Nonlinear Analysis: Hybrid Systems 1 (2007) 223243

www.elsevier.com/locate/nahs
Control vector Lyapunov functions for large-scale impulsive
dynamical systems
Sergey G. Nersesov
a,
, Wassim M. Haddad
b,1
a
Department of Mechanical Engineering, Villanova University, Villanova, PA 19085-1681, United States
b
School of Aerospace Engineering, Georgia Institute of Technology, Atlanta, GA 30332-0150, United States
Received 15 May 2006; accepted 15 May 2006
Abstract
Vector Lyapunov theory has been developed to weaken the hypothesis of standard Lyapunov theory in order to enlarge the
class of Lyapunov functions that can be used for analyzing system stability. In this paper, we provide generalizations to the
recent extensions of vector Lyapunov theory for continuous-time systems to address stability and control design of impulsive
dynamical systems via vector Lyapunov functions. Specically, we provide a generalized comparison principle involving hybrid
comparison dynamics that are dependent on the comparison system states as well as the nonlinear impulsive dynamical system
states. Furthermore, we develop stability results for impulsive dynamical systems that involve vector Lyapunov functions and
hybrid comparison inequalities. Based on these results, we show that partial stability for state-dependent impulsive dynamical
systems can be addressed via vector Lyapunov functions. Furthermore, we extend the recently developed notion of control vector
Lyapunov functions to impulsive dynamical systems. Using control vector Lyapunov functions, we construct a universal hybrid
decentralized feedback stabilizer for a decentralized afne in the control nonlinear impulsive dynamical system that possesses
guaranteed gain and sector margins in each decentralized input channel. These results are then used to develop hybrid decentralized
controllers for large-scale impulsive dynamical systems with robustness guarantees against full modeling and input uncertainty.
c 2006 Elsevier Ltd. All rights reserved.
Keywords: Vector Lyapunov functions; Hybrid comparison principle; Partial stability; Control vector Lyapunov functions; Hybrid decentralized
control; Large-scale impulsive systems
1. Introduction
The mathematical descriptions of many hybrid dynamical systems can be characterized by impulsive differential
equations [16]. Impulsive dynamical systems can be viewed as a subclass of hybrid systems and consist of
three elementsnamely, a continuous-time differential equation, which governs the motion of the dynamical
system between impulsive or resetting events; a difference equation, which governs the way the system states are
instantaneously changed when a resetting event occurs; and a criterion for determining when the states of the system
are to be reset. Since impulsive systems can involve impulses at variable times, they are in general time-varying

Corresponding author. Tel.: +1 610 519 8977; fax: +1 610 519 7312.
E-mail addresses: sergey.nersesov@villanova.edu (S.G. Nersesov), wm.haddad@aerospace.gatech.edu (W.M. Haddad).
1
Tel.: +1 404 894 1078; fax: +1 404 894 2760.
1751-570X/$ - see front matter c 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.nahs.2006.10.006
224 S.G. Nersesov, W.M. Haddad / Nonlinear Analysis: Hybrid Systems 1 (2007) 223243
systems, wherein the resetting events are both a function of time and the systems state. In the case where the resetting
events are dened by a prescribed sequence of times which are independent of the system state, the equations are
known as time-dependent differential equations [1,2,4,79]. Alternatively, in the case where the resetting events are
dened by a manifold in the state space that is independent of time, the equations are autonomous and are known as
state-dependent differential equations [1,2,4,79].
Even though impulsive dynamical systems were rst formulated by Milman and Myshkis [10,11], the fundamental
theory of impulsive differential equations is developed in the monographs by Bainov, Lakshmikantham, Perestyuk,
Samoilenko, and Simeonov [1,2,4,5,12]. These monographs develop qualitative solution properties, existence of
solutions, asymptotic properties of solutions, and stability theory of impulsive dynamical systems. In the recent series
of papers [8,9,13,14], stability, dissipativity, and optimality results have been developed for impulsive dynamical
systems which include invariant set stability theorems, partial stability, dissipativity theory, and optimal control
design.
In this paper, we provide generalizations to the recent extensions of vector Lyapunov theory for continuous-
time systems [15] to address stability and control design of impulsive dynamical systems via vector Lyapunov
functions. Vector Lyapunov theory has been developed to weaken the hypothesis of standard Lyapunov theory in
order to enlarge the class of Lyapunov functions that can be used for analyzing system stability. Lyapunov methods
have also been used by control system designers to obtain stabilizing feedback controllers for nonlinear systems.
In particular, for smooth feedback, Lyapunov-based methods were inspired by Jurdjevic and Quinn [16] who give
sufcient conditions for smooth stabilization based on the ability of constructing a Lyapunov function for the closed-
loop system. More recently, Artstein [17] introduced the notion of a control Lyapunov function whose existence
guarantees a feedback control law which globally stabilizes a nonlinear dynamical system. In general, the feedback
control law is not necessarily smooth, but can be guaranteed to be at least continuous at the origin in addition to being
smooth everywhere else. Even though for certain classes of nonlinear dynamical systems a universal construction
of a feedback stabilizer can be obtained using control Lyapunov functions [18,19], there does not exist a unied
procedure for nding a Lyapunov function candidate that will stabilize the closed-loop system for general nonlinear
systems.
In an attempt to simplify the construction of Lyapunov functions for the analysis and control design of nonlinear
dynamical systems, several researchers have resorted to vector Lyapunov functions as an alternative to scalar
Lyapunov functions. Vector Lyapunov functions were rst introduced by Bellman [20] and Matrosov [21], and further
developed in [2225], with [2224,2629] exploiting their utility for analyzing large-scale systems. The use of vector
Lyapunov functions in dynamical system theory offers a very exible framework since each component of the vector
Lyapunov function can satisfy less rigid requirements as compared to a single scalar Lyapunov function. Weakening
the hypothesis on the Lyapunov function enlarges the class of Lyapunov functions that can be used for analyzing
system stability. In particular, each component of a vector Lyapunov function need not be positive denite with a
negative or even negative-semidenite derivative. Alternatively, the time derivative of the vector Lyapunov function
need only satisfy an element-by-element inequality involving a vector eld of a certain comparison system. Since in
this case the stability properties of the comparison system imply the stability properties of the dynamical system, the
use of vector Lyapunov theory can signicantly reduce the complexity (i.e., dimensionality) of the dynamical system
being analyzed. Extensions of vector Lyapunov function theory that include relaxed conditions on standard vector
Lyapunov functions as well as matrix Lyapunov functions appear in [2830].
The results of this paper closely parallel those in [15] and include a generalized comparison principle involving
hybrid comparison dynamics that are dependent on the comparison system states as well as the nonlinear impulsive
dynamical system states. Next, we develop stability theorems based on hybrid comparison inequalities as well as
partial stability results for impulsive systems using vector Lyapunov functions. Furthermore, we extend the newly
developed notion of control vector Lyapunov functions presented in [15] to impulsive dynamical systems and show
that in the case of a scalar comparison system the denition of a control vector Lyapunov function collapses into a
combination of the classical denition of a control Lyapunov function for continuous-time dynamical systems given
in [17] and the denition of a control Lyapunov function for discrete-time dynamical systems given in [31,32]. In
addition, using control vector Lyapunov functions, we present a universal hybrid decentralized feedback stabilizer for
a decentralized afne in the control nonlinear impulsive dynamical system with guaranteed gain and sector margins.
These results are then used to develop hybrid decentralized controllers for large-scale impulsive dynamical systems
with robustness guarantees against full modeling and input uncertainty.
S.G. Nersesov, W.M. Haddad / Nonlinear Analysis: Hybrid Systems 1 (2007) 223243 225
2. Mathematical preliminaries
In this section we introduce notation and denitions needed for developing stability analysis and synthesis results
for nonlinear impulsive dynamical systems via vector Lyapunov functions. Let R denote the set of real numbers, Z
+
denote the set of nonnegative integers, R
n
denote the set of n 1 column vectors, and ()
T
denote transpose. For
v R
q
we write v 0 (respectively, v 0) to indicate that every component of v is nonnegative (respectively,
positive). In this case, we say that v is nonnegative or positive, respectively. Let R
q
+
and R
q
+
denote the nonnegative
and positive orthants of R
q
, that is, if v R
q
, then v R
q
+
and v R
q
+
are equivalent, respectively, to v 0 and
v 0. Furthermore, let

D
and D denote the interior and the closure of the set D R
n
, respectively. Finally, we write
for an arbitrary spatial vector norm in R
n
, V

(x) for the Fr echet derivative of V at x, B

(), R
n
, > 0, for the
open ball centered at with radius , e R
q
for the ones vector, that is, e [1, . . . , 1]
T
, and x(t ) Mas t
to denote that x(t ) approaches the set M, that is, for each > 0 there exists T > 0 such that dist (x(t ), M) < for
all t > T, where dist ( p, M) inf
xM
p x.
The following denition introduces the notion of class W functions involving quasimonotone increasing functions.
Denition 2.1 ([26]). Afunction w = [w
1
, . . . , w
q
]
T
: R
q
V R
q
, where V R
s
, is of class W
c
if for every xed
y V R
s
, w
i
(z

, y) w
i
(z

, y), i = 1, . . . , q, for all z

, z

R
q
such that z

j
z

j
, z

i
= z

i
, j = 1, . . . , q, i = j ,
where z
i
denotes the i th component of z.
If w(, y) W
c
we say that w satises the Kamke condition [33,34]. Note that if w(z, y) = W(y)z, where
W : V R
qq
, then the function w(, y) is of class W
c
if and only if W(y) is essentially nonnegative for all y V,
that is, all the off-diagonal entries of the matrix function W() are nonnegative. Furthermore, note that it follows from
Denition 2.1 that any scalar (q = 1) function w(z, y) is of class W
c
.
Finally, we introduce the notion of class W
d
functions involving nondecreasing functions.
Denition 2.2 ([25]). A function w = [w
1
, . . . , w
q
]
T
: R
q
V R
q
, where V R
s
, is of class W
d
if for every
xed y V R
s
, w(z

, y) w(z

, y) for all z

, z

R
q
such that z

.
If w(z, y) = W(y)z, where W : V R
qq
, then the function w(, y) is of class W
d
if and only if W(y) is
nonnegative for all y V, that is, all the entries of the matrix function W() are nonnegative. Note that if w(, y) W
d
,
then w(, y) W
c
.
Next, consider the nonlinear comparison system given by
z(t ) = w(z(t ), y(t )), z(t
0
) = z
0
, t I
z
0
, (1)
where z(t ) Q R
q
, t I
z
0
, is the comparison system state vector, y : T V R
s
is a given continuous
function, I
z
0
T R
+
is the maximal interval of existence of a solution z(t ) of (1), Q is an open set, 0 Q, and
w : QV R
q
. We assume that w(, y(t )) is continuous in t and satises the Lipschitz condition
w(z

, y(t )) w(z

, y(t )) Lz

, t T , (2)
for all z

, z

(z
0
), where > 0 and L > 0 is a Lipschitz constant. Hence, it follows from Theorem 2.2 of [35]
that there exists > 0 such that (1) has a unique solution over the time interval [t
0
, t
0
+].
Theorem 2.1 ([15]). Consider the nonlinear comparison system (1). Assume that the function w : Q V R
q
is
continuous and w(, y) is of class W
c
. If there exists a continuously differentiable vector function V = [v
1
, . . . , v
q
]
T
:
I
z
0
Q such that

V(t ) w(V(t ), y(t )), t I

z
0
, (3)
then V(t
0
) z
0
, z
0
Q, implies
V(t ) z(t ), t I
z
0
, (4)
where z(t ), t I
z
0
, is the solution to (1).
Next, we present a stronger version of Theorem 2.1 where the strict inequalities are replaced by soft inequalities.
226 S.G. Nersesov, W.M. Haddad / Nonlinear Analysis: Hybrid Systems 1 (2007) 223243
Theorem 2.2 ([15]). Consider the nonlinear comparison system (1). Assume that the function w : Q V R
q
is
continuous and w(, y) is of class W
c
. Let z(t ), t I
z
0
, be the solution to (1) and [t
0
, t
0
+ ] I
z
0
be a compact
interval. If there exists a continuously differentiable vector function V : [t
0
, t
0
+] Q such that

