practical systems, such as manufacturing systems, telecommunication and economic systems etc. Since time-delay is an important source of insta- bility and poor performance, considerable attention has been paid to the problem of stability analysis and controller synthesis for continuous time-delay systems (see e.g. [3-5, 10, 11, 17-21, 23-26] and the reference therein). Inversely, less attention has been drawn to the corresponding results for discrete-time delay systems (see e.g. [1, 2, 6-9, 12-15, 22, 24]). This is mainly due to the fact that such systems can be transformed into augmented systems without delay. This augmentation of the system is, however, inappropriate for systems with unknown delays and for systems with time-varying delay which are the subject analysis in this work. Recently, increasing attention has been devoted to the problem of delay-dependent stability of linear systems with time-varying delay, including continuous-time (see e.g. [5, 10, 11, 18-21, 23], 25, 26]) and discrete-time systems (see e.g. [1, 2, 7-9, 12, 14, 15, 24]) and a great number of delay-dependent stability criteria were derived. The key point for deriving the delay-dependent stability criterions is the choice of an appropriate Lyapunov–Krasovskii functional (LKF). It is known that the existence of a complete quadratic Lyapunov-Krasovskii functional (CQLKF) is a sufficient and necessary condition for asymptotic stability of the time-delay system. Using the CQLKF, one can obtain the maximum allowable upper bound (MAUB) of delay which is very close to the analytical delay limit for stability. However, the CQLKF leads to a complicated system of partial differential equations, yielding infinite dimensional linear matrix inequalities (LMIs). Therefore, to develop simpler stability criteria, many authors have used special forms of LKF rather than CQLKF, which give LMIs with finite order and a reduced value of MAUB. Further, to reduce the conservativeness of the existing results, some new analysis methods have been proposed, such as descriptor system transformation method [3-5], free weighting matrix Copyright c⃝ 2006-2012 by CCC Publications Stability of Discrete-Time Systems with Time-Varying Delay: Delay Decomposition Approach 777 method [8, 11, 23], matrix inequality method [10, 17, 18] and input–output approach [19]. Using these methods, many stability criteria were derived by checking a variation of LKF in a whole interval of the time-delay. Contrary to this approach, in [24, 25], in order to obtain some less conservative stability conditions, the interval of the time delay is divided into multiple equidistant subintervals and interval delay-dependent LKF (ID-D LKF) is constructed. By checking the variation of the ID-D LKF defined on the subintervals, some new delay-dependent stability criteria are derived. It is worth pointing out that the main difference between LKF and ID-D LKF lies in that the former allows taking different weighing matrices on different subintervals. Therefore, ID-D LKF, as expected, will yield less conservative delay-dependent stability criteria. Inspired by the idea of Zhu and Yang [26] on splitting the delay of continuous-time systems into two unequal subintervals, a new method is developed in this paper for stability