You are on page 1of 7

Applied Mathematics Letters 20 (2007) 839–845 www.elsevier.

com/locate/aml

Impulsive stabilization of delay differential systems via the Lyapunov–Razumikhin method$
Qing Wang, Xinzhi Liu ∗
Department of Applied Mathematics, University of Waterloo, Waterloo, ON, Canada, N2L 3G1 Received 20 April 2006; accepted 14 August 2006

Abstract This work studies global exponential stability of impulsive delay differential systems. By employing the Razumikhin technique and Lyapunov functions, several global exponential stability criteria are established for general impulsive delay differential equations. Our results show that delay differential equations may be exponentially stabilized by impulses. An example and its simulation are also given to illustrate our results. c 2006 Elsevier Ltd. All rights reserved.
Keywords: Razumikhin technique; Lyapunov function; Impulsive delay differential equations; Global exponential stability

1. Introduction Impulsive delay differential equations arise in many applied fields such as control technology, communication networks, and biological population management and hence they have attracted considerable attention. See [1–4,6, 9–13] and the references therein. In recent years, stability of differential equations has been extensively studied. One of the most investigated problems in the stability analysis of such systems is exponential stability since it has played an important role in many areas such as designs and applications of neural networks and synchronization in secure communication [5,7,8]. On the other hand, various methods, such as LMI tools, Laplace transform, and Lyapunov functional or function methods (combined with the Razumikhin technique) and so on, have been successfully utilized in the investigation of exponential stability; see [1,4–6,12] for example. And the well-known Razumikhin technique has been successfully applied in the study of asymptotic and exponential stability of impulsive delay differential equations; see [9–13] and relevant references cited therein. However, to the best of our knowledge, there have been few results obtained for impulsive exponential stabilization of delay differential equations [8,13]. The aim of this work is to establish global exponential stability criteria for impulsive delay systems by employing the Razumikhin technique which illustrate that impulses do contribute to the stabilization of some delay differential systems.
$ Research supported by NSERC-Canada.
∗ Corresponding author.

E-mail address: xzliu@uwaterloo.ca (X. Liu). 0893-9659/$ - see front matter c 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.aml.2006.08.016

