Applied Mathematics Letters 20 (2007) 839845 www.elsevier.

com/locate/aml

Impulsive stabilization of delay differential systems via the LyapunovRazumikhin method$

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 elds such as control technology, communication networks, and biological population management and hence they have attracted considerable attention. See [14,6, 913] 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 [1,46,12] 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 [913] 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: [email protected] (X. Liu). 0893-9659/$ - see front matter c 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.aml.2006.08.016

840

Q. Wang, X. Liu / Applied Mathematics Letters 20 (2007) 839845

2. Preliminaries Given a constant > 0, we equip the linear space PC([, 0], Rn ) with the norm sup s0 (s) . Consider the following impulsive system: x (t) = F(t, xt ), t = tk , x(tk ) = Ik (tk , xt ), k N , k xt0 = ,

dened by

(2.1)

where F, Ik : R+ PC([, 0], Rn ) Rn ; PC([, 0], Rn ); 0 t0 < t1 < t2 < < tk < , with tk as k ; x(t) = x(t)x(t ); and xt , xt PC([, 0], Rn ) are dened by xt (s) = x(t +s), xt (s) = x(t +s) for s 0, respectively. In this work, we assume that functions F, Ik , k N , satisfy all necessary conditions for the global existence and uniqueness of solutions for all t t0 [2]. Denote by x(t) = x(t, t0 , ) the solution of (2.1) such that xt0 = . We further assume that all the solutions x(t) of (2.1) are continuous except at tk , k N , at which x(t) is right continuous, + i.e., x(tk ) = x(tk ), k N . Denition 2.1. Function V : R+ Rn R+ is said to belong to the class 0 if (i) V is continuous in each of the sets [tk1 , tk ) Rn and for each x Rn , t [tk1 , tk ), k N , lim(t,y)(t ,x) V (t, y) = V (tk , x) exists; and k (ii) V (t, x) is locally Lipschitzian in all x Rn , and for all t t0 , V (t, 0) 0. Denition 2.2. Given a function V : R+ Rn R+ , the upper right-hand derivative of V with respect to system (2.1) is dened by 1 D + V (t, (0)) = lim sup [V (t + h, (0) + h F(t, )) V (t, (0))], h0+ h for (t, ) R+ PC([, 0], Rn ). Denition 2.3. The trivial solution of system (2.1) is said to be globally exponentially stable if there exist some constants > 0 and M 1 such that for any initial data xt0 = x(t, t0 , ) M
(tt0 ) , e

t t0 ,

where (t0 , ) R+ PC([, 0], Rn ). 3. The LyapunovRazumikhin method In this section, we shall present some Razumikhin-type theorems on global exponential stability for system (2.1) based on the LyapunovRazumikhin method. Our results show that impulses play an important role in stabilizing delay differential systems. Theorem 3.1. Assume that there exist a function V 0 and constants p, c, c1 , c2 > 0 and > , > c such that (i) c1 x p V (t, x) c2 x p , for any t R+ and x Rn ; (ii) D + V (t, (0)) cV (t, (0)), for all t [tk1 , tk ), k N , whenever q V (t, (0)) V (t + s, (s)) for s [, 0], where q e2 is a constant; (iii) V (tk , (0) + Ik (tk , )) dk V (tk , (0)), where dk > 0, k N , are constants; (iv) tk tk1 and ln(dk ) + < (tk+1 tk ).
p.

Then the trivial solution of the impulsive system (2.1) is globally exponentially stable and the convergence rate is

Q. Wang, X. Liu / Applied Mathematics Letters 20 (2007) 839845

841

Proof. Choose M 1 such that c2

<M

p (t1 t0 ) c e e

<M

p (t1 t0 ) e

qc2

p .

(3.1)

Let x(t) = x(t, t0 , ) be any solution of system (2.1) with xt0 = , and v(t) = V (t, x). We shall show v(t) M We rst show that v(t) M v(t) c2 x v(t ) > M > c2 v(t ) = M and there exists t v(t ) = c2
p (t1 t0 ) , e p (tk t0 ) , e

t [tk1 , tk ), k N . t [t0 , t1 ). <M

(3.2)

(3.3)

From condition (i) and (3.1), we have, for t [t0 , t0 ]

c2

If (3.3) is not true, then there must exist some t (t0 , t1 ) such that
p (t1 t0 ) >M e p v(t0 + s), p (t1 t0 ) c e e

s [, 0],
p (t1 t0 ) , e

(3.4) t0 t t ;

which implies that there exists some t (t0 , t ) such that

and v(t) M
p ,

(3.5)

[t0
p ,

, t )

such that t t t . (3.6)

and v(t) c2 [t , t ] qc2

Then we obtain, for any t v(t + s) M

qv(t),

s [, 0],

(3.7)

and thus by condition (ii), we get D + v(t) cv(t) for t [t , t ], and then we have v(t ) v(t )ec , i.e., p p c2 M e(t1 t0 ) ec , which contradicts (3.1). Hence (3.3) holds and then (3.2) is true for k = 1. Now we assume that (3.2) holds for k = 1, 2, . . . , m (m N , m 1), i.e. v(t) M
p (tk t0 ) , e

t [tk1 , tk ), k = 1, 2, . . . , m.

