ABSTRACT
Denition a The state X of the system is absolutely stable on
N
L with the nite attraction time
a
if and only if it is globally stable with
the nite attraction time
a
a
a
X
f for every f N
L
b The state X of the system is absolutely stable with the nite attrac
tion time
a
on N
L if and only if a holds and sup
a
X
f f N
L
for every X
n
Denition a A set A is absolutely stable on N
L with the nite attraction
time
a
if and only if it is globally stable with the nite attraction time
a
a
a
X
f for every f N
L
b A set A is absolutely stable with the nite attraction time
a
on N
L if and
only if a holds and sup
a
X
f f N
L for every X
n
Comments on the practical stability concept
Introductory comments
The practical stability concept aims at determining closeness in a practical engi
neering sense at clarifying the system behaviour either in the free or in a forced
regime over a time interval that can be bounded and for initial states of practical
engineering signicance The closeness can be dened via a suitably accepted dis
be
Figure The state portrait of the system of Example and the sets X
A
and X
A
in the sequel for the system
dX
dt
fX i f
n
l
n
where X
n
and i
l
i is an input vector function
Since the Lyapunov stability concept cannot meet the goals described above then
Chetaev emphasised the need for a nonLyapunov stability concept La Salle
and Lefschetz were probably the rst to introduce a nonLyapunov stability
concept in the western literature
Denition of practical stability
By referring to LaSalle and Lefschetz Weiss and Infante Michel
# Gruji'c # and Martynyuk we shall accept the following prac
tical stability denition in which
and I is the family of inputs of our
interest Besides a set X
is assumed to have a nonempty interior
X
Denition The system is practically stable with respect to f X
A
X
A
Ig if and only if its motions obey XtX
i X
A
for every tX
i
X
A
I
Comment Let the system of Example Section be reconsidered
dX
jx
j jx
j X
Figure The state portrait of the system of Comment and the sets X
A
and X
A
If
X
A
fX X
jx
j jx
j g
X
A
fX X
jx
j jx
j g
and
I f g
then for the system to be practically stable with respect to fX
A
X
A
Ig it is both
necessary and sucient that is suciently small
If it is not satised
ie if Fig then the system is not practically stable with respect to
fX
A
X
A
Ig Notice again that X is asymptotically stable
Comment Let X
and let the system be specied in the following
form
dX
dt
jx
j jx
j X
Evidently I f g If
X
A
fX X
jx
j jx
j g
and
X
A
fX X
jx
j jx
j g
then the system is practically stable with respect to fX
A
X
A
Ig for any
despite X is unstable Fig The same holds if we redene X
A
and X
A
so
that X
A
X
A
fX X
jx
j jx
j g for any
Figure The system is practically stable with respect to fX
A
X
A
Ig
Figure The system is practically contractive with the settling time
s
with respect
to fX
F
X
F
Ig
Denition of practical contraction with settling time
The notion of practical stability does not reect any contraction property which is
expressed by the following in which T
s
s
Denition Motions of the system are practically contractive or for
short the system is practically contractive with settling time
s
with respect
to fX
F
X
F
Ig if and only if XtX
i X
F
for every tX
i T
s
X
F
I
The dierence between practical stability and practical contraction with settling
time
s
is illustrated by Fig and Fig
Figure The system is practically stable with the settling time
s
with respect to
fX
X
A
X
F
Ig
Denition of practical stability with settling time
Denitions and combined imply the following notion
Denition The system is practically contractively stable for short
practically stable with the settling time
s
with respect to f X
X
A
X
F
Ig
X
F
X
A
if and only if both
a XtX
i X
A
for every t
and
b XtX
i X
F
for every t T
s
hold for every X
i X
I
This denition is illustrated by Fig
BLANK PAGE
Chapter
Stability domain concepts
Introductory comments
In order to get complete information about the causality between initial states and
systems motions concepts of domains of various stability properties were intro
duced In the framework of the Lyapunov stability the notion of attraction domain
of the origin was dened by Zubov and Hahn and notions of the stabil
ity domain and the asymptotic stability domain were dened by Gruji'c