where q ≥ e2λα is a constant. − lim(t. whenever q V (t.1) such that xt0 = φ. Preliminaries Given a constant τ > 0. − (iii) V (tk . at which x(t) is right continuous. k ∈ N . satisfy all necessary conditions for the global existence and uniqueness of solutions for all t ≥ t0 [2]. ϕ(0)).1) is defined by 1 D + V (t. c2 > 0 and α > τ . x) exists. k ∈ N . ϕ(0)). tk ). ϕ(0)) ≥ V (t + s. t ∈ [tk−1 . Assume that there exist a function V ∈ ν0 and constants p. k ∈ N . Definition 2. ϕ)) ≤ dk V (tk . c1 . x(tk ) = x(tk ). 0].3. x(tk ) = Ik (tk . xt − ∈ PC([−τ. k ∈ N . In this work. · τ defined by ψ τ = (2.1) based on the Lyapunov–Razumikhin method. h→0+ h for (t. V (t. φ ∈ PC([−τ. for all t ∈ [tk−1 . Liu / Applied Mathematics Letters 20 (2007) 839–845 2. + i. Denote by x(t) = x(t.y)→(t − . Function V : R+ × Rn → R+ is said to belong to the class ν0 if (i) V is continuous in each of the sets [tk−1 . where (t0 . where dk > 0. 0) ≡ 0. for any t ∈ R+ and x ∈ Rn . 0]. Rn ) are defined by xt (s) = x(t +s). t0 .2. and k (ii) V (t.1) is globally exponentially stable and the convergence rate is . (iv) τ ≤ tk − tk−1 ≤ α and ln(dk ) + λα < −λ(tk+1 − tk ). Then the trivial solution of the impulsive system (2. 0 ≤ t0 < t1 < t2 < · · · < tk < · · ·. Rn ). λ p. x) is locally Lipschitzian in all x ∈ Rn . 0]. k ∈ N . ψ(0)) = lim sup [V (t + h. we shall present some Razumikhin-type theorems on global exponential stability for system (2. ϕ(0) + Ik (tk . c. ϕ(s)) for s ∈ [−τ. y) = V (tk . respectively. Theorem 3. are constants. ψ) ∈ R+ × PC([−τ.1.x) V (t. Ik : R+ × PC([−τ. xt ).1. we equip the linear space PC([−τ. ∀k ∈ N . φ) ∈ R+ × PC([−τ. λ > c such that (i) c1 x p ≤ V (t. We further assume that all the solutions x(t) of (2. Rn ) with the norm sup−τ ≤s≤0 ψ(s) . 3. we assume that functions F.e. tk ). ψ(0))]. and for all t ≥ t0 . Consider the following impulsive system: ⎧ ⎨x (t) = F(t. 0]. (ii) D + V (t. ϕ(0)) ≤ cV (t. φ) ≤ M φ −α(t−t0 ) . ψ(0) + h F(t. Rn ). The Lyapunov–Razumikhin method In this section. the upper right-hand derivative of V with respect to system (2. with tk → ∞ as k → ∞. 0]. t = tk . 0]. Definition 2. xt − (s) = x(t − +s) for −τ ≤ s ≤ 0. Rn ) → Rn . Wang. Definition 2.840 Q. Given a function V : R+ × Rn → R+ . k ∈ N . Ik . k ⎩ xt0 = φ.1) where F. and xt . xt − ). 0]. x(t) = x(t)−x(t − ).. tk ) × Rn and for each x ∈ Rn . t0 . Our results show that impulses play an important role in stabilizing delay differential systems. τe t ≥ t0 . φ) the solution of (2.1) are continuous except at tk . ψ)) − V (t.1) is said to be globally exponentially stable if there exist some constants α > 0 and M ≥ 1 such that for any initial data xt0 = φ x(t. X. x) ≤ c2 x p . Rn ). The trivial solution of system (2.