V(t ) w(V(t ), y(t )), t [t

0
, t
0
+], (5)
then V(t
0
) z
0
, z
0
Q, implies V(t ) z(t ), t [t
0
, t
0
+].
Next, consider the nonlinear dynamical system given by
x(t ) = f (x(t )), x(t
0
) = x
0
, t I
x
0
, (6)
where x(t ) D R
n
, t I
x
0
, is the system state vector, I
x
0
is the maximal interval of existence of a solution x(t )
of (6), D is an open set, 0 D, and f () is Lipschitz continuous on D. The following result is a direct consequence
of Theorem 2.2.
Corollary 2.1 ([15]). Consider the nonlinear dynamical system (6). Assume there exists a continuously differentiable
vector function V : D Q R
q
such that
V

(x) f (x) w(V(x), x), x D, (7)

where w : QD R
q
is a continuous function, w(, x) W
c
, and
z(t ) = w(z(t ), x(t )), z(t
0
) = z
0
, t I
z
0
,x
0
, (8)
has a unique solution z(t ), t I
z
0
,x
0
, where x(t ), t I
x
0
, is a solution to (6). If [t
0
, t
0
+ ] I
x
0
I
z
0
,x
0
is a
compact interval, then V(x
0
) z
0
, z
0
Q, implies V(x(t )) z(t ), t [t
0
, t
0
+].
If in (6) f : R
n
R
n
is globally Lipschitz continuous, then (6) has a unique solution x(t ) for all t t
0
. A more
restrictive sufcient condition for global existence and uniqueness of solutions to (6) is continuous differentiability
of f : R
n
R
n
and uniform boundedness of f

(x) on R
n
. Note that if the solutions to (6) and (8) are globally
dened for all x
0
D and z
0
Q, then the result of Corollary 2.1 holds for any arbitrarily large but compact interval
[t
0
, t
0
+] R
+
. For the remainder of this paper we assume that the solutions to the systems (6) and (8) are dened
for all t t
0
. Continuous differentiability of f () and w(, ) provides a sufcient condition for the existence and
uniqueness of solutions to (6) and (8) for all t t
0
.
3. Stability of impulsive systems via vector Lyapunov functions
In this section, we consider state-dependent impulsive dynamical systems [8] given by
x(t ) = f
c
(x(t )), x(t
0
) = x
0
, x(t ) Z, t I
x
0
, (9)
x(t ) = f
d
(x(t )), x(t ) Z, (10)
where x(t ) D R
n
, t I
x
0
, is the system state vector, I
x
0
is the maximal interval of existence of a solution x(t )
to (9) and (10), D is an open set, 0 D, f
c
: D R
n
is Lipschitz continuous and satises f
c
(0) = 0, f
d
: D R
n
is continuous, x(t ) x(t
+
) x(t ), and Z D R
n
is the resetting set. For a particular trajectory x(t ), t 0, we
let t
k
=
k
(x
0
), x
0
D, denote the kth instant of time at which x(t ) intersects Z. To ensure the well-posedness of the
resetting times we make the following assumptions [8]:
A1. If x(t ) Z \ Z, then there exists > 0 such that, for all 0 < < , x(t +) Z.
A2. If x(t
k
) Z Z, then the system states reset to x
+
(t
k
) x(t
k
) + f
d
(x(t
k
)), according to the resetting law (10),
which serves as the initial condition for the continuous-time dynamics (9).
Assumption A1 ensures that, if a trajectory reaches the closure of Z at a point that does not belong to Z, then the
trajectory must be directed away fromZ, that is, a trajectory cannot enter Z through a point that belongs to the closure
of Z but not to Z. Furthermore, A2 ensures that when a trajectory intersects the resetting set Z, it instantaneously
exits Z. Finally, we note that if x
0
Z then the system initially resets to x
+
0
= x
0
+ f
d
(x
0
) Z, which serves as the
initial condition for continuous-time dynamics (9). Since the resetting times are well dened and distinct, and since
the solution to (9) exists and is unique, it follows that the solution of the impulsive dynamical system (9) and (10)
S.G. Nersesov, W.M. Haddad / Nonlinear Analysis: Hybrid Systems 1 (2007) 223243 227
also exists and is unique over a forward time interval. However, it is important to note that the analysis of impulsive
dynamical systems can be quite involved. In particular, such systems can exhibit Zenoness and beating, as well as
conuence, wherein solutions exhibit innitely many resettings in a nite time, encounter the same resetting surface
a nite or innite number of times in zero time, and coincide after a certain point in time. In this paper we allow for
the possibility of conuence and Zeno solutions; however, A2 precludes the possibility of beating. Furthermore, since
not every bounded solution of an impulsive dynamical system over a forward time interval can be extended to innity
due to Zeno solutions, we assume that existence and uniqueness of solutions are satised in forward time. For details
see [1,2,4].
The next result presents a generalization of the comparison principle given in Corollary 2.1 to impulsive dynamical
systems.
Theorem 3.1. Consider the impulsive dynamical system (9) and (10). Assume there exists a continuously
differentiable vector function V : D Q R
q
such that
V

(x) f
c
(x) w
c
(V(x), x), x Z, (11)
V(x + f
d
(x)) V(x) +w
d
(V(x), x), x Z, (12)
where w
c
: QD R
q
and w
d
: QZ R
q
are continuous functions, w
c
(, x) W
c
, w
d
(, x) W
d
, and the
comparison impulsive dynamical system
z(t ) = w
c
(z(t ), x(t )), z(t
0
) = z
0
, x(t ) Z, t I
z
0
,x
0
, (13)
z(t ) = w
d
(z(t ), x(t )), x(t ) Z, (14)
has a unique solution z(t ), t I
z
0
,x
0
, where x(t ), t I
x
0
, is a solution to (9) and (10). If [t
0
, t
0
+] I
x
0
I
z
0
,x
0
is a compact interval, then
V(x
0
) z
0
, z
0
Q, (15)
implies
V(x(t )) z(t ), t [t
0
, t
0
+]. (16)
Proof. Without loss of generality, let x
0
Z, x
0
D. If x
0
Z, then by Assumption A2, x
0
+ f
d
(x
0
) Z
serves as the initial condition for the continuous-time dynamics. If for x
0
Z the solution x(t ) Z for all
t [t
0
, t
0
+], then the result follows from Corollary 2.1. Next, suppose the interval [t
0
, t
0
+] contains the resetting
times
k
(x
0
) <
k+1
(x
0
), k {1, 2, . . . , m}. Consider the compact interval [t
0
,
1
(x
0
)] and let V(x
0
) z
0
. Then it
follows from (11) and Corollary 2.1 that
V(x(t )) z(t ), t [t
0
,
1
(x
0
)], (17)
where z(t ), t I
z
0
, is the solution to (13). Now, since w
d
(, x) W
d
it follows from (12) and (17) that
V(x(
+
1
(x
0
))) V(x(
1
(x
0
))) +w
d
(V(x(
1
(x
0
))), x(
1
(x
0
)))
z(
1
(x
0
)) +w
d
(z(
1
(x
0
)), x(
1
(x
0
)))
= z(
+
1
(x
0
)). (18)
Consider the compact interval [
+
1
(x
0
),
2
(x
0
)]. Since V(x(
+
1
(x
0
))) z(
+
1
(x
0
)), it follows from (11) that
V(x(t )) z(t ), t [
+
1
(x
0
),
2
(x
0
)]. (19)
Repeating the above arguments for t [
+
k
(x
0
),
k+1
(x
0
)], k = 3, . . . , m, yields (16). Finally, in the case of
innitely many resettings over the time interval [t
0
, t
0
+], let lim
k

k
(x
0
) =

(x
0
) (t
0
, t
0
+]. In this case,
[t
0
,

(x
0
)] = [t
0
,
1
(x
0
)]
_

k=1
[
k
(x
0
),
k+1
(x
0
)]
_
. Repeating the above arguments, the result can be shown for
the interval [t
0
,

(x
0
)].
Note that if the solutions to (9), (10), (13) and (14) are globally dened for all x
0
D and z
0
Q, then the result of
Theorem 3.1 holds for any arbitrarily large but compact interval [t
0
, t
0
+] R
+
. For the remainder of this paper we
228 S.G. Nersesov, W.M. Haddad / Nonlinear Analysis: Hybrid Systems 1 (2007) 223243
assume that the solutions to the systems (9), (10), (13) and (14) are dened for all t t
0
. Next, consider the cascade
nonlinear impulsive dynamical system given by
z(t ) = w
c
(z(t ), x(t )), z(t
0
) = z
0
, x(t ) Z, t t
0
, (20)
x(t ) = f
c
(x(t )), x(t
0
) = x
0
, x(t ) Z, (21)
z(t ) = w
d
(z(t ), x(t )), x(t ) Z, (22)
x(t ) = f
d
(x(t )), x(t ) Z, (23)
where z
0
Q R
q
, x
0
D R
n
, [z
T
(t ), x
T
(t )]
T
, t t
0
, is the solution to (20)(23), w
c
: Q D R
q
and
w
d
: Q Z R
q
are continuous, w
c
(, x) W
c
, w
d
(, x) W
d
, w
c
(0, 0) = 0, f
c
: D R
n
is Lipschitz
continuous on D, f
c
(0) = 0, and f
d
: D R
n
is continuous. The following denition introduces several types of
partial stability of the nonlinear state-dependent impulsive dynamical system (20)(23).
Denition 3.1 ([6]). (i) The nonlinear impulsive dynamical system (20)(23) is Lyapunov stable with respect to z if,
for every > 0 and x
0
R
n
, there exists = (, x
0
) > 0 such that z
0
< implies that z(t ) < for all t 0.
(ii) The nonlinear impulsive dynamical system (20)(23) is Lyapunov stable with respect to z uniformly in x
0
if, for
every > 0, there exists = () > 0 such that z
0
< implies that z(t ) < for all t 0 and for all x
0
R
n
.
(iii) The nonlinear impulsive dynamical system (20)(23) is asymptotically stable with respect to z if it is Lyapunov
stable with respect to z and, for every x
0
R
n
, there exists = (x
0
) > 0 such that z
0
< implies that
lim
t
z(t ) = 0.
(iv) The nonlinear impulsive dynamical system (20)(23) is asymptotically stable with respect to z uniformly in
x
0
if it is Lyapunov stable with respect to z uniformly in x
0
and there exists > 0 such that z
0
< implies that
lim
t
z(t ) = 0 uniformly in z
0
and x
0
for all x
0
R
n
.
(v) The nonlinear impulsive dynamical system (20)(23) is globally asymptotically stable with respect to z if it is
Lyapunov stable with respect to z and lim
t
z(t ) = 0 for all z
0
R
q
and x
0
R
n
.
(vi) The nonlinear impulsive dynamical system (20)(23) is globally asymptotically stable with respect to z
uniformly in x
0
if it is Lyapunov stable with respect to z uniformly in x
0
and lim
t
z(t ) = 0 uniformly in z
0
and x
0
for all z
0
R
q
and x
0
R
n
.
(vii) The nonlinear impulsive dynamical system (20)(23) is exponentially stable with respect to z uniformly in x
0
if there exist scalars , , > 0 such that z
0
< implies that z(t ) z
0
e
t
, t 0, for all x
0
R
n
.
(viii) The nonlinear impulsive dynamical system (20)(23) is globally exponentially stable with respect to z
uniformly in x
0
if there exist scalars , > 0 such that z(t ) z
0
e
t
, t 0, for all z
0
R
q
and x
0
R
n
.
Theorem 3.2. Consider the impulsive dynamical system (9) and (10). Assume that there exist a continuously
differentiable vector function V : D Q R
q
+
and a positive vector p R
q
+
such that V(0) = 0, the scalar
function v : D R
+
dened by v(x) p
T
V(x), x D, is such that v(x) > 0, x = 0, and
V