analysis for discrete-time systems with time-varying delay. The delay interval [k −hM, k −1] in the ID-D LKF is divided into two unequal subintervals: [k −hM, k −α−1] and [k −α, k −1], where 0 < α < hM is a tuning parameter. The new ID-D LKF is constructed with different weighing matrices in various subintervals. Free-weighting matrices and model transformation are not used in order to derive delay dependent criterion. It is shown that the presented stability condition is much less conservative than the existing ones [1, 2, 6-9, 13-15, 22, 24], because it has a lower value of MAUB. The derived condition can be seen as an extension of the methods in [24, 25], wherein the whole delay range is divided to n ≥ 2 equal subintervals. As the number of subintervals in [24, 25] is greater than two, the decomposition approach is more complex and the resulting stability conditions are more conservative and difficult to implement. To demonstrate the effectiveness of the proposed method, numerical examples are given in section 3. Notation: ℜn and Z+ denote the n-dimensional Euclidean space and positive integers. Notation P > 0 (P ≥ 0) means that matrix P is real symmetric and positive definite (semi- definite). For real symmetric matrices P and Q, the notation P > Q (P ≥ Q) means that matrix P − Q is positive definite (positive semi-definite). I is an identity matrix with an appropriate dimension. Superscript “T” represents the transpose. In symmetric block matrices or complex matrix expressions, we use an asterisk (∗) to represent a term which is induced by symmetry. If dimensions of matrices are not explicitly given, then they are assumed to be compatible for algebraic operations. 2 Main results Consider the following system with an interval time-varying delay: x(k + 1) = Ax(k) + Bx(k −h(k)) (1) where x(k) ∈ℜn is the state at instant k, matrices A ∈ℜn×nand B ∈ℜn×n are constant matrices and h(k)is the positive integer representing the time delay of the system that we assume to be time dependent and satisfies the following: 0 ≤ h(k) ≤ hM (2) where hM is known to be a positive and finite integer. The aim of this article is to establish the sufficient condition that guarantee the delay- dependent stability of the system (1), which is less conservative than the existing results in literature. We first introduce the following result, which will be used in the proof of our main results. 778 S.B. Stojanovic, D.LJ. Debeljkovic, N. Dimitrijevic Lemma 1. Let y(k) = x(k + 1)−x(k). For any matrix R > 0 [24] −(hM −hm) k−1−hm∑ m=k−hM yT (m)Ry(m) ≤ [ x(k −hm) x(k −hM) ]T [ −R R R −R ] [ x(k −hm) x(k −hM) ] = − [x(k −hm)−x(k −hM)] T R [x(k −hm)−x(k −hM)] (3) Theorem 2. For given scalars hM(hM > 0) and α(0 < α < hM), the system described by (1)-(2) is asymptotically stable if there exists matrices P = PT > 0, Qi = QTi ≥ 0 and Zi = Z T i ≥ 0 (i = 1,2,3), such that the following LMIs hold: Φ =   Φ11 Φ12 0 0 Φ15 ∗ Φ22 Φ23 0 Φ25 ∗ ∗ Φ33 Φ34 0 ∗ ∗ ∗ Φ44 0 ∗ ∗ ∗ ∗ Φ55   < 0 (4) Ψ =   Ψ11 Ψ12 Ψ13 0 Ψ15 ∗ Ψ22 Ψ23 Ψ24 Ψ25 ∗ ∗ Ψ33 0 0 ∗ ∗ ∗ Ψ44 0 ∗ ∗ ∗ ∗ Ψ55   < 0 (5) where Φ11 = A T PA−P + Q1 + Q3 − 1α (Z1 + Z3) , Φ12 = A T PB + 1 α (Z1 + Z3) , Φ15 = (A− I)T U1, Φ22 = B T PB −Q3 − 1α (2Z1 + Z3) , Φ23 = 1 α Z1, Φ25 = B T U1, Φ33 = −Q1 + Q2 − 1αZ1 − 1 hM −α