# and used by Gruji'c et al # The concept of practical stability
domains was introduced by Gruji'c
In the literature eg LaSalle and Lefschetz and Zubov the notion of
region of asymptotic stability has been used in the sense of the attraction domain
In what follows the dierence between them will be claried
Domains of Lyapunov stability properties
The notion of domain
The term domain denotes a set that can be but need not be open or closed
Domains of Lyapunov stability properties will be called for short Lyapunov
stability domains in a general sense incorporating domains of stability of attraction
and of asymptotic stability In the closer sense the notion Lyapunov stability
domain will be used for the domain of stability for short the stability domain
Lyapunov stability domains will be studied herein in the framework of time
invariant continuoustime nonlinear systems governed by
dX
dt
fX
with possibly certain specic features that will be described when they are needed
In the literature eg LaSalle and Lefschetz the notion region has been
to a
property
Therefore we shall use the term domain in general rather than region In case
a domain is open and connected then we can also call it a region
Denitions of stability domains
The denitions of stability domains were introduced to comply with the denitions
of stability of a state and of a set
Denition a The state X of the system has the stability domain
denoted by D
s
if and only if both
i for every
there is a neighbourhood D
s
of X such that
kXtX
k for all t
holds provided only that X
D
s
and
ii D
s
D
s
b The state X of the system has the strict domain of stability the
strict stability domain D
sc
if and only if
i it has the domain D
s
of stability
ii D
sc
is the largest connected neighbourhood of X which is subset
of D
s
D
sc
D
s
for every
iii D
sc
D
sc
Comment The state X of the system has the stability domain i
it is stable due to Denition and Denition Section Moreover
it is globally stable i its strict stability domain D
sc
is its whole state space
n
D
sc
n
The signicance of the latter is the following for arbitrary X
n
there is possibly suciently large
such that kXtX
k for all t
In other words system motions are bounded on
for every initial state X
Since D
s
can be disconnected and D
sc
cannot then the latter rather than the
former corresponds to B
for every
In other words D
sc
is the domain
of stability in the strict Lyapunov sense while D
s
is the domain of stability in a
wider more practical Lyapunov sense
In order to illustrate the preceding denition let us consider the following ex
ample
Example Let a second order system be in the following form
dX
dt
jx
j jx
j X
It has the innite set S
e
of equilibrium states
jx
j
For any
the maximal denoted by
M
Denition Section obeys
M
For any
the maximal
M
However
kXtX
k t
i X
B
B
S
S
where
S
X X
jx
j jx
j
which is illustrated by Fig Evidently S
S
e
O
It is now obvious Fig that for
p
D
s
B
B
S
S
so that
D
s
D
s
S
In this example the stability domain is the closed set S
ie
D
s
X X
jx
j jx
j
which is showed in Fig f The strict stability domain D
sc
equals D
s
D
sc
D
s
Denition a A set A
n
of states of the system has the stability
domain denoted by D
s
A if and only if both
i for every
there is a neighbourhood D
s
A of A such that
XtX
A for all t
provided only X
D
s
A
and
ii D
s
A D
s
A
b A set A
n
of the states of system has the strict domain of stability
the strict stability domain denoted by D
sc
A if and only if
i it has the stability domain D
s
ii D
sc
A is the largest connected neighbourhood of A which is a subset of
D
s
A D
sc
A D
s
A for all
iii D
sc
A D
sc
A
a b
D
s
B
p
D
s
B
c d
D
s
B
S
D
s
B
S
e f
D
s
S
D
s
D
s
S
Figure Dependence of D
s
on
Figure The state portrait of the system
X jx
j jx
j jx
j jx
j X The
shaded area together with its boundary represents the set A
Example Let the second order system be specied by
dX
dt
jx
j jx
j jx
j jx
j X
We are interested in stability of the set A
A
X X
jx
j jx
j
The state portrait of the system is given in Fig The set S
e
of the equilibrium
states of the system is found as
S
e
X X
X or jx
j jx
j or jx
j jx
j
The graphical analysis of dependence of D
s
A on is presented in Fig
In this example Denition Section
M
p
p
p
However
kXtX
k t
i X
N A
p
N A S
p