(x) f
c
(x) w
c
(V(x), x), x Z, (24)
V(x + f
d
(x)) V(x) +w
d
(V(x), x), x Z, (25)
where w
c
: QD R
q
and w
d
: QZ R
q
are continuous, w
c
(, x) W
c
, w
d
(, x) W
d
, and w
c
(0, 0) = 0.
Then the following statements hold:
(i) If the nonlinear impulsive dynamical system (20)(23) is Lyapunov stable with respect to z uniformly in x
0
, then
the zero solution x(t ) 0 to (9) and (10) is Lyapunov stable.
(ii) If the nonlinear impulsive dynamical system (20)(23) is asymptotically stable with respect to z uniformly in x
0
,
then the zero solution x(t ) 0 to (9) and (10) is asymptotically stable.
(iii) If D = R
n
, Q = R
q
, v : R
n
R
+
is radially unbounded, and the nonlinear impulsive dynamical system
(20)(23) is globally asymptotically stable with respect to z uniformly in x
0
, then the zero solution x(t ) 0 to
(9) and (10) is globally asymptotically stable.
(iv) If there exist constants 1, > 0, and > 0 such that v : D R
+
satises
x

v(x) x

, x D, (26)
and the nonlinear impulsive dynamical system (20)(23) is exponentially stable with respect to z uniformly in x
0
,
then the zero solution x(t ) 0 to (9) and (10) is exponentially stable.
S.G. Nersesov, W.M. Haddad / Nonlinear Analysis: Hybrid Systems 1 (2007) 223243 229
(v) If D = R
n
, Q = R
q
, there exist constants 1, > 0, and > 0 such that v : R
n
R
+
satises (26), and
the nonlinear impulsive dynamical system (20)(23) is globally exponentially stable with respect to z uniformly
in x
0
, then the zero solution x(t ) 0 to (9) and (10) is globally exponentially stable.
Proof. Assume there exist a continuously differentiable vector function V : D Q R
q
+
and a positive vector
p R
q
+
such that v(x) = p
T
V(x), x D, is positive denite, that is, v(0) = 0 and v(x) > 0, x = 0. Since
v(x) = p
T
V(x) max
i =1,...,q
{ p
i
}e
T
V(x), x D, where e [1, . . . , 1]
T
, the function e
T
V(x), x D, is also
positive denite. Thus, there exist r > 0 and class K functions , : [0, r] R
+
such that B
r
(0) D and
(x) e
T
V(x) (x), x B
r
(0). (27)
(i) Let > 0 and choose 0 < < min{, r}. It follows from Lyapunov stability of the nonlinear impulsive
dynamical system (20)(23) with respect to z uniformly in x
0
that there exists = ( ) = () > 0 such that if
z
0

1
< , where z
1


q
i =1
|z
i
| and z
i
is the i th component of z, then z(t )
1
< ( ), t t
0
, for any x
0
D.
Now, choose z
0
= V(x
0
) 0, x
0
D. Since V(x), x D, is continuous, the function e
T
V(x), x D, is also
continuous. Hence, for = ( ) > 0 there exists = (( )) = () > 0 such that < and if x
0
< , then
e
T
V(x
0
) = e
T
z
0
= z
0

1
< , which implies that z(t )
1
< ( ), t t
0
. Now, with z
0
= V(x
0
) 0, x
0
D,
and the assumption that w
c
(, x) W
c
and w
d
(, x) W
d
, it follows from (24) and (25), and Theorem 3.1 that
0 V(x(t )) z(t ) on any compact interval [t
0
, t
0
+], and hence, e
T
z(t ) = z(t )
1
, [t
0
, t
0
+]. Let > t
0
be
such that x(t ) B
r
(0), t [t
0
, t
0
+] for all x
0
B

(0). Thus, using (27), it follows that for x

0
< ,
(x(t )) e
T
V(x(t )) e
T
z(t ) < ( ), t [t
0
, t
0
+], (28)
which implies x(t ) < < , t [t
0
, t
0
+ ]. Now, suppose, ad absurdum, that for some x
0
B

(0) there exists

t > t
0
+ such that x(

t ) . Then, for z
0
= V(x
0
) and the compact interval [t
0
,

t ] it follows from (24) and (25),

and Theorem 3.1 that V(x(

t )) z(

t ), which implies that ( ) (x(

t )) e
T
V(x(

t )) e
T
z(

t ) < ( ). This is
a contradiction, and hence, for a given > 0 there exists = () > 0 such that for all x
0
B

(0), x(t ) < , t t

0
,
which implies Lyapunov stability of the zero solution x(t ) 0 to (9) and (10).
(ii) It follows from (i) and the asymptotic stability of the nonlinear impulsive dynamical system (20)(23) with
respect to z uniformly in x
0
that the zero solution x(t ) 0 to (9) and (10) is Lyapunov stable, and there exists > 0
such that if z
0

1
< , then lim
t
z(t ) = 0 for any x
0
D. As in (i), choose z
0
= V(x
0
) 0, x
0
D. It follows
from Lyapunov stability of the zero solution x(t ) 0 to (9) and (10), and the continuity of V : D Q R
q
+
that
there exists = () > 0 such that if x
0
< , then x(t ) < r, t t
0
, and e
T
V(x
0
) = e
T
z
0
= z
0

1
< . Thus, by
asymptotic stability of (20)(23) with respect to z uniformly in x
0
, for any arbitrary > 0 there exists T = T() > t
0
such that z(t )
1
< (), t T. Thus, it follows from (24) and (25), and Theorem 3.1 that 0 V(x(t )) z(t )
on any compact interval [T, T +], and hence, e
T
z(t ) = z(t )
1
, t [T, T +], and, by (27),
(x(t )) e
T
V(x(t )) e
T
z(t ) < (), t [T, T +]. (29)
Now, suppose, ad absurdum, that for some x
0
B

(0), lim
t
x(t ) = 0, that is, there exists a sequence {t
n
}

n=1
,
with t
n
as n , such that x(t
n
) , n Z
+
, for some 0 < < r. Choose = and the interval
[T, T +] such that at least one t
n
[T, T +]. Then it follows from (29) that () (x(t
n
)) < (), which is
a contradiction. Hence, there exists > 0 such that for all x
0
B

(0), lim
t
x(t ) = 0 which along with Lyapunov
stability implies asymptotic stability of the zero solution x(t ) 0 to (9) and (10).
(iii) Suppose D = R
n
, Q = R
q
, v : R
n
R
+
is radially unbounded, and the nonlinear impulsive dynamical
system (20)(23) is globally asymptotically stable with respect to z uniformly in x
0
. In this case, V : R
n
R
q
+
satises (27) for all x R
n
, where the functions , : R
+
R
+
are of class K

[35]. Furthermore, Lyapunov

stability of the zero solution x(t ) 0 to (9) and (10) follows from (i). Next, for any x
0
R
n
and z
0
= V(x
0
) R
q
+
,
identical arguments as in (ii) can be used to show that lim
t
x(t ) = 0, which proves global asymptotic stability of
the zero solution x(t ) 0 to (9) and (10).
(iv) Suppose (26) holds. Since p R
q
+
, then
x

e
T
V(x)

x

, x D, (30)
230 S.G. Nersesov, W.M. Haddad / Nonlinear Analysis: Hybrid Systems 1 (2007) 223243
where

=/ max
i =1,...,q
{ p
i
} and

=/ min
i =1,...,q
{ p
i
}. It follows from the exponential stability of the nonlinear
impulsive dynamical system (20)(23) with respect to z uniformly in x
0
that there exist positive constants , , and
such that if z
0

1
< , then
z(t )
1
z
0

1
e
(t t
0
)
, t t
0
, (31)
for all x
0
D. Choose z
0
= V(x
0
) 0, x
0
D. By continuity of V : D Q R
q
+
, there exists = () > 0
such that for all x
0
B

(0), e
T
V(x
0
) = e
T
z
0
= z
0

1
< . Furthermore, it follows from (24), (25), (30) and (31),
and Theorem 3.1 that for all x
0
B

(0) the inequality

x(t )

e
T
V(x(t )) e
T
z(t ) z
0

1
e
(t t
0
)


x
0

e
(t t
0
)
(32)
holds on any compact interval [t
0
, t
0
+]. This in turn implies that for any x
0
B

(0),
x(t )
_

_1

x
0
e

(t t
0
)
, t [t
0
, t
0
+]. (33)
Now, suppose, ad absurdum, that for some x
0
B

(0) there exists

t > t
0
+ such that
x(

t ) >
_

_1

x
0
e

( t t
0
)
. (34)
Then for the compact interval [t
0
,

t ], it follows from (33) that x(

t )
_

_
1

x
0
e

( t t
0
)
, which is a
contradiction. Thus, inequality (33) holds for all t t
0
establishing exponential stability of the zero solution x(t ) 0
to (9) and (10).
(v) The proof is identical to the proof of (iv).
If V : D Q R
q
+
satises the conditions of Theorem 3.2 we say that V(x), x D, is a vector Lyapunov
function [26]. Note that for stability analysis each component of a vector Lyapunov function need not be positive
denite, nor does it need to have a negative denite time derivative along the trajectories of (9) and (10) between
resettings and negative semi-denite difference across the resettings. This provides more exibility in searching for
a vector Lyapunov function as compared to a scalar Lyapunov function for addressing the stability of impulsive
dynamical systems.
Next, we use the vector Lyapunov stability results of Theorem 3.2 to develop partial stability analysis results for
nonlinear impulsive dynamical systems [14,36]. Specically, consider the nonlinear impulsive dynamical system (9)
and (10) with partitioned dynamics
2
given by
x
I
(t ) = f
Ic
(x
I
(t ), x
II
(t )), x
I
(t
0
) = x
I0
, x(t ) Z, t t
0
, (35)
x
II
(t ) = f
IIc
(x
I
(t ), x
II
(t )), x
II
(t
0
) = x
II0
, x(t ) Z, (36)
x
I
(t ) = f
Id
(x
I
(t ), x
II
(t )), x(t ) Z, (37)
x
II
(t ) = f
IId
(x
I
(t ), x
II
(t )), x(t ) Z, (38)
where x
I
(t ) D
I
, t t
0
, D
I
R
n
I
is an open set such that 0 D
I
, x
II
(t ) R
n
II
, t t
0
, x
I
(t ) = x
I
(t
+
) x
I
(t ),
x
II
(t ) = x
II
(t
+
) x
II
(t ), f
Ic
: D
I
R
n
II
R
n
I
is such that for all x
II
R
n
II
, f
Ic
(0, x
II
) = 0 and f
Ic
(, x
II
) is
locally Lipschitz in x
I
, f
IIc
: D
I
R
n
II
R
n
II
is such that for every x
I
D
I
, f
IIc
(x
I
, ) is locally Lipschitz in x
II
,
f
Id
: D
I
R
n
II
R
n
I
is continuous and f
Id
(0, x
II
) = 0 for all x
II
R
n
II
, f
IId
: D
I
R
n
II
R
n
II
is continuous,
Z D
I
R
n
II
, x(t ) [x
T
I
(t ), x
T
II
(t )]
T
D = D
I
R
n
II
R
n
, t t
0
, x
0
[x
T
I0
, x
T
II0
]
T
, and n
I
+ n
II
= n. For the
nonlinear impulsive dynamical system (35)(38) the denitions of partial stability given in [14] hold. Furthermore,
for a particular trajectory x(t ) = (x
I
(t ), x
II
(t )), t 0, we let t
k
(=
k
(x
I0
, x
II0
)) denote the kth instant of time at
2
Here we use the Roman subscripts I and II as opposed to Arabic subscripts 1 and 2 for denoting the partial states of x in order not to confuse
the partial states with the component states of the vector Lyapunov function.
S.G. Nersesov, W.M. Haddad / Nonlinear Analysis: Hybrid Systems 1 (2007) 223243 231
which x(t ) intersects Z and we assume that Assumptions 1 and 2 hold for x(t ) = (x
I
(t ), x
II
(t )), t 0. Note that
for the impulsive dynamical system (9) and (10), f
c
(x
I
, x
II
) = [ f
T
Ic
(x
I
, x
II
), f
T
IIc
(x
I
, x
II
)]
T
, (x
I
, x
II
) D
I
R
n
II
, and
f
d
(x
I
, x
II
) = [ f
T
Id
(x
I
, x
II
), f
T
IId
(x
I
, x
II
)]
T
, (x
I
, x
II
) D
I
R
n
II
. For the following result dene
V(x
I
, x
II
) V(x
I
+ f
Id
(x
I
, x
II
), x
II
+ f
IId
(x
I
, x
II
)) V(x
I
, x
II
), (39)
for a given vector function V : D
I
R
n
II
R
q
.
Theorem 3.3. Consider the nonlinear impulsive dynamical system (35)(38). Assume that there exist a continuously
differentiable vector function V : D
I
R
n
II
Q R
q
+
, a positive vector p R
q
+
, and class K functions () and
() such that the scalar function v : D
I
R
n
II
R
+
dened by v(x
I
, x
II
) p
T
V(x
I
, x
II
) satises
(x
I
) v(x
I
, x
II
) (x
I
), (x
I
, x
II
) D
I
R
n
II
, (40)
and
V