11) From (3. .3) From condition (i) and (3. p p c2 φ τ ≥ M φ τ e−λ(t1 −t0 ) e−αc . τe p −λ(tm −t0 ) τe (3.1) with xt0 = φ. p −λ(t1 −t0 ) −αc e τe s ∈ [−τ. We shall show v(t) ≤ M φ We first show that v(t) ≤ M φ v(t) ≤ c2 x ¯ v(t ) > M φ > c2 φ v(t ∗ ) = M φ and there exists t ∗∗ v(t ∗∗ ) = c2 φ p −λ(t1 −t0 ) . m. x)..4) t0 − τ ≤ t ≤ t ∗ . τe t ∈ [tm . ¯ t ∈ [t ∗ . . .7) and thus by condition (ii). i.e. we shall show that (3. s ∈ [−τ. τe and v(t) ≤ M φ p τ. Wang. which contradicts (3. t0 ] p ≤ c2 φ p τ ¯ If (3. and then we have v(t ∗∗ ) ≥ v(t ∗ )e−αc . . we know t = tm . tk ). we have and v(t) ≤ M φ ¯ < v(t ).8). tm+1 ). (3. Hence (3. X. m ≥ 1). 0]. t ]. 2. t1 ). and v(t) = V (t. t ]. tm+1 )|v(t) > M φ ¯ v(t ) = M φ p −λ(tm+1 −t0 ) τe ¯ From (3.9). k ∈ N . Now we assume that (3. Liu / Applied Mathematics Letters 20 (2007) 839–845 841 Proof.9). 0]. By the continuity of v(t) in the interval [tm . p −λ(tm+1 −t0 ) }. suppose (3. (3.Q. . i. v(t) ≤ M φ p −λ(tk −t0 ) . (3. we have − v(tm ) ≤ dm v(tm ) < e−λα e−λ(tm+1 −tm ) M φ < M φ v(t) ≤ M φ p −λ(tm+1 −t0 ) . k = 1. then there must exist some t ∈ (t0 . we have.1) Let x(t) = x(t. tm+1 ).3) is not true.9) Next. m (m ∈ N .2) holds for k = 1. we get D + v(t) ≤ cv(t) for t ∈ [t ∗∗ . ¯ which implies that there exists some t ∗ ∈ (t0 . i. for t ∈ [t0 − τ. t ∗) such that t ∗∗ ≤ t ≤ t ∗ . ¯ and v(t ∗ ) ≤ v(t) ≤ v(t ).8) From condition (iii) and (3.2) (3. τe p −λ(tk −t0 ) . τe t ∈ [tk−1 . 2. t0 . (3. tk ). Choose M ≥ 1 such that c2 φ p τ <M φ p −λ(t1 −t0 ) −αc e τe <M φ p −λ(t1 −t0 ) τe ≤ qc2 φ p τ.e. t ) such that p −λ(t1 −t0 ) . . Then we define ¯ t = inf{t ∈ [tm .5) ∈ [t0 p τ.10) is not true.2) is true for k = 1.10) For the sake of contradiction.6) and v(t) ≥ c2 φ [t ∗∗ . τe (3. we have v(tm ) < e−λα M φ v(t ∗ ) = e−λα M φ p −λ(tm+1 −t0 ) τe ¯ which implies that there exists some t ∗ ∈ (tm . φ) be any solution of system (2. p −λ(tm+1 −t0 ) . τe ¯ t ∈ [tm . t ∗ ] ≤ qc2 φ p −λ(t1 −t0 ) τe Then we obtain. τe (3. t1 ) such that p −λ(t1 −t0 ) >M τe p τ ≥ v(t0 + s). . . t ∗ ]. for any t ∈ v(t + s) ≤ M φ p τ ≤ qv(t). <M φ φ p −λ(t1 −t0 ) −αc e . t ∈ [t0 .e.12) .1). p −λ(tm+1 −t0 ) .3) holds and then (3.2) holds for k = m + 1. (3. . τe t ∈ [tk−1 . t ) such that p −λ(tm+1 −t0 ) τe (3.1). τe (3. p −λ(t1 −t0 ) . (3.

