3.

Fundamental properties

Nonlinear Control Systems

António Pedro Aguiar

[email protected]

3. Fundamental properties

IST-DEEC PhD Course

http://users.isr.ist.utl.pt/%7Epedro/NCS2012/

2012

1
3. Fundamental properties

Example

Consider the system

ẋ = f (t, x) , x (t0 ) = x0

• Does it have a solution over an interval [t0 , t1 ]?

That is, does exist a continuous function x : [t0 , t1 ] → Rm such that ẋ(t) is
defined and satisfies ẋ(t) = f (t, x(t)), ∀t ∈ [t0 , t1 ]?

• Is it unique? or is possible to have more than one solution?

• ... and if we restrict f (t, x) to be continuous in x and piecewise continuous in t.

Is this sufficient to guarantee existence and uniqueness?
No!, e.g.
1/3
ẋ = x , x(0) = 0
` 2t ´3/2
It has solution x(t) = 3 .
But is not unique, since x(t) = 0 is another solution!

2
3. Fundamental properties

Lipschitz condition

A function f (x) is said to be locally Lipschitz on a domain (open and connected set)
D ⊂ Rn if each point of D has a neighborhood D0 such that f (.) satisfies

kf (x) − f (y)k ≤ L kx − yk , ∀x, y ∈ D0

(with the same Lipschitz constant L). The same terminology is extended to a function
f (t, x), provided that the Lipschitz constant holds uniformly in t for all t in a given
interval.

Remark: For f : R → R, we have

|f (x) − f (y)|
≤L
|x − y|

which means that a straight line joining any two points of f (.) cannot have a slope
whose absolute value is greater than L.

Example
1 −2/3
f (x) = x1/3 is not locally Lipschitz at x = 0 since f 0 (x) = 3
x → ∞ as x → 0.

3
3. Fundamental properties

Existence and Uniqueness

Theorem 3.1 - Local Existence and Uniqueness

Let f (t, x) be piecewise continuous in t and satisfy the Lipschitz condition

kf (t, x) − f (t, y)k ≤ L kx − yk

∀x, y ∈ B = {x ∈ Rn : kx − x0 k ≤ r} , ∀t ∈ [t0 , t1 ]. Then, there exists some δ > 0

such that the state equation ẋ = f (t, x) with x(t0 ) = x0 has a unique solution over
[t0 , t0 + δ].

Theorem 3.2 - Global Existence and Uniqueness

Suppose that f (t, x) is piecewise continuous in t and satisfies

kf (t, x) − f (t, y)k ≤ L kx − yk , ∀x, y ∈ Rn , ∀t ∈ [t0 , t1 ]

Then, the state equation ẋ = f (t, x), with x(t0 ) = x0 , has a unique solution over
[t0 , t1 ].

4
3. Fundamental properties

Existence and Uniqueness

Lemma 3.2
h i
∂f
If f (t, x) and ∂x (t,x)
are continuous on [a, b] × D for some domain D ⊂ Rn , then f
is locally Lipschitz in x on [a, b] × D.

Lemma 3.3
h i
∂f
If f (t, x) and are continuous on [a, b] × Rn , then f is globally Lipschitz in x
∂x (t,x)
h i
on [a, b] × Rn if and only if ∂f
∂x
is uniformly bounded on [a, b] × Rn .

5
3. Fundamental properties

Examples

1.
ẋ = A(t)x + g(t) = f (t, x) (1)
with A(t), g(t) piecewise continuous functions of t.

kf (t, x) − f (t, y)k = kA (t) x + g (t) − (A (t) y + g (t))k

= kA (t) (x − y)k ≤ kA (t)k k(x − y)k

Note that for any finite interval of time [t0 , t1 ], the elements of A(t) are
bounded. Thus kA(t)k ≤ a for any induced norm and

kf (t, x) − f (t, y)k ≤ a k(x − y)k

Therefore, from Theorem 3.1 we conclude that (1) has a unique solution over
[t0 , t1 ]. Since t1 can be arbitrarily large it follows that the system has a unique
solution ∀t ≥ t0 . There is no finite escape time.

6
3. Fundamental properties

Examples

2.
ẋ = −x3 = f (x), x∈R (2)
Is it globally Lipschitz?
No! From Lemma 3.3, f (x) is continuous but the Jacobian ∂f ∂x
= −3x2 is not
globally bounded. Nevertheless, ∀x(t0 ) = x0 , (2) has the unique solution
s
x20
x(t) = sgn(x0 )
1 + 2x20 (t − t0 )

3.
ẋ = −x2 , x(0) = −1 (3)
From Lemma 3.2 we conclude that is locally Lipschitz in any compact subset of R
because f (x) and ∂f
∂x
are continuos. Hence, there exists a unique solution over
[0, δ] for some δ > 0. In particular the solution is

1
x(t) =
t−1
and only exists over [0, 1), i.e, there is finite escape time at t = 1!