(3.8)

From condition (iii) and (3.8), we have

< e e(tm+1 tm ) M < M v(t) M

p (tm t0 ) e

(3.9)

Next, we shall show that (3.2) holds for k = m + 1, i.e.

t [tm , tm+1 ).
p (tm+1 t0 ) }. e

(3.10)

For the sake of contradiction, suppose (3.10) is not true. Then we dene t = inf{t [tm , tm+1 )|v(t) > M v(t ) = M
p (tm+1 t0 ) e

From (3.9), we know t = tm . By the continuity of v(t) in the interval [tm , tm+1 ), we have and v(t) M < v(t ), and v(t ) v(t) v(t ), t [t , t ].
p (tm+1 t0 ) , e

t [tm , t ].

(3.11)

From (3.9), we have v(tm ) < e M v(t ) = e M

which implies that there exists some t (tm , t ) such that

(3.12)

842

Q. Wang, X. Liu / Applied Mathematics Letters 20 (2007) 839845

Then we know t + s [tm1 , t ] for t [t , t ] and s [, 0] since tk tk1 . By (3.8) and (3.11), we get, for t [t , t ], v(t + s) M = M e M
p (tm t0 ) e p (tm+1 t0 ) (tm+1 tm ) e e p (tm+1 t0 ) e

= e2 v(t ) qv(t), v(t ) v(t )ec = e M

s [, 0]. < v(t ),

Then from condition (ii), we get D + v(t) cv(t); since > c, we have, from (3.12)
p (tm+1 t0 ) c e e

which is a contradiction. This implies the assumption is not true, and hence (3.2) holds for k = m + 1. Thus by mathematical induction, we obtain that (3.2) holds, and hence we have v(t) M
p (tt0 ) , e

t [tk1 , tk ).

Then by condition (i), we get x M

t [tk1 , tk ), k N ,

M where M max{1, [ c1 ] p }; this implies that the trivial solution of system (2.1) is globally exponentially stable with convergence rate p .

Remark 3.1. If the condition > c is removed in Theorem 3.1, then we have to require q max{ec , e2 } in condition (ii) and condition (iv) to be strengthened. The details are stated in the following result whose proof is similar and thus omitted. Theorem 3.2. Assume that there exist a function V 0 and constants p, c, c1 , c2 , > 0 and > such that (i) c1 x p V (t, x) c2 x p , for any t R+ and x Rn ; (ii) D + V (t, (0)) cV (t, (0)), for all t [tk1 , tk ), k N , whenever q V (t, (0)) V (t + s, (s)) for s [, 0], where q max{ec , e2 } is a constant; (iii) V (tk , (0) + Ik (tk , )) dk V (tk , (0)), where dk > 0, k N , are constants; (iv) tk tk1 and ln(dk ) + ( + c) < (tk+1 tk ).
p.

Then the trivial solution of the impulsive system (2.1) is globally exponentially stable and the convergence rate is

Remark 3.2. It is well known that, in the stability theory of delay differential equations, the condition D + V (t, x) cV (t, x) allows the derivative of the Lyapunov function to be positive which may not even guarantee the stability of a delay differential system (see [9,13] and Example 4.1). However, as we can see from Theorems 3.1 and 3.2, impulses have played an important role in exponentially stabilizing a delay differential system. Next, we shall apply the previous theorems to the following linear impulsive delay system: x(t) = Ax(t) + Bx(t (t)), t [tk1 , tk ), x(t) = Ck x(t ), t = tk , k N , xt0 = , where t (t) is strictly increasing on R+ and 0 (t) . Corollary 3.1. Suppose there exist some constants , > 0 such that (i) for some constant q e2 , max (A) + q 2 B < ; 2
1

(3.13)

Q. Wang, X. Liu / Applied Mathematics Letters 20 (2007) 839845

843

(ii) tk tk1 and ln I + Ck + < (tk+1 tk ). 2 2 (3.14)

Then system (3.13) is globally exponentially stable and its convergence rate is . 2 Proof. It follows from Theorem 3.1 on choosing V (x) = x
2.