This implies the assumption is not true. However.1). as we can see from Theorems 3. for any t ∈ R+ and x ∈ Rn .12) p −λ(tm+1 −t0 ) αc e τe which is a contradiction. λmax (A) + q 2 B < λ . 0]. The details are stated in the following result whose proof is similar and thus omitted. e2λα } is a constant. and hence we have v(t) ≤ M φ p −λ(t−t0 ) . − (iii) V (tk . Liu / Applied Mathematics Letters 20 (2007) 839–845 ¯ ¯ Then we know t + s ∈ [tm−1 . ¯ for t ∈ [t ∗ . the condition D + V (t. (ii) D + V (t.2. x) allows the derivative of the Lyapunov function to be positive which may not even guarantee the stability of a delay differential system (see [9. ϕ(0)). c. x) ≤ cV (t. we get x ≤ M∗ φ − λ (t−t0 ) p . ¯ v(t ) ≤ v(t ∗ )eαc = e−λα M φ s ∈ [−τ. X. t ∈ [tk−1 . where t − τ (t) is strictly increasing on R+ and 0 ≤ τ (t) ≤ τ . k ∈ N . Then the trivial solution of the impulsive system (2. then we have to require q ≥ max{eαc . t ] for t ∈ [t ∗ . (iv) τ ≤ tk − tk−1 ≤ α and ln(dk ) + (λ + c)α < −λ(tk+1 − tk ). Then from condition (ii).1 and 3. 0] since τ ≤ tk − tk−1 ≤ α. ϕ(s)) for s ∈ [−τ.1. Corollary 3. tk ). ϕ(0)) ≤ cV (t. λ > 0 and α > τ such that (i) c1 x p ≤ V (t. ϕ(0)) ≥ V (t + s. Remark 3.2. t ]. Then by condition (i).13) . Theorem 3. we get D + v(t) ≤ cv(t). λ p. ϕ(0)). Thus by mathematical induction. c2 . ϕ(0) + Ik (tk .2.1) is globally exponentially stable with λ convergence rate p . this implies that the trivial solution of system (2. in the stability theory of delay differential equations. k ∈ N .1. from (3. τe t ∈ [tk−1 .8) and (3. x) ≤ c2 x p . we obtain that (3. we shall apply the previous theorems to the following linear impulsive delay system: ⎧ ˙ ⎨x(t) = Ax(t) + Bx(t − τ (t)). It is well known that. we get. we have. ¯ < v(t ). Assume that there exist a function V ∈ ν0 and constants p. where q ≥ max{eαc . tk ).2) holds.1. Suppose there exist some constants α. ⎩ xt0 = φ. are constants. e2λα } in condition (ii) and condition (iv) to be strengthened. By (3. and hence (3. 2 1 (3. k ∈ N . Wang. τe 1 t ∈ [tk−1 . for all t ∈ [tk−1 . ∀k ∈ N . If the condition λ > c is removed in Theorem 3. [ c1 ] p }. v(t + s) ≤ M φ = M φ ≤ eλα M p −λ(tm −t0 ) τe p −λ(tm+1 −t0 ) λ(tm+1 −tm ) e τe p −λ(tm+1 −t0 ) φ τe ∗ = e2λα v(t ) ≤ qv(t). M where M ∗ ≥ max{1.13] and Example 4. t = tk . c1 . since λ > c. whenever q V (t. Next.1) is globally exponentially stable and the convergence rate is Remark 3.2) holds for k = m + 1. where dk > 0. tk ). x(t) = Ck x(t − ). λ > 0 such that (i) for some constant q ≥ e2λα . t ] and s ∈ [−τ. 0].842 Q.11). impulses have played an important role in exponentially stabilizing a delay differential system. tk ). ϕ)) ≤ dk V (tk .

1 0.1 ⎤ 0 0 ⎦.12 0 ⎤ 0.24 and ⎡ ⎡ −0.14) Then system (3. e2λα } = 1. λmax (A) + q 2 B = 0.15) 1 Then system (3. An example In this section. 2 2 1 . ⎪ ⎩ xt0 = φ. −0.05 = τ ≤ tk − tk−1 ≤ α = 0. Example 4.17.05 ⎦ 0. B = [λmax (B B T )] 2 = 0.6808 < − λ (tk+1 − tk ) = −0. 0.2 0.05 = τ ≤ tk − tk−1 ≤ α = 0. we give an example and its simulation to illustrate our results.2 A = ⎣0. 2 λ 2 = 0. Choosing q = 2.2.2 0.4388. where ⎡ ⎤ −0.2905 and I + Ck = 0.85.8496 < (ii) 0. ln I + Ck + λα 2 1 = −0.6992.05. 2 Proof.13) is globally exponentially stable and its convergence rate is λ . τ = 0.9739.Q. Thus by Corollary 3. (ii) τ ≤ tk − tk−1 ≤ α and ln I + Ck + λ α (λ + c) < − (tk+1 − tk ). we can also find that the conditions of Corollary 3.2.15 0 0.5 0 −0.6.2 hold: (i) c = 2(λmax (A) + q 2 B ) = 1. It follows from Theorem 3. where c = 2(λmax (A) + q 2 B ). (4.12 B = ⎣ 0.1) is globally exponentially stable with convergence rate 0.1. It follows from Theorem 3.85. 2 2 (3. k ∈ N . Consider the following linear impulsive delay system: ⎧ ⎪ ⎪x(t) = Ax(t) + Bx t − 1 (1 + e−t ) .17.1 on choosing V (x) = x 2. q = 2 ≥ max{ecα . λ = 1. e2λα }.1. λ > 0 such that (i) there exists some constant q > 0 such that q ≥ max{ecα . we know that the trivial solution of (4. Wang.1 hold: (i) q = 2 ≥ e2λα = 1.2.3 ⎦ .8 Ck = ⎣ 0 0 0 Then λmax (A) = 0. we find that the conditions of Corollary 3. t = k ⎨˙ 40 ⎪ x(t) = Ck x(t − ).14 −0. ln I + Ck + (λ+c)α = −0.4 1 2.7. Liu / Applied Mathematics Letters 20 (2007) 839–845 843 (ii) τ ≤ tk − tk−1 ≤ α and ln I + Ck + λα λ < − (tk+1 − tk ).9739. 2 Proof.2.03 0 −0.13) is globally exponentially stable and its convergence rate is λ . Suppose there exist some constants α. α = 0. 2 2 (3.1 −0. Corollary 3. t = k. Furthermore.1 0.1709 < − λ (tk+1 − tk ) = −0.2 on choosing V (x) = x 4. (ii) 0.1) 0. X.

