Consider the autonomous system where
is a locally Lipschitz map from a domain into and there is at least one
equilibrium point , that is
.
Goal: Stability analysis of the equilibrium point
.
Without loss of generality, we consider that . If it is not, then consider
the change of variables . Then
Definition:
The equilibrium point
is
stable if, for each , there is such that
unstable if it is not stable
asymptotically stable if it is stable and can be chosen such that
A function
is said to be
positive definite if
and
positive semidefinite if and
negative definite (resp. negative semi definite) if is definite positive (resp.
definite semi positive).
In particular, for (quadratic form), where is a real symmetric matrix, is positive (semi)definite if and
only if all the eigenvalues of
are positive (nonnegative), which is true if all leading principal
minors (顺序主子式) of are
positive (all principal minors of
are nonnegative). (阶矩阵
为正定矩阵的充要条件是A的所有顺序主子式 , 为实对称矩阵)
Example: The leading principal minors of are Therefore,
if .
Lyapunov's stability theorem
Theorem: Lyapunov's stability theorem
Let be an equilibrium
point for and be a domain
containing . Let be a continuously
differentiable function such that
Then, is stable. Moreover,
if
then is asymptotically
stable.
Examples: Consider the Pendulum example without friction Assume the following energy function Clearly, and
. Thus the origin is stable. Since , we can also conclude that
the origin is not asymptotically stable.
Examples: Pendulum equation, but this time with friction Consider
Then, which is negative semidefinite. why? because for irrespective of the value of
. We can only conclude that the
origin is stable! (use Corollary of La Salle's
Theorem)
Theorem: Globally Asymptotically Stability (GAS)
Let be an equilibrium
point for and be a domain
containing . Let be a
continuously differentiable function such that
Then, is globally
asymptotically stable (GAS).
La Salle's Theorem
Theorem 4.11 - La Salle's Theorem
Let
be a compact
positively invariant set.
be a
continuously differentiable function such that in .
be the largest invariant set
in .
Then every solution starting in approaches as .
Note that is not needed to
be positive definite.
When is the origin?
Corollary 4.1
Let be an equilibrium
point of . Let be a positive definite function containing
the origin such that in . Let and suppose that no solution can stay
identically in , other than the
trivial solution . Then,
the origin is asymptotically stable.
Corollary 4.2
Let be an equilibrium
point of . Let be a
, radially unbounded, positive
definite function such that for all . Let and suppose that no solution can stay identically
in , other than the trivial
solution . Then, the origin
is global asymptotically stable (GAS).
Example: Consider Clearly and
as . i.e, negative semidefinite. Is it GAS?
Let be a solution that
belongs identically to : Therefore, the only solution that can stay identically in
is the trivial solution . Thus, is GAS.
Comparison Functions
Definition:
A continuous function is said to belong to class if it is strictly increasing
and . It is of class
if and as .
Definition:
A continuous function is said to belong to class if, fore each fixed , the mapping belong to class with respect to and, for each fixed , the mapping is decreasing with respect to
and as .
Examples:
Class : Class : Class :
Lemma 4.4
Consider the scalar autonomous differential equation where is a locally
Lipschitz class
function defined on . Then,
for all , the
solution is unique and defined for all . Moreover, where is defined on .
Consider the nonautonomous system Definition: The origin is an equilibrium point at if
The equilibrium point x = 0 of the nonautonomous system is
Stable if
Uniformly Stable (US) if is independent of
Asymptotically Stable if it is stable and as
Uniformly Asymptotically Stable (UAS) if is independent of and the convergence is uniformly in
, that is
Globally Uniformly Asymptotic Stable (GUAS) if can be chosen to
satisfies and
The following lemma make the above definitions more clear.
Lemma 4.5
The equilibrium point of
the nonautonomous system is
US there
exist and
(independent of ) such that
UAS there
exist and
(independent of ) such that
GUAS the
inequality in UAS holds for all , that is
Definition
The equilibrium point is
exponentially stable if there exists positive constants and such that and globally exponentially stable if this inequality is
satisfied
Theorem 4.8:
Let be an equilibrium
point that belongs to . Let be a
function such that and where are
continuous positive functions on .
Then, is uniformly
stable.
Theorem 4.9 Same assumptions as Theorem
4.8 but with where is a
continuous positive definite function on . Then, is UAS.
Moreover, if and are such that Then for some .
If in addition
and is radially unbounded
then is GUAS.
Definition:
is said to be positive semidefinite if
is said to be positive definite if ,
for some positive definite
function
is said to be unbounded if is radially unbounded.
Theorem 4.10:
Let be an equilibrium
point for the nonautonomous system that belongs to some . Let be a function such that where and a
are positive constants. Then
is exponentially stable. If the assumption holds globally. Then is GES.
Proof: Hence, Examples 1 : Consider Then, and ,
thus is GUAS. Note we cannot
conclude exponential because is not the same .
Example 2 Consider Note that Thus is positive
definite and radially unbounded. Therefore, , and are positive definite quadratic
functions () and so we conclude that
is GES.
Boundedness and Ultimate Boundedness
Until now we have used Lyapunov theory to study the behavior of the
system about the equilibrium point. What happens when the system does
not have any equilibrium point? We will see that Lyapunov analysis can
be used to show boundedness of the solution of the state equation.
Example: There are no equilibrium points!
Nevertheless, with Note that , which means that the set with is an
invariant set because . Hence, the solutions are uniformly
bounded.
Moreover for some in
the set which
shows that will reach in finite
time and the solution enter the set .
Thus, we can conclude that the solution is uniformly ultimately
bounded with ultimate bound .
Definition:
The solutions of
are
Uniformly Bounded (UB) if there exists a , independent of such that
Globally Uniformly Bounded (GUB) if can be arbitrarily large
Uniformly Ultimately Bounded (UUB) with ultimate
bound , if there exists (independent of ) such that
Globally Uniformly Ultimately Bounded (GUUB) if
can be arbitrarily
large.
Theorem 4.18 (Global)
Let be a
function such that where and . Then, there exists a and
such that and
Example - Mass-spring system and assuming certain numerical values we get Choosing Then Therefore which shows that the solutions are GUUB. To compute the
ultimate bound, we have to find .