Z2, Φ34 = 1 hM −α Z2, Φ44 = −Q2 − 1hM −αZ2, Φ55 = −U1, Ψ11 = Φ11, Ψ12 = A T PB, Ψ13 = 1 α (Z1 + Z3) , Ψ15 = (A− I)T U2, Ψ22 = B T PB −Q3 − 1hM −α (2Z2 + Z3) , Ψ23 = 1 hM −α (Z2 + Z3) , Ψ24 = 1 hM −α Z2, Ψ25 = B T U2, Ψ33 = −Q1 + Q2 − 1α(Z1 + Z3)− 1 hM −α (Z2 + Z3) , Ψ44 = −Q2 − 1hM −αZ2, Ψ55 = −U2, U1 = αZ1 + (hM −α)Z2 + αZ3, U2 = αZ1 + (hM −α)Z2 + hMZ3 Proof: Construct the interval delay-dependent LKF as V (k) = V1(k) + V2(k) + V3(k) (6) where V1(k) = x T (k)Px(k) (7) V2(k) = k−1∑ i=k−α xT (i)Q1x(i) + k−1−α∑ i=k−hM xT (i)Q2x(i) + k−1∑ i=k−h(k) xT (i)Q3x(i) (8) V3(k) = −1∑ i=−α k−1∑ j=k+i yT (j)Z1y(j) + −1−α∑ i=−hM k−1∑ j=k+i yT (j)Z2y(j) + −1∑ i=−h(k) k−1∑ j=k+i yT (j)Z3y(j) (9) Stability of Discrete-Time Systems with Time-Varying Delay: Delay Decomposition Approach 779 where P = PT > 0, Qi = QTi ≥ 0 and Zi = Z T i > 0 (i = 1,2,3). Note, the delay interval [k −hM, k −1] in the LKF is divided into two unequal subintervals: [k −hM, k −α−1] and [k −α, k −1], where 0 < α < hM is a tuning parameter. Taking the difference of ∆Vi(k) = Vi(k + 1)−Vi(k), we can obtain ∆V1(k) = x T (k) ( AT PA−P ) x(k) + 2xT (k)AT PBx(k −h(k)) +xT (k −h(k))BT PBx(k −h(k)) (10) ∆V2(k) = x T (k)Q1x(k)−xT (k −α)Q1x(k −α) +xT (k −α)Q2x(k −α)−xT (k −hM)Q2x(k −hM) +xT (k)Q3x(k)−xT (k −h(k))Q3x(k −h(k)) (11) ∆V3(k) = −1∑ i=−α [ yT (k)Z1y(k)−yT (k + i)Z1y(k + i) ] + −1−α∑ i=−hM [ yT (k)Z2y(k)−yT (k + i)Z2y(k + i) ] + −1∑ i=−h(k) [ yT (k)Z3y(k)−yT (k + i)Z3y(k + i) ] = αyT (k)Z1y(k)− −1∑ i=−α yT (k + i)Z1y(k + i) + (hM −α)yT (k)Z2y(k)− −1−α∑ i=−hM yT (k + i)Z2y(k + i) +h(k)yT (k)Z3y(k)− −1∑ i=−h(k) yT (k + i)Z3y(k + i) = yT (k) [αZ1 + (hM −α)Z2 + h(k)Z3]y(k) − k−1∑ m=k−α yT (m)Z1y(m)− k−1−α∑ m=k−hM yT (m)Z2y(m) − k−1∑ m=k−h(k) yT (m)Z3y(m) = yT (k) [αZ1 + (hM −α)Z2 + h(k)Z3]y(k) − k−1∑ m=k−α yT (m)Z1y(m)− k−1−α∑ m=k−hM yT (m)Z2y(m) − k−1∑ m=k−h(k) yT (m)Z3y(m) (12) It is know from (2) that, for any k ∈ Z+, h(k) ∈ [0,α−1] or h(k) ∈ [α,hM]. Define two sets Ω1 = { k : h(k) ∈ [0,α] , k ∈ Z+ } (13) Ω2 = { k : h(k) ∈ [α + 1,hM] , k ∈ Z+ } (14) In the following, we will discuss the variation of ∆V (k) for two cases (k ∈ Ω1 and k ∈ Ω2). Case 1. For k ∈ Ω1, i.e. 0 ≤ h(k) ≤ α. k−1∑ m=k−α yT (m)Z1y(m) = k−1−h(k)∑ m=k−α yT (m)Z1y(m) + k−1∑ m=k−h(k) yT (m)Z1y(m) (15) 780 S.B. Stojanovic, D.LJ. Debeljkovic, N. Dimitrijevic ∆V3(k) = y T (k) [αZ1 + (hM −α)Z2 + h(k)Z3]y(k)− k−1−h(k)∑ m=k−α yT (m)Z1y(m) − k−1∑ m=k−h(k) yT (m) (Z1 + Z3)y(m)− k−1−α∑ m=k−hM yT (m)Z2y(m) (16) Because Z1 + Z3 > 0, h(k) ≤ α and α−h(k) ≤ α, using Lemma 1, it follows − k−1∑ m=k−h(k) yT (m) (Z1 + Z3)y(m) ≤− 1 h(k) [x(k)−x(k −h(k))]T (Z1 + Z3) [x(k)−x(k −h(k))] ≤ 1 α xT (k) (−Z1 −Z3)x(k) + 1α2x T (k) (Z1 + Z3)x(k −h(k)) + 1 α xT (k −h(k)) (−Z1 −Z3)x(k −h(k)) (17) − k−1−h(k)∑ m=k−α yT (m)Z1y(m) ≤− 1 α−h(k) [x(k −h(k))−x(k −α)] T Z1 [x(k −h(k))−x(k −α)] ≤ 1 α xT (k −h(k)) (−Z1)x(k −h(k)) + 1α2x T (k −h(k))Z1x(k −α) + 1 α xT (k −α) (−Z1)x(k −α) (18) − k−1−α∑ m=k−hM yT (m)Z2y(m) ≤− 1 hM −α [x(k −α)−x(k −hM)] T Z2 [x(k −α)−x(k −hM)] ≤ 1 hM −α xT (k −α) (−Z2)x(k −α) + 1hM −α2x T (k −α)Z2x(k −hM) + 1 hM −α xT (k −hM) (−Z2)x(k −hM) (19) Combining (10)-(19), it yields ∆V (k) ≤ ξT (k)Φ̂ξ(k) Φ̂ =   Φ11 + (A− I)T U1(A− I) Φ12 + (A− I)T U1B 0 0 ∗ Φ22 + BT U1B Φ23 0 ∗ ∗ Φ33 Φ34 ∗ ∗ ∗ Φ44   ξ(k) = [ xT (k) xT (k −h(k)) xT (k −α) xT (k −hM) ]T (20) Obviously, ∆V (k) < 0 for k ∈ Ω1 if Φ̂ < 0. Using the Schur complement, it is easy to see that ∆V (k) < 0 holds if Φ < 0 and h(k) ∈ [0,α]. Case 2. Similarly, for k ∈ Ω2, i.e. α + 1 ≤ h(k) ≤ hM , using the Schur complement, it is easy to see that ∆V (k) < 0 holds if Ψ < 0. From the above discussions, we can see that for all k ∈ Z+ if (4)-(5) hold, ∆V (k) < 0, which completes the proof. 