(x
I
, x
II
) f (x
I
, x
II
) w
c
(V(x
I
, x
II
), x
I
, x
II
), (x
I
, x
II
) Z, (41)
V(x
I
, x
II
) w
d
(V(x
I
, x
II
), x
I
, x
II
), (x
I
, x
II
) Z, (42)
where w
c
: QD
I
R
n
II
R
q
and w
d
: QZ R
q
, where Z D
I
R
n
II
, are continuous, w
c
(, x
I
, x
II
) W
c
,
w
d
(, x
I
, x
II
) W
d
, and w
c
(0, 0, 0) = 0. Then the following statements hold:
(i) If the nonlinear impulsive dynamical system (20), (22) and (35)(38) is Lyapunov (respectively, asymptotically)
stable with respect to z uniformly in (x
I0
, x
II0
), then the nonlinear impulsive dynamical system (35)(38) is
Lyapunov (respectively, asymptotically) stable with respect to x
I
uniformly in x
II0
.
(ii) If D
I
= R
n
I
, Q = R
q
, the functions () and () are class K

, and the nonlinear impulsive dynamical system

(20), (22) and (35)(38) is globally asymptotically stable with respect to z uniformly in (x
I0
, x
II0
), then the
nonlinear impulsive dynamical system (35)(38) is globally asymptotically stable with respect to x
I
uniformly in
x
II0
.
(iii) If there exist constants 1, > 0, and > 0 such that v : D
I
R
n
II
R
+
satises
x
I

v(x
I
, x
II
) x
I

, (x
I
, x
II
) D
I
R
n
II
, (43)
and the nonlinear impulsive dynamical system (20), (22) and (35)(38) is exponentially stable with respect to
z uniformly in (x
I0
, x
II0
), then the nonlinear impulsive dynamical system (35)(38) is exponentially stable with
respect to x
I
uniformly in x
II0
.
(iv) If D
I
= R
n
I
, Q = R
q
, there exist constants 1, > 0, and > 0 such that v : R
n
I
R
n
II
R
+
satises
(43), and the nonlinear impulsive dynamical system (20), (22) and (35)(38) is globally exponentially stable
with respect to z uniformly in (x
I0
, x
II0
), then the nonlinear impulsive dynamical system (35)(38) is globally
exponentially stable with respect to x
I
uniformly in x
II0
.
Proof. Since p R
q
+
is a positive vector it follows from (40) that
(x
I
)/ max
i =1,...,q
{ p
i
} e
T
V(x
I
, x
II
) (x
I
)/ min
i =1,...,q
{ p
i
}, (x
I
, x
II
) D
I
R
n
II
. (44)
Next, let > 0 and note that it follows from Lyapunov stability of the nonlinear impulsive dynamical system (20),
(22) and (35)(38) with respect to z uniformly in (x
I0
, x
II0
) that there exists = () > 0 such that if z
0

1
< ,
where z
1


q
i =1
|z
i
| and z
i
is the i th component of z, then z(t )
1
< ()/ max
i =1,...,q
{ p
i
}, t t
0
, for any
(x
I0
, x
II0
) D
I
R
n
II
. Now, choose z
0
= V(x
I0
, x
II0
) 0, (x
I0
, x
II0
) D
I
R
n
II
. Since V(, ) is continuous, the
function e
T
V(, ) is also continuous. Moreover, it follows from the continuity of () that for = () there exists
= (()) = () > 0 such that < and if x
I0
< , then (x
I0
)/ min
i =1,...,q
{ p
i
} < which, by (44), implies
that e
T
V(x
I0
, x
II0
) = e
T
z
0
= z
0

1
< for all x
II0
R
n
II
, and hence, z(t )
1
< ()/ max
i =1,...,q
{ p
i
}, t t
0
. In
addition, it follows from (41) and (42), and Theorem 3.1 that 0 V(x
I
(t ), x
II
(t )) z(t ) on any compact interval
[t
0
, t
0
+ ], and hence, e
T
z(t ) = z
1
, [t
0
, t
0
+ ]. Thus, it follows from (44) that for all x
I0
< , x
II0
R
n
II
, and
t [t
0
, t
0
+],
(x
I
(t ))/ max
i =1,...,q
{ p
i
} e
T
V(x
I
(t ), x
II
(t )) e
T
z(t ) < ()/ max
i =1,...,q
{ p
i
}, (45)
232 S.G. Nersesov, W.M. Haddad / Nonlinear Analysis: Hybrid Systems 1 (2007) 223243
which implies that x
I
(t ) < , t [t
0
, t
0
+].
Next, suppose, ad absurdum, that for some x
I0
D
I
with x
I0
< and for some x
II0
R
n
II
there exists

t > t
0
+ such that x
I
(

t ) . Then, for z
0
= V(x
I0
, x
II0
) and the compact interval [t
0
,

t ] it follows from
Theorem 3.1 that V(x
I
(

t ), x
II
(

t )) z(

t ) which implies that ()/ max

i =1,...,q
{ p
i
} (x
I
(

t ))/ max
i =1,...,q
{ p
i
}
e
T
V(x
I
(

t ), x
II
(

t )) e
T
z(

t ) < ()/ max

i =1,...,q
{ p
i
}. This is a contradiction and hence, for a given > 0, there
exists = () > 0 such that for all x
I0
D
I
with x
I0
< and for all x
II0
R
n
II
, x
I
(t ) < , t t
0
, which
implies Lyapunov stability of the nonlinear impulsive dynamical system (35)(38) with respect to x
I
uniformly in x
II0
.
The remainder of the proof involves similar arguments as above and as in the proof of parts (ii)(v) of Theorem 3.2
and, hence, is omitted.
Remark 3.1. Note that Theorem 3.3 allows us to address stability of time-dependent nonlinear impulsive dynamical
systems via vector Lyapunov functions. In particular, with x
I
(t ) x(t ), x
II
(t ) t , n
I
= n, n
II
= 1, f
Ic
(x
I
, x
II
) =
f
c
(x(t ), t ), f
IIc
(x
I
(t ), x
II
(t )) = 1, f
Id
(x
I
, x
II
) = f
d
(x(t ), t ), f
IId
(x
I
(t ), x
II
(t )) = 0, and Z = D T , with
T {t
1
, t
2
, . . .}, Theorem 3.3 can be used to establish stability results for the nonlinear time-dependent impulsive
dynamical system given by
x(t ) = f
c
(x(t ), t ), x(t
0
) = x
0
, t = t
k
, t t
0
, (46)
x(t ) = f
d
(x(t ), t ), t = t
k
, (47)
where x
0
D R
n
. For details on the unication between partial stability of state-dependent impulsive systems and
stability theory for time-dependent impulsive systems see [14].
4. Control vector Lyapunov functions for impulsive systems
In this section, we consider a feedback control problem and generalize the notion of a control vector Lyapunov
function introduced in [15] to nonlinear impulsive dynamical systems. Specically, consider the nonlinear controlled
impulsive dynamical system given by
x(t ) = F
c
(x(t ), u
c
(t )), x(t
0
) = x
0
, x(t ) Z, t t
0
, (48)
x(t ) = F
d
(x(t ), u
d
(t )), x(t ) Z, (49)
where x
0
D, D R
n
is an open set with 0 D, u
c
(t ) U
c
R
m
c
, t t
0
, u
d
(t
k
) U
d
R
m
d
, t
k
denotes the kth
instant of time at which x(t ) intersects Z for a particular trajectory x(t ) and input (u
c
(), u
d
()), F
c
: D U
c
R
n
is Lipschitz continuous for all (x, u
c
) D U
c
and satises F
c
(0, 0) = 0, and F
d
: D U
d
R
n
is continuous.
Here, we assume that u
c
() and u
d
() are restricted to the class of admissible control inputs consisting of measurable
functions such that (u
c
(t ), u
d
(t
k
)) U
c
U
d
for all t t
0
and k Z
[t
0
,t )
{k : t
0
t
k
< t }, where the constraint
set U
c
U
d
is given with (0, 0) U
c
U
d
. Furthermore, we assume that u
c
() and u
d
() satisfy sufcient regularity
conditions such that the nonlinear impulsive dynamical system (48) and (49) has a unique solution forward in time.
Let
c
: D U
c
be such that
c
(0) = 0 and let
d
: Z U
d
. If (u
c
(t ), u
d
(t
k
)) = (
c
(x(t )),
d
(x(t
k
))), where x(t ),
t t
0
, satises (48) and (49), then (u
c
(), u
d
()) is called a hybrid feedback control.
Denition 4.1. If there exist a continuously differentiable vector function V = [v
1
, . . . , v
q
]
T
: D Q R
q
+
,
continuous functions w
c
= [w
c1
, . . . , w
cq
]
T
: Q D R
q
and w
d
= [w
d1
, . . . , w
dq
]
T
: Q Z R
q
,
and a positive vector p R
q
+
such that V(0) = 0, v(x) p
T
V(x), x D, is positive denite, w
c
(, x) W
c
,
w
d
(, x) W
d
, w
c
(0, 0) = 0, F
c
(x)
q
i =1
F
ci
(x) = , x D, x Z, x = 0, F
d
(x)
q
i =1
F
di
(x) = , x Z,
where F
ci
(x) {u
c
U
c
: v

i
(x)F
c
(x, u
c
) < w
ci
(V(x), x)}, x D, x Z, x = 0, i = 1, . . . , q, and
F
di
(x) {u
d
U
d
: v
i
(x + F
d
(x, u
d
)) v
i
(x) w
di
(V(x), x)}, x Z, i = 1, . . . , q, then the vector function
V : D Q R
q
+
is called a control vector Lyapunov function candidate.
It follows from Denition 4.1 that if there exists a control vector Lyapunov function candidate, then there exists a
hybrid feedback control law
c
: D U
c
and
d
: Z U
d
such that
V

(x)F
c
(x,
c
(x)) w
c
(V(x), x), x D, x Z, x = 0, (50)
V(x + F
d
(x,
d
(x))) V(x) +w
d
(V(x), x), x Z. (51)
S.G. Nersesov, W.M. Haddad / Nonlinear Analysis: Hybrid Systems 1 (2007) 223243 233
Moreover, if the nonlinear impulsive dynamical system
z(t ) = w
c
(z(t ), x(t )), z(t
0
) = z
0
, x(t ) Z, t t
0
, (52)
x(t ) = F
c
(x(t ),
c
(x(t ))), x(t
0
) = x
0
, x(t ) Z, (53)
z(t ) = w
d
(z(t ), x(t )), x(t ) Z, (54)
x(t ) = F
d
(x(t ),
d
(x(t ))), x(t ) Z, (55)
where z
0
Qand x
0
D, is asymptotically stable with respect to z uniformly in x
0
, then it follows from Theorem 3.2
that the zero solution x(t ) 0 to (53) and (55) is asymptotically stable. In this case, the vector function V : D R
q
+
given in Denition 4.1 is called a control vector Lyapunov function for impulsive dynamical system (48) and (49).
This is a generalization of the notion of control vector Lyapunov functions introduced in [15] for continuous-time
systems. Furthermore, if D = R
n
, Q = R
q
, U
c
= R
m
c
, U
d
= R
m
d
, v : R
n
R
+
is radially unbounded, and the
system (52)(55) is globally asymptotically stable with respect to z uniformly in x
0
, then the zero solution x(t ) 0
to (48) and (49) is globally asymptotically stabilizable.
Remark 4.1. If in Denition 4.1 w
c
(z, x) = w
c
(z), w
d
(z, x) = w
d
(z), and the zero solution z(t ) 0 to
z(t ) = w
c
(z(t )), z(t
0
) = z
0
, x(t ) Z, t t
0
, (56)
z(t ) = w
d
(z(t )), x(t ) Z, (57)
where z
0
Q, is asymptotically stable, then it follows from Theorem 3.2, with w
c
(z, x) = w
c
(z), w
d
(z, x) = w
d
(z),
that V : D QR
q
+
is a control vector Lyapunov function. In this case, the nonlinear impulsive comparison system
(56) and (57) is a time-dependent impulsive dynamical system [8]. To see this, note that the resetting times
k
(x
0
),
k Z
+
, where x(
k
(x
0
)) Z and x(t ) is the solution to (48) and (49), are determined by the state of (48) and (49),
and hence provide a prescribed sequence of the resetting times for (56) and (57) since the dynamics of the impulsive
system (48) and (49) and the comparison system (56) and (57) are decoupled.
In the case where q = 1, w
c
(z, x) w
c
(z), and w
d
(z, x) w
d
(z), Denition 4.1 implies the existence of a
positive-denite continuously differentiable function v : D Q R
+
and continuous functions w
c
: Q R and
w
d
: Q R, where Q R, such that w
c
(0) = 0, F
c
(x) = {u
c
U
c
: v