As we see from the above pictures. Liu. 2. E. Fig. Exponential stability of linear delay impulsive differential equations. the graph of the solution of the corresponding system without impulses is given in Fig. Berezansky.844 Q. X.1H (t). [2] G. J.85. the trivial solution of system (4. References [1] A. is given in Fig. however. L. .1. 1.2 that the trivial solution of (4.1) is globally exponentially stable with convergence rate 0. This implies that impulses may be used to exponentially stabilize some delay differential systems. 2. 5 (1999) 579–591. Dyn. Existence and uniqueness results for impulsive delay differential equations. Syst. Discrete Impuls. Impulse-stabilized system. 2. Braverman. Ballinger. Anokhin. Appl. System without impulses. the trivial solution becomes globally exponentially stable. Then we know from Corollary 3. X. Wang. Liu / Applied Mathematics Letters 20 (2007) 839–845 Fig. 1. The numerical simulation of this impulsive delay differential equation with the initial function (3.1) without impulses is unstable. after impulsive control. where H (t) is the Heaviside step function.7H (t). Math. 193 (1995) 923–941.502H (t))T . −2. Anal. Remark 4. Contin.

Exponential stability of some scalar impulsive delay differential equation. IMA J. Math. Practical stability of impulsive delay differential equations and applications to control problems. 2001. Yan. .T. Commun. Appl. Appl. Math. 9 (1994) 163–174. Ballinger. J. Impulsive stabilization of functional differential equations with infinite delays.Z. Dyn. Nonlinear Anal. G. Liu. Syst.M.. J. Stamova. Stab. Math. Luo. in: Optimization Methods and Applications. L. Comput. [5] T. Wang. Shen. J. X. Liu.P. Math. Liu. Shen. Control Inform.Z. J. X. Math. 130 (2001) 163–171. Appl. Berezansky. Wang. Neural Networks 14 (2001) 977–980. Lett. [8] X. Impulsive stabilization of nonlinear systems. Kluwer Acad. Math. 41 (2001) 903–915. Razumikhin-type theorems on exponential stability of impulsive delay systems. Anal. 309 (2005) 462–473. Idels. K. [6] B. Publ. [12] Q. 10 (1993) 11–19. Appl. Exponential stability for impulsive delay differential equations by Razumikhin method.Z. Comput. Chen. [10] J. Stability results for impulsive differential systems with applications to population growth models.Q.Z. Razumikhin type stability theorems for impulsive functional differential equations. Math. Q. Appl. Uniform asymptotic stability of impulsive delay differential equations. 71 (2006) 47–61. [4] L. Teo. G. Wang. X. Liu. 33 (1998) 519–531. Lyapunov–Razumikhin method for impulsive functional equations and applications to the population dynamics.Z. 2 (1998) 301–309. IMA J. Liu.Z. [13] Z. Liu. Stamov. Appl. Anal. [11] I. 16 (2003) 695–701. Dordrecht. Ballinger. Liu. Liu / Applied Mathematics Letters 20 (2007) 839–845 845 [3] G. X. [9] X. [7] X. Global exponential stability of delayed Hopfield neural networks.