2 Remark 3. Theorem 2 presents a stability result which depends on the maximum delay bound hM . The conditions in Theorem 2 are expressed in terms of LMIs, and therefore, they can be easily checked by using standard numerical software. Remark 4. The delay interval [k −hM, k −1] in the ID-D LFK is divided into two unequal subintervals: [k −hM, k −α−1] and [k −α, k −1], where 0 < α < hM is a tuning parameter. Stability of Discrete-Time Systems with Time-Varying Delay: Delay Decomposition Approach 781 Consequently, the different weighing matrices in the Lyapunov functional are used in various subintervals and the information of delayed state x(k −α) can be taken into full consideration. Further, using the subintervals and Lemma 1, the upper bounds of some terms in ∆V3(k) are more accurately estimated than by using the previous methods since the upper bound hM of delay h(k) on the interval 0 ≤ h(k) ≤ hM is substituted with two less conservative upper bounds α and hM on the subintervals 0 ≤ h(k) ≤ α and α < h(k) ≤ hM , respectively. So the decomposition method presented in Theorem 2 can reduce the value of MAUB. An algorithm for seeking a corresponding values of α (0 < α < hM) subject to (4)-(5), such that the MAUB of hM has maximal value, can easily be obtained. Algorithm 5. Step 1. Let h = 0 and α = 0. Step 2. h = h + 1. Step 3. α = α + 1. Step 4. If inequalities (4)-(5) is feasible, then αm = α, α = 0 and go to step 2; otherwise, go to step 5. Step 5. If α = h−1, go to step 6; otherwise, go to step 3. Step 6. The maximal delay is hM = h−1 and the minimal value of tuning parameterαis αm. 3 Numerical examples In this section, two examples are presented. The obtained results have been compared with several existing criteria in the literature. Example 1. Consider system (1) with the time-varying delay h(k) satisfying (2) and A = [ 0.8 0 0 λ ] , B = [ −0.1 0 −0.1 −0.1 ] , λ ∈{0.91, 0.97} Case 1 (λ = 0.91). This system was considered in [13] and [22]. Table 1 lists the MAUB of delay obtained from Theorem 2 of this paper. For comparison, results from [13, 22] are also listed in the table. It is clear that Theorem 2 leads to better results than those in [13, 22]. Table 1. Comparison of MAUB of delay based on different ex- isting methods for λ = 0.91 Method hM [13] 41 [22, Corollary 1] 42 Theorem 2 in this paper 46 for α = 19, · · · ,30 Case 2 (λ = 0.97). For comparison, the results from [2, 6, 7, 9] and this paper are listed in Table 2. It is clear that Theorem 2 gives much better results than the existing delay-dependent criteria. Example 2. Consider system (1) with the time-varying delay h(k) satisfying (2) and Case 1. [1, 14, 15] A = [ 0.6 0 0.35 0.7 ] , B = [ 0.1 0 0.2 0.1 ] Case 2. [8, 9, 15, 24] A = [ 0.8 0 0.05 0.9 ] , B = [ −0.1 0 −0.2 −0.1 ] 782 S.B. Stojanovic, D.LJ. Debeljkovic, N. Dimitrijevic Table 2. Comparison of