(x)F
c
(x, u
c
) < w
c
(v(x))} = , x D,
x Z, x = 0, and F
d
(x) = {u
d
U
d
: v(x + F
d
(x, u
d
)) v(x) w
d
(v(x))} = , x Z, which implies
inf
u
c
U
c
v

(x)F
c
(x, u
c
) < w
c
(v(x)), x Z, x = 0, (58)
inf
u
d
U
d
[v(x + F
d
(x, u
d
)) v(x)] w
d
(v(x)), x Z. (59)
Now, the fact that F
c
(x) = , x D, x Z, x = 0, and F
d
(x) = , x Z, implies the existence of a hybrid
feedback control law
c
: D U
c
and
d
: Z U
d
such that v

(x)F
c
(x,
c
(x)) < w
c
(v(x)), x D, x Z, x = 0,
and v(x + F
d
(x,
d
(x))) v(x) w
d
(v(x)), x Z. Moreover, if v : D R
+
is a control vector Lyapunov
function (with q = 1), then it follows from Remark 4.1 that the zero solution z(t ) 0 to the system (56) and (57) is
asymptotically stable and, since q = 1, this implies that w
c
(z) < 0, z Q R
+
, z = 0. Thus, since v() is positive
denite, (58) and (59) can be rewritten as
inf
u
c
U
c
v

(x)F
c
(x, u
c
) < 0, x Z, x = 0, (60)
inf
u
d
U
d
[v(x + F
d
(x, u
d
)) v(x)] w
d
(v(x)), x Z, (61)
which can be regarded as a denition of a scalar control Lyapunov function for impulsive dynamical systems. Even
though such a result has never been addressed in the literature, it can be seen that this is a combination of the classical
denition of a control Lyapunov function for continuous-time dynamical systems given in [17] and the denition of a
control Lyapunov function for discrete-time dynamical systems given in [31,32].
Next, consider the case where the control input to (48) and (49) possesses a decentralized control architecture so
that the dynamics of (48) and (49) are given by
x
i
(t ) = F
ci
(x(t ), u
ci
(t )), x(t ) Z, t t
0
, i = 1, . . . , q, (62)
234 S.G. Nersesov, W.M. Haddad / Nonlinear Analysis: Hybrid Systems 1 (2007) 223243
x
i
(t ) = F
di
(x(t ), u
di
(t )), x(t ) Z, i = 1, . . . , q, (63)
where x
i
(t ) R
n
i
, x(t ) = [x
T
1
(t ), . . . , x
T
q
(t )]
T
, u
ci
(t ) U
ci
R
m
ci
, t t
0
, u
di
(t
k
) U
di
R
m
di
, k Z
+
,
t
k
denotes the kth resetting time for a particular trajectory of (62) and (63),

q
i =1
n
i
= n,

q
i =1
m
ci
= m
c
, and

q
i =1
m
di
= m
d
. Note that x
i
(t ) R
n
i
, t t
0
, i = 1, . . . , q, as long as x(t ) D, t t
0
, and the sets of control
inputs are given by U
c
= U
c1
U
cq
R
m
c
and U
d
= U
d1
U
dq
R
m
d
. In the case of a component
decoupled control vector Lyapunov function candidate, that is, V(x) = [v
1
(x
1
), . . . , v
q
(x
q
)]
T
, x D, it sufces to
require in Denition 4.1 that, for all i = 1, . . . , q,
F
ci
(x) = {u
c
U
c
: v

i
(x
i
)F
ci
(x, u
ci
) < w
ci
(V(x), x)} = , x D, x Z, x = 0, (64)
F
di
(x) = {u
d
U
d
: v
i
(x
i
+ F
di
(x, u
di
)) v
i
(x) w
di
(V(x), x)} = , x Z, (65)
to ensure that F
c
(x) =
_
q
i =1
F
ci
(x) = , x D, x Z, x = 0, and F
d
(x) =
_
q
i =1
F
di
(x) = , x Z. Note that
for a component decoupled control vector Lyapunov function V : D Q R
q
+
, (64) is equivalent to
inf
u
c
U
c
V

(x)F
c
(x, u
c
) w
c
(V(x), x), x D, x Z, x = 0, (66)
and (65) implies
inf
u
d
U
d
[V(x + F
d
(x, u
d
)) V(x)] w
d
(V(x), x), x Z, (67)
where the inmumin (66) and (67) is taken componentwise, that is, for each component of (66) and (67) the inmumis
calculated separately. It follows from the fact that F
c
(x) = , x D, x Z, x = 0, and F
d
(x) = , x Z, that there
exists a hybrid feedback control law
c
: D U
c
and
d
: Z U
d
such that
c
(x) = [
T
c1
(x), . . . ,
T
cq
(x)]
T
, x D,

d
(x) = [
T
d1
(x), . . . ,
T
dq
(x)]
T
, x Z, where
ci
: D U
ci
,
di
: Z U
di
, v

i
(x
i
)F
ci
(x,
ci
(x)) <
w
ci
(V(x), x), x D, x Z, x = 0, and v
i
(x
i
+ F
di
(x,
di
(x))) v
i
(x
i
) w
di
(V(x), x), x Z, i = 1, . . . , q.
Remark 4.2. If w
ci
(V(x), x) = 0 for x D with x
i
= 0, then condition (64) holds for all x D such that x
i
= 0.
Next, we consider the special case of a nonlinear impulsive dynamical system of the form (62) and (63) with afne
control inputs given by
x
i
(t ) = f
ci
(x(t )) + G
ci
(x(t ))u
ci
(t ), x(t ) Z, t t
0
, i = 1, . . . , q, (68)
x
i
(t ) = f
di
(x(t )) + G
di
(x(t ))u
di
(t ), x(t ) Z, i = 1, . . . , q, (69)
where f
ci
: R
n
R
n
i
satisfying f
ci
(0) = 0 and G
ci
: R
n
R
n
i
m
ci
are smooth functions (at least continuously
differentiable mappings) for all i = 1, . . . , q, f
di
: R
n
R
n
i
and G
di
: R
n
R
n
i
m
di
are continuous for all
i = 1, . . . , q, u
ci
(t ) R
m
ci
, t t
0
, and u
di
(t
k
) R
m
di
, k Z
+
, for all i = 1, . . . , q.
Theorem 4.1. Consider the controlled nonlinear impulsive dynamical system given by (68) and (69). If there
exist a continuously differentiable, component decoupled vector function V : R
n
R
q
+
, continuous functions
P
1ui
: R
n
R
1m
di
, P
2ui
: R
n
R
m
di
m
di
, i = 1, . . . , q, w
c
= [w
c1
, . . . , w
cq
]
T
: R
q
+
R
n
R
q
,
w
d
= [w
d1
, . . . , w
dq
]
T
: R
q
+
Z R
q
, and a positive vector p R
q
+
such that V(0) = 0, the scalar function
v : R
n
R
+
dened by v(x) p
T
V(x), x R
n
, is positive denite and radially unbounded, w
c
(, x) W
c
,
w
d
(, x) W
d
, w
c
(0, 0) = 0, and, for all i = 1, . . . , q,
v
i
(x
i
+ f
di
(x) + G
di
(x)u
di
) = v
i
(x
i
+ f
di
(x)) + P
1ui
u
di
+u
T
di
P
2ui
(x)u
di
, x R
n
, u
di
R
m
di
, (70)
v

i
(x
i
) f
ci
(x) < w
ci
(V(x), x), x R
i
, (71)
v
i
(x
i
+ f
di
(x)) v
i
(x
i
)
1
4
P
1ui
(x)P

2ui
(x)P
T
1ui
(x) w
di
(V(x), x), x Z, (72)
where R
i
{x R
n
, x = 0 : v

i
(x
i
)G
ci
(x) = 0}, i = 1, . . . , q, and P

2ui
() is MoorePenrose generalized inverse of
P
2ui
() [37], then V : R
n
R
q
+
is a control vector Lyapunov function candidate. If, in addition, there exist
c
: R
n

R
m
c
and
d
: Z R
m
d
such that
c
(x) = [
T
c1
(x), . . . ,
T
cq
(x)]
T
, x R
n
,
d
(x) = [
T
d1
(x), . . . ,
T
dq
(x)]
T
, x Z,
S.G. Nersesov, W.M. Haddad / Nonlinear Analysis: Hybrid Systems 1 (2007) 223243 235
and the system (52)(55) is globally asymptotically stable with respect to z uniformly in x
0
, then the zero solution
x(t ) 0 to (53) and (55) is globally asymptotically stable and V : R
n
R
q
+
is a control vector Lyapunov function.
Proof. Note that for all i = 1, . . . , q,
inf
u
ci
R
m
ci
v

i
(x
i
)( f
ci
(x) + G
ci
(x)u
ci
) =
_
, x R
i
,
v

i
(x
i
) f
ci
(x), x R
i
,
< w
ci
(V(x), x), x R
n
, x = 0, (73)
which implies that F
ci
(x) = , x R
n
, x = 0, i = 1, . . . , q. Next, note that it follows from a Taylor series
expansion about x

i
= x
i
+ f
di
(x) that P
1ui
(x) = v

i
(x

i
)G
di
(x) and P
2ui
(x) =
1
2
G
T
di
(x)

2
v
i
x
2
i
|
x
i
=x

i
G
di
(x). Since
V() is continuously differentiable it follows that

2
v
i
x
2
i
|
x
i
=x

i
is symmetric, and hence, P
2ui
() is symmetric for all
i = 1, . . . , q. Next, with u
di
=
1
2
P

2ui
(x)P
T
1ui
(x), x Z, i = 1, . . . , q, it follows from (70) and (72) that
v
i
(x
i
+ f
di
(x) + G
di
(x)u
di
) v
i
(x
i
) = v
i
(x
i
+ f
di
(x)) v
i
(x
i
) + P
1ui
(x)u
di
+u
T
di
P
2ui
(x)u
di
= v
i
(x
i
+ f
di
(x)) v
i
(x
i
)
1
4
P
1ui
(x)P

2ui
(x)P
T
1ui
(x)
w
di
(V(x), x), x Z, i = 1, . . . , q, (74)
which implies that F
di
(x) = , x Z. Now, the result is a direct consequence of the denition of a control vector
Lyapunov function by noting that, for component decoupled vector Lyapunov functions, (64) and (65) are equivalent
to F
c
(x) = , x R
n
, x = 0, and F
d
(x) = , x Z, respectively.
Using Theorem 4.1 we can construct an explicit feedback control law that is a function of the control vector
Lyapunov function V(). Specically, consider the hybrid feedback control law
c
(x) = [
T
c1
(x), . . . ,
T
cq
(x)]
T
, x
R
n
, and
d
(x) = [
T
d1
(x), . . . ,
T
dq
(x)]
T
, x Z, given by

ci
(x) =
_

_
_
c
0i
+
(
i
(x) w
ci
(V(x), x)) +
_
(
i
(x) w
ci
(V(x), x))
2
+(
T
i
(x)
i
(x))
2