Corollary 3.2. Suppose there exist some constants , > 0 such that (i) there exists some constant q > 0 such that q max{ec , e2 }, where c = 2(max (A) + q 2 B ); (ii) tk tk1 and ln I + Ck + ( + c) < (tk+1 tk ). 2 2 (3.15)
1

Then system (3.13) is globally exponentially stable and its convergence rate is . 2 Proof. It follows from Theorem 3.2 on choosing V (x) = x 4. An example In this section, we give an example and its simulation to illustrate our results. Example 4.1. Consider the following linear impulsive delay system: x(t) = Ax(t) + Bx t 1 (1 + et ) , t = k 40 x(t) = Ck x(t ), t = k, k N , xt0 = , where 0.1 0.3 , 0.1 0 0 . 0.4
1

2.

(4.1)

0.1 0.2 A = 0.2 0.15 0 0.24 and

0.12 B = 0.12 0

0.03 0 0.2 0.05 0.14 0.1

0.5 0 0.8 Ck = 0 0 0

Then max (A) = 0.4388, B = [max (B B T )] 2 = 0.2905 and I + Ck = 0.6. Choosing q = 2, = 1.7, = 0.05, = 0.2, we nd that the conditions of Corollary 3.1 hold: (i) q = 2 e2 = 1.9739, max (A) + q 2 B = 0.8496 < (ii) 0.05 = tk tk1 = 0.2, ln I + Ck +
2
1

= 0.6808 < (tk+1 tk ) = 0.17. 2

= 0.85;

Thus by Corollary 3.1, we know that the trivial solution of (4.1) is globally exponentially stable with convergence rate 0.85. Furthermore, we can also nd that the conditions of Corollary 3.2 hold: (i) c = 2(max (A) + q 2 B ) = 1.6992, q = 2 max{ec , e2 } = 1.9739; (ii) 0.05 = tk tk1 = 0.2, ln I + Ck + (+c) = 0.1709 < (tk+1 tk ) = 0.17. 2 2
1

844

Q. Wang, X. Liu / Applied Mathematics Letters 20 (2007) 839845

Fig. 1. Impulse-stabilized system.

Fig. 2. System without impulses.

Then we know from Corollary 3.2 that the trivial solution of (4.1) is globally exponentially stable with convergence rate 0.85. The numerical simulation of this impulsive delay differential equation with the initial function (3.7H (t), 2.1H (t), 2.502H (t))T , where H (t) is the Heaviside step function, is given in Fig. 1; the graph of the solution of the corresponding system without impulses is given in Fig. 2. Remark 4.1. As we see from the above pictures, the trivial solution of system (4.1) without impulses is unstable; however, after impulsive control, the trivial solution becomes globally exponentially stable. This implies that impulses may be used to exponentially stabilize some delay differential systems. References
[1] A. Anokhin, L. Berezansky, E. Braverman, Exponential stability of linear delay impulsive differential equations, J. Math. Anal. Appl. 193 (1995) 923941. [2] G. Ballinger, X. Liu, Existence and uniqueness results for impulsive delay differential equations, Dyn. Contin. Discrete Impuls. Syst. 5 (1999) 579591.

Q. Wang, X. Liu / Applied Mathematics Letters 20 (2007) 839845

845

[3] G. Ballinger, X.Z. Liu, Practical stability of impulsive delay differential equations and applications to control problems, in: Optimization Methods and Applications, Kluwer Acad. Publ., Dordrecht, 2001. [4] L. Berezansky, L. Idels, Exponential stability of some scalar impulsive delay differential equation, Commun. Appl. Math. Anal. 2 (1998) 301309. [5] T.P. Chen, Global exponential stability of delayed Hopeld neural networks, Neural Networks 14 (2001) 977980. [6] B. Liu, X.Z. Liu, K. Teo, Q. Wang, Razumikhin-type theorems on exponential stability of impulsive delay systems, IMA J. Appl. Math. 71 (2006) 4761. [7] X.Z. Liu, Stability results for impulsive differential systems with applications to population growth models, Dyn. Stab. Syst. 9 (1994) 163174. [8] X.Z. Liu, Impulsive stabilization of nonlinear systems, IMA J. Math. Control Inform. 10 (1993) 1119. [9] X.Z. Liu, G. Ballinger, Uniform asymptotic stability of impulsive delay differential equations, Comput. Math. Appl. 41 (2001) 903915. [10] J. Shen, J. Yan, Razumikhin type stability theorems for impulsive functional differential equations, Nonlinear Anal. 33 (1998) 519531. [11] I.M. Stamova, G.T. Stamov, LyapunovRazumikhin method for impulsive functional equations and applications to the population dynamics, J. Comput. Appl. Math. 130 (2001) 163171. [12] Q. Wang, X.Z. Liu, Exponential stability for impulsive delay differential equations by Razumikhin method, J. Math. Anal. Appl. 309 (2005) 462473. [13] Z. Luo, J. Shen, Impulsive stabilization of functional differential equations with innite delays, Appl. Math. Lett. 16 (2003) 695701.