MAUB of delay based on different ex- isting methods for λ = 0.97 Method hM [9, Theorem 1] 4 [7, Theorem 3] 8 [6, Lemma 2] 8 [2, Theorem 1] 10 Theorem 2 in this paper 17 for α = 9,10,11 For the above systems, the LMIs conditions for delay-independent stability [16] P = PT > 0, Q = QT > 0,   −P + Q 0 AT P ∗ −Q BT P ∗ ∗ −P   < 0 are feasible, from which we deduce that both systems are stable for 0 ≤ h(k) < ∞. Using the existing delay-dependent criteria, it can be obtained only the finite value of MAUB, which will guarantee the stability of the given systems. Tables 3 (Case 1) and 4 (Case 2) list the intervals of time delay for different methods. Based on Theorem 2 in this paper, large numerical values of MAUB are obtained (hM → ∞). Hence, using of Theorem 2 leads to better results than those in [1, 8, 9, 14, 15, 25]. Table 3. Interval of time delay for Case 1 Method Interval [1, Th. 3.1] 2 ≤ h(k) ≤ 10 [15, Theorem 1], [15, Theorem 2] 2 ≤ h(k) ≤ 13 [14, Theorem 3.2] 0 ≤ h(k) ≤ 12 [2, Theorem 1] 2 ≤ h(k) ≤ 15 Theorem 2 in this paper 0 ≤ h(k) ≤ 10 ·1021 Table 4. Interval of time delay for Case 2 Method Interval stability [9, Theorem 1] 0 ≤ h(k) ≤ 6 [15, Theorem 2] 0 ≤ h(k) ≤ 10 [8, Theorem 1] 0 ≤ h(k) ≤ 12 [24, Theorem 5] 2 ≤ h(k) ≤ 19 [24, Theorem 7] 2 ≤ h(k) ≤ 20 Theorem 2 in this paper 0 ≤ h(k) ≤ 9.61 ·108 Stability of Discrete-Time Systems with Time-Varying Delay: Delay Decomposition Approach 783 4 Conclusion In this paper, the problem of obtaining delay dependent stability conditions for a class of systems with interval time-varying delay is discussed. A new interval delay-dependent Lyapunov– Krasovskii functional is constructed by splitting the delay interval into two unequal intervals by tuning parameter α. The free-weighting matrices and model transformation are not used in order to derive delay-dependent criteria. Numerical examples show that the results proposed in this paper are much less conservative while comparing the maximum allowable upper bound of delay with the existing results in the literature. Bibliography [1] E.K. Boukas, Discrete-time systems with time-varying time delay: stability and stabilizability, Mathematical Problems in Engineering, Article ID 42489:1-10, 2006. [2] K.F. Chen and I-K Fong, Stability of discrete-time uncertain systems with a time-varying state delay, Proc. IMechE, Part I: J. Systems and Control Engineering, 222: 493-500, 2008. [3] E. Fridman and U. Shaked, A descriptor system approach to H∞ control of linear time-delay systems, IEEE Transactions on Automatic Control, 47(2): 253–270, 2002. [4] E. Fridman and U. Shaked, H∞ control of linear state delay descriptor systems: an LMI approach, Linear Algebra and its Applications, 351–352: 271–302, 2002. [5] E. Fridman, New Lyapunov–Krasovskii functionals for stability of linear retarded and neutral type systems, Systems and Control Letters, 43: 309–319, 2001. [6] E. Fridman and U. Shaked, Delay-Dependent H∞ Control of Uncertain Discrete Delay Sys- tems, European Journal of Control, 11: 29-37, 2005. [7] E. Fridman and U. Shaked, Stability and guaranteed cost control of uncertain discrete delay systems, International Journal of Control, 78(4): 235-246, 2005. [8] H. Gao and T. Chen, New results on stability of discrete-time systems with time-varying