T
i
(x)
i
(x)
_
_

i
(x),
i
(x) = 0,
0,
i
(x) = 0,
(75)
and

di
(x) =
1
2
P

2ui
(x)P
T
1ui
(x), x Z, (76)
where
i
(x) v

i
(x
i
) f
ci
(x), x R
n
,
i
(x) G
T
ci
(x)v
T
i
(x
i
), x R
n
, and c
0i
> 0, i = 1, . . . , q. The derivative

V()
along the trajectories of the dynamical system (68), with u
c
=
c
(x), x R
n
, given by (75), is given by
v
i
(x
i
) = v

i
(x
i
)( f
ci
(x) + G
ci
(x)
ci
(x))
=
i
(x) +
T
i
(x)
ci
(x)
=
_

_
c
0i

T
i
(x)
i
(x)
_
(
i
(x) w
ci
(V(x), x))
2
+(
T
i
(x)
i
(x))
2
+w
ci
(V(x), x),
i
(x) = 0,

i
(x),
i
(x) = 0,
< w
ci
(V(x), x), x R
n
. (77)
In addition, using (70), the difference of V(x) at the resetting instants with u
d
=
d
(x), x Z, given by (76), is given
by
v
i
(x
i
) = v
i
(x
i
+ f
di
(x) + G
di
(x)
di
(x)) v
i
(x
i
)
= v
i
(x
i
+ f
di
(x)) v
i
(x
i
)
1
4
P
1ui
(x)P

2ui
(x)P
T
1ui
(x)
w
di
(V(x), x), x Z. (78)
236 S.G. Nersesov, W.M. Haddad / Nonlinear Analysis: Hybrid Systems 1 (2007) 223243
Thus, if the zero solution z(t ) 0 to (52)(55) is globally asymptotically stable with respect to z uniformly in
x
0
, then it follows from Theorem 3.2 that the zero solution x(t ) 0 to (68) and (69) with u
c
=
c
(x) =
[
T
c1
(x), . . . ,
T
cq
(x)]
T
, x R
n
, given by (75) and u
d
=
d
(x) = [
T
d1
(x), . . . ,
T
dq
(x)]
T
, x Z, given by (76) is
globally asymptotically stable.
Remark 4.3. If in Theorem 4.1 w
c
(z, x) = w
c
(z), w
d
(z, x) = w
d
(z), and the zero solution z(t ) 0 to (56) and
(57) is globally asymptotically stable, then it follows from Theorem 3.2 that the hybrid feedback control law given
by (75) and (76) is a globally asymptotically stabilizing controller for the nonlinear impulsive dynamical system (68)
and (69).
Remark 4.4. In the case where q = 1, the functions w(, ) and w
d
(, ) in Theorem 4.1 can be set to be identically
zero, that is, w
c
(z, x) 0 and w
d
(z, x) 0. In this case, the feedback control law (75) specializes to Sontags
universal formula [18], the feedback control law (76) specializes to the discrete-time control law given in [32], and
the hybrid feedback control law (75) and (76) is a global stabilizer for the nonlinear impulsive dynamical system (68)
and (69).
Since f
ci
() and G
ci
() are smooth and v
i
() is continuously differentiable for all i = 1, . . . , q, it follows that

i
(x) and
i
(x), x R
n
, i = 1, . . . , q, are continuous functions, and hence,
ci
(x) given by (75) is continuous for
all x R
n
if either
i
(x) = 0 or
i
(x) w
ci
(V(x), x) < 0 for all i = 1, . . . , q. Hence, the feedback control law
given by (75) is continuous everywhere except for the origin. The following result provides necessary and sufcient
conditions under which the feedback control law given by (75) is guaranteed to be continuous at the origin in addition
to being continuous everywhere else.
Proposition 4.1 ([15]). The feedback control law
c
(x) given by (75) is continuous on R
n
if and only if for every
> 0, there exists > 0 such that for all 0 < x < there exists u
ci
R
m
ci
such that u
ci
< and

i
(x) +
T
i
(x)u
ci
< w
ci
(V(x), x), i = 1, . . . , q.
Remark 4.5. If the conditions of Proposition 4.1 are satised, then the feedback control law
c
(x) given by (75) is
continuous on R
n
. However, it is important to note that for a particular trajectory x(t ), t 0, of (68) and (69),
c
(x(t ))
is a piecewise continuous function of time due to state resettings.
5. Stability margins and inverse optimality
In this section, we construct a hybrid feedback control law that is robust to sector bounded input nonlinearities.
Specically, we consider the nonlinear impulsive dynamical system (68) and (69) with nonlinear uncertainties in the
input so that the impulsive dynamics of the system are given by
x
i
(t ) = f
ci
(x(t )) + G
ci
(x(t ))
ci
(u
ci
(t )), x(t ) Z, t t
0
, i = 1, . . . , q, (79)
x
i
(t ) = f
di
(x(t )) + G
di
(x(t ))
di
(u
di
(t )), x(t ) Z, i = 1, . . . , q, (80)
where
ci
()
ci
{
ci
: R
m
ci
R
m
ci
:
ci
(0) = 0 and
1
2
u
T
ci
u
ci

T
ci
(u
ci
)u
ci
< , u
ci
R
m
ci
},
di
()
di

{
di
: R
m
di
R
m
di
:
di
(0) = 0 and
d
u
2
di j

di j
(u
di
)u
di j
<
d
u
2
di j
, u
di
R
m
di
}, i = 1, . . . , q,
di j
() R
and u
di j
R, j = 1, . . . , m
di
, are the j th components of
di
() and u
di
, respectively, and 0
d
< 1 <
d
< .
In addition, we show that for the dynamical system (68) and (69) the hybrid feedback control law to be dened in
Theorem 5.1 is inverse optimal in the sense that it minimizes a derived hybrid performance functional over the set of
stabilizing controllers S(x
0
) {(u
c
(), u
d
()) : u
c
() and u
d
() are admissible andx(t ) 0 as t }.
Theorem 5.1. Consider the nonlinear impulsive dynamical system (68) and (69) and assume that the conditions of
Theorem 4.1 hold with (72) replaced by
v
i
(x
i
+ f
di
(x)) v
i
(x
i
)
1
4
P
1ui
(x)(R
2di
(x) + P
2ui
(x))
1
P
T
1ui
(x) w
di
(V(x), x), x Z, (81)
where R
2di
: Z R
m
di
m
di
is positive denite, w
c
(z, x) w
c
(z), w
d
(z, x) w
d
(z), and with the zero solution
z(t ) 0 to (56) and (57) being globally asymptotically stable. Then the hybrid feedback control law (
ci
(),
di
())
S.G. Nersesov, W.M. Haddad / Nonlinear Analysis: Hybrid Systems 1 (2007) 223243 237
given by (75) and

di
(x) =
1
2
(R
2di
(x) + P
2ui
(x))
1
P
T
1ui
(x), x Z, i = 1, . . . , q, (82)
minimizes the performance functional given by
J(x
0
, u
c
(), u
d
()) =
_

t
0
q

i =1
[L
1ci
(x(t )) +u
T
ci
(t )R
2ci
(x(t ))u
ci
(t )]dt
+

kZ
[t
0
,)
q

i =1
[L
1di
(x(t
k
)) +u
T
di
(t
k
)R
2di
(x(t
k
))u
di
(t
k
)] (83)
in the sense that
J(x
0
,
c
(x()),
d
(x())) = min
(u
c
(),u
d
())S(x
0
)
J(x
0
, u
c
(), u
d
()), x
0
R
n
, (84)
where Z
[t
0
,)
= {k : t
0
t
k
< }, L
1ci
(x)
i
(x) +

i
(x)
2

T
i
(x)
i
(x), x R
n
,
R
2ci
(x)
_
_
_
1
2
i
(x)
I
m
c
i
,
i
(x) = 0,
0,
i
(x) = 0,
(85)

i
(x)
_

_
c
0i
+
(
i
(x) w
ci
(V(x))) +
_
(
i
(x) w
ci
(V(x)))
2
+(
T
i
(x)
i
(x))
2

T
i
(x)
i
(x)
> 0,
i
(x) = 0,
0,
i
(x) = 0,
(86)
L
1di
(x)
T
di
(x)(R
2di
(x) + P
2ui
(x))
di
(x) v
i
(x
i
+ f
di
(x)) + v
i
(x
i
), x Z, and P
2ui
() and v
i
() are given in
Theorem 4.1. Furthermore, J(x
0
,
c
(x()),
d
(x())) = e
T
V(x
0
), x
0
R
n
, where V : R
n
R
q
+
is a control vector
Lyapunov function for the impulsive dynamical system (68) and (69). In addition, the nonlinear impulsive dynamical
system (79) and (80) with the hybrid feedback control law given by (75) and (82) is globally asymptotically stable
for all
ci
()
ci
and
di
()
di
, i = 1, . . . , q, where
d
= max
i =1,...,q
_
1
1+
di
_
,
d
= min
i =1,...,q
_
1
1
di
_
,

di

_

di

di
,
di
inf
xZ

min
(R
2di
(x)),
di
sup
xZ

max
(R
2di
(x) + P
2ui
(x)), i = 1, . . . , q.
Proof. To show that the hybrid feedback control law (75) and (82) minimizes (83) in the sense of (84), dene the
Hamiltonian
H(x, u
c
, u
d
) = H
c
(x, u
c
) + H
d
(x, u
d
), (87)
where
H
c
(x, u
c
)
q

i =1
[L
1ci
(x) +u
T
ci
R
2ci
(x)u
ci
+v

i
(x
i
)( f
ci
(x) + G
ci
(x)u
ci
)], (88)
H
d
(x, u
d
)
q

i =1
[L
1di
(x) +u
T
di
R
2di
(x)u
di
+v
i
(x
i
+ f
di
(x) + G
di
(x)u
di
) v
i
(x
i
)], (89)
and note that H(x,
c
(),
d
()) = 0 and H(x, u
c
, u
d
) 0, x R
n
, u
c
R
m
c
, u
d
R
m
d
, since
H(x, u
c
, u
d
) =
q

i =1
(u
ci

ci
(x))
T
R
2ci
(x)(u
ci

ci
(x))
+
q

i =1
(u
di

di
(x))
T
(R
2di
(x) + P
2ui
(x))(u
di

di
(x)),
x R
n
, u
c
R
m
c
, u
d
R
m
d
. (90)
238 S.G. Nersesov, W.M. Haddad / Nonlinear Analysis: Hybrid Systems 1 (2007) 223243
Thus,
J(x
0
, u
c
(), u
d
()) =
_

t
0
_
H
c
(x(t ), u
c
(t ))
q

i =1
v

i
(x
i
(t ))( f
ci
(x(t )) + G
ci
(x(t ))u
ci
(t ))
_
dt
+

kZ
[t
0
,)
_
H
d
(x(t
k
), u
d
(t
k
))
+
q

i =1
(v
i
(x
i
(t
k
))) v
i
(x
i
(t
k
) + f
di
(x(t
k
)) + G
di
(x(t
k
))u
di
(t
k
))
_
= lim
t
e
T
V(x(t )) +e
T
V(x
0
) +
_

t
0
H
c
(x(t ), u
c
(t ))dt +

kZ
[t
0
,)
H
d
(x(t
k
), u
d
(t
k
))
e
T
V(x
0
)
= J(x
0
,
c
(x()),
d
(x())), (91)
which yields (84).
Next, we show that the uncertain nonlinear impulsive dynamical system (79) and (80) with the hybrid feedback
control law (75) and (82) is globally asymptotically stable for all
ci
()
ci
() and
di
()
di
(). It follows
from Theorem 4.1 that the hybrid feedback control law (75) and (82) globally asymptotically stabilizes the impulsive
dynamical system (68) and (69) and the vector function V : R
n
R
q
+
is a control vector Lyapunov function for the
impulsive dynamical system (68) and (69). Note that with (86) the continuous-time feedback control law (75) can be
rewritten as
ci
(x) =
i
(x)
i
(x), x R
n
, i = 1, . . . , q. Let the control vector Lyapunov function V : R
n
R
q
+
for (68) and (69) be a vector Lyapunov function candidate for (79) and (80). Then the vector Lyapunov derivative
components along the trajectories of (79) are given by
v
i
(x
i
) = v

i
(x
i
)( f
ci
(x) + G
ci
(x)
ci
(
ci
(x))) =
i
(x) +
T
i
(x)
ci
(
ci
(x)), x Z, i = 1, . . . , q. (92)
Note that
ci
(x) = 0, and hence,
ci
(
ci
(x)) = 0 whenever
i
(x) = 0 for all i = 1, . . . , q. In this case, it
follows from (71) that v
i
(x
i
) < w
ci
(V(x)), x Z,
i
(x) = 0, x = 0, i = 1, . . . , q. Next, consider the case where

i
(x) = 0, i = 1, . . . , q. In this case, note that

i
(x) w
ci
(V(x))

i
(x)
2

T
i
(x)
i
(x) =
c
0i

T
i
(x)
i
(x)
2
+
(
i
(x) w
ci
(V(x)))
_
(
i
(x) w
ci
(V(x)))
2
+(
T
i
(x)
i
(x))
2
2
< 0, x Z,
i
(x) = 0, (93)
for all i = 1, . . . , q. Thus, the vector Lyapunov derivative components given by (92) satisfy
v
i
(x
i
) < w
ci
(V(x)) +

i
(x)
2

T
i
(x)
i
(x) +
T
i
(x)
ci
(
ci
(x))
= w
ci
(V(x)) +
1
2
i
(x)

T
ci
(x)
ci
(x)
1

i
(x)

T
ci
(x)
ci
(
ci
(x))
= w
ci
(V(x)) +
1

i
(x)
_

T
ci
(x)
ci
(x)
2

T
ci
(x)
ci
(
ci
(x))
_
w
ci
(V(x)), x Z,
i
(x) = 0, (94)
for all
ci
()
ci
and i = 1, . . . , q.
Next, consider the Lyapunov difference of each component of V() at the resetting instants for the resetting
dynamics (80) with u
di
=
di
(x) given by (82). Note that it follows from (81) that L
1di
(x) + w
di
(V(x), x) 0,
x Z, i = 1, . . . , q, and hence, since
di
()
di
it follows that
S.G. Nersesov, W.M. Haddad / Nonlinear Analysis: Hybrid Systems 1 (2007) 223243 239
v
i
(x
i
) = v
i
(x
i
+ f
di
(x) + G
di
(x)
di
(
di
(x))) v
i
(x
i
)
v
i
(x
i
+ f
di
(x)) v
i
(x
i
) + P
1ui
(x)
di
(
di
(x))
+
T
di
(
di
(x))P
2ui
(x)
di
(
di
(x)) + L
1di
(x) +w
di
(V(x), x)
= P
1ui
(x)
di
(
di
(x)) +
T
di
(
di
(x))P
2ui
(x)
di
(
di
(x))
+
T
di
(x)(R
2di
(x) + P
2ui
(x))
di
(x) +w
di
(V(x), x)
=
T
di
(x)(R
2di
(x) + P
2ui
(x))
di
(x) +
T
di
(
di
(x))P
2ui
(x)
di
(
di
(x))
2
T
di
(x)(R
2di
(x) + P
2ui
(x))
di
(
di
(x)) +w
di
(V(x), x)
= [
di
(
di
(x))
di
(x)]
T
(R
2di
(x) + P
2ui
(x))[
di
(
di
(x))
di
(x)]

T
di
(
di
(x))R
2di
(x)
di
(
di
(x)) +w
di
(V(x), x)

di
[
di
(
di
(x))
di
(x)]
T
[
di
(
di
(x))
di
(x)]
di

T
di
(
di
(x))
di
(
di
(x)) +w
di
(V(x), x)
=
di
(1
2
di
)
_

di
(
di
(x))
1
1 +
di

di
(x)
_
T
_

di
(
di
(x))
1
1
di

di
(x)
_
+w
di
(V(x), x)
=
di
(1
2
di
)
m
di

j =1
_

di j
(
di
(x))
1
1 +
di

di j
(x)
__

di j
(
di
(x))
1
1
di

di j
(x)
_
+w
di
(V(x), x)
w
di
(V(x), x), x Z, i = 1, . . . , q. (95)
Since the impulsive dynamical system (56) and (57) is globally asymptotically stable it follows from Theorem 3.2
that the nonlinear impulsive dynamical system (79) and (80) is globally asymptotically stable for all
ci
()
ci
and

di
()
di
, i = 1, . . . , q.
Remark 5.1. It follows from Theorem 5.1 that with the hybrid feedback stabilizing control law (75) and (82) the
nonlinear impulsive dynamical system (68) and (69) has a sector (and hence gain) margin ((
1
2
, ), (
1
1+
di
,
1
1
di
)),
i = 1, . . . , q, in each decentralized input channel. For details on stability margins for nonlinear dynamical systems,
see [3840].
6. Decentralized control for large-scale impulsive dynamical systems
In this section, we apply the proposed hybrid control framework to decentralized control of large-scale nonlinear
impulsive dynamical systems. Specically, we consider the large-scale dynamical systemG involving energy exchange
between n interconnected subsystems. Let x
i
: [0, ) R
+
denote the energy (and hence a nonnegative quantity)
of the i th subsystem, let u
ci
: [0, ) R denote the control input to the i th subsystem, let
ci j
: R
n
+
R
+
, i =
j, i, j = 1, . . . , n, denote the instantaneous rate of energy ow from the j th subsystem to the i th subsystem between
resettings, let
di j
: R
n
+
R
+
, i = j , i, j = 1, . . . , n, denote the amount of energy transferred from the j th
subsystem to the i th subsystem at the resetting instant, and let Z R
n
+
be a resetting set for the large-scale impulsive
dynamical system G.
An energy balance for each subsystem G
i
, i = 1, . . . , q, yields [6,41]
x
i
(t ) =
n

j =1, j =i
[
ci j
(x(t ))
c j i
(x(t ))] + G
ci
(x(t ))u
ci
(t ), x(t
0
) = x
0
, x(t ) Z, t t
0
, (96)
x
i
(t ) =
n

j =1, j =i
[
di j
(x(t ))
d j i
(x(t ))] + G
di
(x(t ))u
di
(t ), x(t ) Z, (97)
or, equivalently, in vector form for the large-scale impulsive dynamical system G
x(t ) = f
c
(x(t )) + G
c
(x(t ))u
c
(t ), x(t
0
) = x
0
, x(t ) Z, t t
0
, (98)
x(t ) = f
d
(x(t )) + G
d
(x(t ))u
d
(t ), x(t ) Z, (99)
240 S.G. Nersesov, W.M. Haddad / Nonlinear Analysis: Hybrid Systems 1 (2007) 223243
where x(t ) = [x
1
(t ), . . . , x
n
(t )]
T
, t t
0
, f
ci
(x) =

n
j =1, j =i

ci j
(x), where
ci j
(x)
ci j
(x)
c j i
(x), x R
n
+
,
i = j, i, j = 1, . . . , q, denotes the net energy ow from the j th subsystem to the i th subsystem between resettings,
G
c
(x) = diag[G
c1
(x), . . . , G
cn
(x)] = diag[x
1
, . . . , x
n
], x R
n
+
, G
d
(x) = diag[G
d1
(x), . . . , G
dn
(x)], x Z,
G
di
: R
n
R, i = 1, . . . , n, u
c
(t ) R
n
, t t
0
, u
d
(t
k
) R
n
, k Z
+
, f
di
(x) =

n
j =1, j =i

di j
(x), where

di j
(x)
di j
(x)
d j i
(x), x Z, i = j , i, j = 1, . . . , q, denotes the net amount of energy transferred
from the j th subsystem to the i th subsystem at the instant of resetting. Here, we assume that
ci j
: R
n
+
R
+
,
i = j , i, j = 1, . . . , n, are locally Lipschitz continuous on R
n
+
,
ci j
(0) = 0, i = j , i, j = 1, . . . , n, and
u
c
= [u
c1
, . . . , u
cn
]
T
: R R
n
is such that u
ci
: R R, i = 1, . . . , n, are bounded piecewise continuous
functions of time. Furthermore, we assume that
ci j
(x) = 0, x R
n
+
, whenever x
j
= 0, i = j , i, j = 1, . . . , n,
and x
i
+

n
j =1, j =i

di j
(x) 0, x Z. In this case, f
c
() is essentially nonnegative [41,42] (i.e., f
ci
(x) 0 for all
x R
n
+
such that x
i
= 0, i = 1, . . . , n) and x + f
d
(x), x Z R
n
+
, is nonnegative (i.e. x
i
+ f
di
(x) 0 for all x Z,
i = 1, . . . , n). The above constraints imply that if the energy of the j th subsystem of G is zero, then this subsystem
cannot supply any energy to its surroundings between resettings and the i th subsystem of G cannot transfer more
energy to its surroundings than it possesses at the instant of resetting. Finally, in order to ensure that the trajectories of
the closed-loop system remain in the nonnegative orthant of the state space for all nonnegative initial conditions, we
seek a hybrid feedback control law (u
c
(), u
d
()) that guarantees the continuous-time closed-loop system dynamics
(98) are essentially nonnegative and the closed-loop system states after the resettings are nonnegative [43].
For the dynamical system G, consider the control vector Lyapunov function candidate V(x) =
[v
1
(x
1
), . . . , v
n
(x
n
)]
T
, x R
n
+
, given by
V(x) = [x
1
, . . . , x
n
]
T
, x R
n
+
. (100)
Note that V(0) = 0 and v(x) e
T
V(x), x R
n
+
, is such that v(0) = 0, v(x) > 0, x = 0, x R
n
+
, and v(x) as
x with x R
n
+
. Also, note that since v
i
(x
i
) = x
i
, x R
n
, i = 1, . . . , n, are linear functions of x, it follows
that P
2ui
(x) 0, i = 1, . . . , n, and hence, by (76),
di
(x) 0, i = 1, . . . , n. Furthermore, consider the functions
w
c
(V(x), x) =
_

c11
(v
1
(x
1
)) +
n

j =1, j =1

c1 j
(x), . . . ,
cnn
(v
n
(x
n
)) +
n

j =1, j =n

cnj
(x)
_
T
,
x R
n
+
, (101)
w
d
(V(x), x) =
_
n

j =1, j =1

d1 j
(x), . . . ,
n

j =1, j =n

dnj
(x)
_
T
, x Z, (102)
where
ci i
: R
+
R
+
, i = 1, . . . , n, are positive denite functions, and note that w
c
(, x) W
c
, x R
n
+
,
w
c
(0, 0) = 0, w
d
(, ) does not depend on V(), and hence w
d
(, x) W
d
, x Z. Also note that it follows from
Remark 4.2 that R
i
{x R
n
+
, x
i
= 0 : V

i
(x
i
)G
ci
(x) = 0} = {x R
n
+
, x
i
= 0 : x
i
= 0} = , and hence,
condition (71) is satised for V() and w
c
(, ) given by (100) and (101), respectively, and condition (72) is satised
as an equality for w
d
(, ) given by (102) and P
2ui
(x) 0, i = 1, . . . , n.
To show that the impulsive dynamical system
z(t ) = w
c
(z(t ), x(t )), z(t
0
) = z
0
, x(t ) Z, t t
0
, (103)
z(t ) = w
d
(z(t ), x(t )), x(t ) Z, (104)
where z(t ) R
n
+
, t t
0
, x(t ), t t
0
, is the solution to (98) and (99), the i th component of w
c
(z, x) is given
by w
ci
(z, x) =
ci i
(z
i
) +

n
j =1, j =i

ci j
(x), z R
n
+
, x R
n
+
, and the i th component of w
d
(z, x) is given by
w
di
(z, x) =

n
j =1, j =i

di j
(x), x Z, is globally asymptotically stable with respect to z uniformly in x
0
, consider
the partial Lyapunov function candidate v(z) = e
T
z, z R
n
+
. Note that v() is radially unbounded, v(0) = 0, v(z) > 0,
z R
n
+
, z = 0,