state delay, IEEE Transactions on Automatic Control, 52: 328–334, 2007. [9] H. Gao, J. Lam and Y. Wang, Delay-dependent output-feedback stabilization of discrete- time systems with time-varying state delay, IEE Proc.: Control Theory Applications, 151(6): 691-698, 2004. [10] Q.L. Han and D. Yue, Absolute stability of Lur’e systems with time-varying delay, IET Control Theory, 1(3): 854–859, 2007. [11] Y. He, M. Wu, J.H. She and G.P. Liu, Parameter-dependent Lyapunov functional for stabil- ity of time-delay systems with polytopic-type uncertainties, IEEE Transactions on Automatic Control, 49: 828–832, 2004. [12] X. Jiang, Q.L. Han and X.H. Yu, Stability criteria for linear discrete-time systems with interval-like time-varying delay, Proc. American Control Conference, New Orleans, USA, 2817–2822, 2005. [13] Y.S. Lee and W.H. Kwon, Delay-dependent robust stabilization of uncertain discrete-time state-delayed systems, Proc. 15thIFAC World Congr., 15(1): Barcelona, Spain 2002. 784 S.B. Stojanovic, D.LJ. Debeljkovic, N. Dimitrijevic [14] V. Leite and M. Miranda, Robust Stabilization of Discrete-Time Systems with Time-Varying Delay: An LMI Approach, Mathematical Problems in Engineering, 2008: Article ID 876509, 15 pages, 2008. [15] X.G. Liu, R.R. Martin, M. Wu and M.L. Tang, Delay-dependent robust stabilization of discrete-time systems with time-varying delay, IEE Proc.: Control Theory and Applications, 153(6): 689–702, 2006. [16] M.S. Mahmood, Robust control and filtering for time-delay systems, Marcel-Dekker, New York, 2000. [17] Y.S. Moon, P. Park and W.H. Kwon, Robust stabilization of uncertain input-delayed systems using reduction method, Automatica, 37: 307–312, 2001. [18] P. Park and J.W. Ko Stability and robust stability for systems with a time-varying delay, Automatica, 43: 1855–1858, 2007. [19] E. Shustin and E. Fridman, On delay-derivative-dependent stability of systems with fast- varying delays, Automatica, 43: 1649–1655, 2007. [20] M. Wu, Y. He, J.H. She and G.P. Liu, Delay-dependent criteria for robust stability of time-varying delay systems, Automatica, 40: 1435–1439, 2004. [21] S. Xu and J. Lam, Improved Delay-dependent Stability Criteria for time-delay systems, IEEE Transactions on Automatic Control, 50(3): 384-387, 2005. [22] S. Xu, J. Lam and Y. Zou, Improved conditions for delay-dependent robust stability and stabilization of uncertain discrete time-delay systems, Asian Journal of Control, 7(3): 344- 348, 2005. [23] D. Yue and Q.L. Han, A delay-dependent stability criterion of neutral systems and its application to a partial element equivalent circuit model, IEEE Transactions on Circuits and Systems-II, 51(12): 685–689, 2004. [24] D. Yue, E. Tian and Y. Zhang, A piecewise analysis method to stability analysis of linear continuous/discrete systems with time-varying delay, International Journal of Robust and Nonlinear Control, 19: 1493–1518, 2009. [25] X.M. Zhang and Q.L. Han, A delay decomposition to delay-dependent stability for linear systems with time-varying delays, International Journal of robust and nonlinear control, 19: 1922–1930, 2009. [26] X.L. Zhu and G.H. Yang, New results of stability analysis for systems with time-varying delay, International Journal of robust and nonlinear control 20: 596–606, 2010.