v(z) =

n
i =1

ci i
(z
i
) +

n
i =1

n
j =1, j =i

ci j
(x) =

n
i =1

ci i
(z
i
) < 0, z R
n
+
, z = 0, x R
n
+
,
and v(z) =

n
i =1

n
j =1, j =i

di j
(x) = 0, z R
n
+
, x Z. Thus, it follows from Theorem 2.1 of [14] that the
impulsive dynamical system (103), (104), (98) and (99) is globally asymptotically stable with respect to z uniformly
S.G. Nersesov, W.M. Haddad / Nonlinear Analysis: Hybrid Systems 1 (2007) 223243 241
Fig. 6.1. Controlled system states versus time.
in x
0
. Hence, it follows from Theorem 4.1 that V(x), x R
n
+
, given by (100) is a control vector Lyapunov function
for the dynamical system (98) and (99).
Next, using (75) with
i
(x) = v

i
(x
i
) f
ci
(x) =

n
j =1, j =i

ci j
(x),
i
(x) = x
i
, x R
n
+
, and c
0i
> 0, i = 1, . . . , n,
we construct a globally stabilizing hybrid decentralized feedback controller for (98) and (99) given by

ci
(x) =
_

_
_
c
0i
+

ci i
(x
i
) +
_

2
ci i
(x
i
) + x
2
i
x
2
i
_
_
x
i
, x
i
= 0,
0, x
i
= 0,
i = 1, . . . , n, (105)
and

di
(x) 0, i = 1, . . . , n. (106)
It can be seen from the structure of the feedback control law (105) and (106) that the continuous-time closed-loop
system dynamics are essentially nonnegative and the closed-loop system states after the resettings are nonnegative.
Furthermore, since
i
(x) w
ci
(V(x), x) =
ci i
(v
i
(x
i
)), x R
n
+
, i = 1, . . . , n, the continuous-time feedback
controller
c
() is fully independent from f
c
(x) which represents the internal interconnections of the large-scale
system dynamics, and hence, is robust against full modeling uncertainty in f
c
(x).
For the following simulation we consider (98) and (99) with
ci j
(x) =
ci j
x
i
x
j
,
ci i
(x) =
ci i
x
2
i
, and

di j
(x) =
di j
x
j
, where
ci j
0, i = j , i, j = 1, . . . , n,
ci i
> 0, i = 1, . . . , n, and
di j
0, i = j , i, j = 1, . . . , n,
and 1

n
j =1, j =i

d j i
, i = 1, . . . , n. Note that in this case the conditions of Proposition 4.1 are satised, and hence
the continuous-time feedback control law (105) is continuous on R
n
+
. For our simulation we set n = 3,
c11
= 0.1,

c22
= 0.2,
c33
= 0.01,
c12
= 2,
c13
= 3,
c21
= 1.5,
c23
= 0.3,
c31
= 4.4,
c32
= 0.6,
d12
= 0.75,

d13
= 0.33,
d21
= 0.2,
d23
= 0.5,
d31
= 0.66,
d32
= 0.2, c
01
= 1, c
02
= 1, c
03
= 0.25, the resetting set
Z = {x R
3
: x
1
+x
2
0.75 = 0}, with initial condition x
0
= [3, 4, 1]
T
. Fig. 6.1 shows the states of the closed-loop
system versus time and Fig. 6.2 shows continuous-time control signal for each decentralized control channel as a
function of time.
7. Conclusion
In this paper we developed vector Lyapunov stability results for nonlinear impulsive dynamical systems.
Specically, we addressed the generalized comparison principle involving hybrid comparison dynamics that are
dependent on comparison system states as well as the nonlinear impulsive dynamical system states. We presented
Lyapunov stability theorem to provide sufcient conditions for several types of stability using vector Lyapunov
242 S.G. Nersesov, W.M. Haddad / Nonlinear Analysis: Hybrid Systems 1 (2007) 223243
Fig. 6.2. Control signals in each decentralized control channel versus time.
functions. In addition, we addressed partial stability of nonlinear impulsive dynamical systems via vector Lyapunov
functions. Furthermore, we generalized the novel notion of control vector Lyapunov functions to impulsive dynamical
systems. Using this notion, we developed universal decentralized hybrid feedback stabilizer for a decentralized
afne in control nonlinear impulsive dynamical system with robustness guarantees against full modeling and input
uncertainty. Finally, we designed decentralized controllers for large-scale impulsive dynamical systems and presented
a numerical example.
Acknowledgments
This research was supported in part by the Air Force Ofce of Scientic Research under Grant F49620-03-1-0178
and by the Naval Surface Warfare Center under Contract N65540-05-C-0028.
References
[1] V. Lakshmikantham, D.D. Bainov, P.S. Simeonov, Theory of Impulsive Differential Equations, World Scientic, Singapore, 1989.
[2] D.D. Bainov, P.S. Simeonov, Systems with Impulse Effect: Stability, Theory and Applications, Ellis Horwood, Chichester, 1989.
[3] S. Hu, V. Lakshmikantham, S. Leela, Impulsive differential systems and the pulse phenomena, J. Math. Anal. Appl. 137 (1989) 605612.
[4] D.D. Bainov, P.S. Simeonov, Impulsive Differential Equations: Asymptotic Properties of the Solutions, World Scientic, Singapore, 1995.
[5] A.M. Samoilenko, N.A. Perestyuk, Impulsive Differential Equations, World Scientic, Singapore, 1995.
[6] W.M. Haddad, V. Chellaboina, S.G. Nersesov, Impulsive and Hybrid Dynamical Systems. Stability, Dissipativity, and Control, Princeton
University Press, Princeton, NJ, 2006.
[7] R.T. Bupp, D.S. Bernstein, V. Chellaboina, W.M. Haddad, Resetting virtual absorbers for vibration control, J. Vib. Control. 6 (2000) 6183.
[8] W.M. Haddad, V. Chellaboina, N.A. Kablar, Nonlinear impulsive dynamical systems. Part I: Stability and dissipativity, Internat. J. Control 74
(2001) 16311658.
[9] W.M. Haddad, V. Chellaboina, N.A. Kablar, Nonlinear impulsive dynamical systems. Part II: Stability of feedback interconnections and
optimality, Internat. J. Control 74 (2001) 16591677.
[10] V.D. Milman, A.D. Myshkis, On the stability of motion in the presence of impulses, Siberian Math. J. 1 (1960) 233237.
[11] V.D. Milman, A.D. Myshkis, Approximate methods of solutions of differential equations, in: RandomImpulses in Linear Dynamical Systems,
Publ. House. Acad. Sci. Ukr., SSR, Kiev, 1963, pp. 6481.
[12] D.D. Bainov, P.S. Simeonov, Impulsive Differential Equations: Periodic Solutions and Applications, Longman Scientic & Technical,
London, 1993.
[13] V. Chellaboina, S.P. Bhat, W.M. Haddad, An invariance principle for nonlinear hybrid and impulsive dynamical systems, Nonlinear Anal.
TMA 53 (2003) 527550.
[14] W. Haddad, V. Chellaboina, S. Nersesov, A unication between partial stability of state-dependent impulsive systems and stability theory for
time-dependent impulsive systems, Int. J. Hybrid Syst. 2 (2) (2002) 155168.
[15] S.G. Nersesov, W.M. Haddad, On the stability and control of nonlinear dynamical systems via vector Lyapunov functions, IEEE Trans.
Automat. Control 51 (2) (2006) 203215.
S.G. Nersesov, W.M. Haddad / Nonlinear Analysis: Hybrid Systems 1 (2007) 223243 243
[16] V. Jurdjevic, J.P. Quinn, Controllability and stability, J. Differential Equations 28 (1978) 381389.
[17] Z. Artstein, Stabilization with relaxed controls, Nonlinear Anal. TMA 7 (1983) 11631173.
[18] E.D. Sontag, A universal construction of Artsteins theorem on nonlinear stabilization, Systems Control Lett. 13 (1989) 117123.
[19] J. Tsinias, Existence of control Lyapunov functions and applications to state feedback stabilizability of nonlinear systems, SIAM J. Control
Optim. 29 (1991) 457473.
[20] R. Bellman, Vector Lyapunov functions, SIAM J. Control 1 (1962) 3234.
[21] V.M. Matrosov, Method of vector Liapunov functions of interconnected systems with distributed parameters (Survey), Avtomat. i Telemekh.
33 (1972) 6375 (in Russian).
[22] A.N. Michel, R.K. Miller, Qualitative Analysis of Large Scale Dynamical Systems, Academic Press, New York, NY, 1977.
[23] L.T. Gruji c, A.A. Martynyuk, M. Ribbens-Pavella, Large Scale Systems: Stability Under Structural and Singular Perturbations, Springer-
Verlag, Berlin, 1987.
[24] J. Lunze, Stability analysis of large-scale systems composed of strongly coupled similar subsystems, Automatica 25 (1989) 561570.
[25] V. Lakshmikantham, V.M. Matrosov, S. Sivasundaram, Vector Lyapunov Functions and Stability Analysis of Nonlinear Systems, Kluwer,
Dordrecht, The Netherlands, 1991.
[26] D.D.

Siljak, Large-Scale Dynamic Systems: Stability and Structure, Elsevier North-Holland, New York, NY, 1978.
[27] D.D.

Siljak, Complex dynamical systems: Dimensionality, structure and uncertainty, Large Scale Syst. 4 (1983) 279294.
[28] A.A. Martynyuk, Stability by Liapunovs Matrix Function Method with Applications, Marcel Dekker, New York, NY, 1998.
[29] A.A. Martynyuk, Qualitative Methods in Nonlinear Dynamics. Novel Approaches to Liapunovs Matrix Functions, Marcel Dekker, New York,
NY, 2002.
[30] Z. Drici, New directions in the method of vector Lyapunov functions, J. Math. Anal. Appl. 184 (1994) 317325.
[31] G. Amicucci, S. Monaco, D. Normand-Cyrot, Control Lyapunov stabilization of afne discrete-time systems, in: Proc. IEEE Conf. Dec.
Contr., San Diego, CA, 1997, pp. 923924.
[32] V. Chellaboina, W. Haddad, Stability margins of discrete-time nonlinear-nonquadratic optimal regulators, in: Proc. IEEE Conf. Dec. Contr.,
Tampa, FL, 1998, pp. 17861791.
[33] E. Kamke, Zur Theorie der Systeme gew ohnlicher DifferentialGleichungen. II, Acta Math. 58 (1931) 5785.
[34] T. Wa zewski, Syst` emes des equations et des in egalit es diff erentielles ordinaires aux deuxi emes membres monotones et leurs applications,
Ann. Soc. Poln. Mat. 23 (1950) 112166.
[35] H.K. Khalil, Nonlinear Systems, Prentice-Hall, Upper Saddle River, NJ, 1996.
[36] V. Chellaboina, W.M. Haddad, A unication between partial stability and stability theory for time-varying systems, Control Syst. Mag. 22 (6)
(2002) 6675; 23 (2003) 101 (erratum).
[37] D.S. Bernstein, Matrix Mathematics, Princeton University Press, Princeton, NJ, 2005.
[38] R. Sepulchre, M. Jankovi c, P. Kokotovi c, Constructive Nonlinear Control, Springer-Verlag, New York, NY, 1997.
[39] V. Chellaboina, W.M. Haddad, Stability margins of nonlinear optimal regulators with nonquadratic performance criteria involving cross-
weighting terms, Systems Control Lett. 39 (2000) 7178.
[40] W.M. Haddad, V. Chellaboina, S.G. Nersesov, On the equivalence between dissipativity and optimality of nonlinear hybrid controllers, Int. J.
Hybrid Syst. 1 (2001) 5165.
[41] W.M. Haddad, V. Chellaboina, S.G. Nersesov, Thermodynamics. A Dynamical Systems Approach, Princeton University Press, Princeton, NJ,
2005.
[42] W.M. Haddad, V. Chellaboina, Stability and dissipativity theory for nonnegative dynamical systems: A unied analysis framework for
biological and physiological systems, Nonlinear Anal. RWA 6 (2005) 3565.
[43] W.M. Haddad, V. Chellaboina, S.G. Nersesov, Hybrid nonnegative and compartmental dynamical systems, J. Math. Probab. Eng. 8 (2002)
493515.