7
3. Fundamental properties

Uniqueness and Existence

Theorem 3.3
Let f (t, x) be piecewise continuous in t and locally Lipschitz in x for all t ≥ t0 and all
x ∈ D ⊂ Rn . Let W be a compact subset of D, x0 ∈ W and suppose it is known that
every solution of
ẋ = f (t, x), x(t0 ) = x0
lies entirely in W . Then, there is a unique solution that is defined for all t ≥ t0 .
Proof.
By Theorem 3.1, there is a unique solution over [t0 , t0 + δ]. Let [t0 , T ) be its
maximum interval of existence. We would like to show that T = ∞. Suppose that is
not, i.e., T is finite. Then the solution must leave any compact subset of D. But this
is a contradiction, becuase x never leaves the compact set W . Thus we can conclude
that T = ∞.

8
3. Fundamental properties

Example

Returning to the example

ẋ = −x3
It is locally Lipschitz in R. For any initial condition x(0) = x0 ∈ R, the solution cannot
leave the compact set W = {x ∈ R : |x| ≤ x0 } because for any instant of time
• if x > 0 then ẋ < 0
• if x < 0 then ẋ > 0
Thus, without computing explicitly the solution we can conclude from Theorem 3.3
that the system has a unique solution for all t ≥ 0.

9
3. Fundamental properties

Continuous dependence on initial conditions and parameters

Consider the following nominal model

ẋ = f (t, x, λ0 ) (4)

where λ0 ∈ Rp denotes the nominal vector of constant parameters of the model and
x ∈ Rn is the state.
• Let y(t) be a solution of (4) that starts at y(t0 ) = y0 and is defined on the
interval [t0 , t1 ].
• Let z(t) be a solution of ẋ = f (t, x, λ) defined on [t0 , t1 ] with z(t0 ) = z0 .

When does z(t) remains close to y(t)?

Or in other words, is the solution continuous dependent on the initial condition and
parameter λ? That is,

∀ε>0 ∃δ>0 : kz0 − y0 k < δ, kλ − λ0 k < δ ⇒ kz (t) − y (t)k < ε, ∀t∈[t0 ,t1 ]

10
3. Fundamental properties

Continuous dependence on initial conditions and parameters

Theorem 3.4
Let f (t, x) be piecewise continuous in t and Lipschitz in x (with a Lipschits constant
L) on [t0 , t1 ] × W , where W ⊂ Rn is an open connected set. Let y(t) and z(t) be
solutions of

ẏ = f (t, y) , y (t0 ) = y0
ż = f (t, z) + g (t, z) , z (t0 ) = z0

such that y(t), z(t) ∈ W, ∀t ∈ [t0 , t1 ]. Suppose that

kg (t, z)k ≤ µ, ∀ (t, x) ∈ [t0 , t1 ] × W

for some µ > 0. Then

µ “ L(t−t0 ) ”
ky (t) − z (t)k ≤ ky0 − z0 k eL(t−t0 ) + e −1 , ∀t ∈ [t0 , t1 ]
L

11
3. Fundamental properties

Proof

Zt
y (t) = y0 + f (τ, y (τ )) dτ
t0

Zt
z (t) = z0 + [f (τ, z (τ )) + g (τ, z (τ ))] dτ
t0

Subtracting and taking norms yields

Zt Zt
ky (t) − z (t)k ≤ ky0 − z0 k + kf (τ, y (τ )) − f (τ, z (τ ))k dτ + kg (τ, z (τ ))k dτ
t0 t0

Zt
≤ ky0 − z0 k + L ky (τ ) − z (τ )k dτ + µ (t − t0 )
| {z }
γ t0

12
3. Fundamental properties

Applying the Gronwall-Bellman inequality, yields

Zt
ky (t) − z (t)k ≤ γ + µ (t − t0 ) + L [γ + µ (τ − t0 )] eL(t−τ ) dτ
t0

Integrating the right-hand side by parts ( uv 0 = uv − vu0 ) we obtain

R R

Zt
ky (t) − z (t)k ≤ γ + µ(t − t0 ) − γ − µ(t − t0 ) + γeL(t−t0 ) + µeL(t−s) ds
t0

L(t−t0 ) µ “ L(t−t0 ) ”
= γe + e −1
L

13
3. Fundamental properties

Theorem 3.5 - Continuity of solutions

Let f (t, x, λ) be continuous in (t, x, λ) and locally Lipschitz in x on
[t0 , t1 ] × D × {kλ − λ0 k ≤ c}, where D ⊂ Rn is an open connected set. Let y(t, λ0 )
be a solution of
ẋ = f (t, x, λ0 ),
with y(t0 , λ0 ) = y0 ∈ D. Suppose y(t, λ0 ) is defined and belongs to D for all
t ∈ [t0 , t1 ]. Then, given  > 0, there is δ > 0 such that if

kz0 − y0 k < δ, kλ − λ0 k < δ

then there is a unique solution z(t, λ) of ẋ = f (t, x, λ) defined on [t0 , t1 ], with

z(t0 , λ) = z0 such that kz(t, λ) − y(t, λ0 )k < , ∀t ∈ [t0 , t1 ]
Proof.
ż = f (t, z, λ0 ) + f (t, z, λ) − f (t, z, λ0 )
| {z }
g(t,z)

By continuity ∀µ > 0, ∃δ > 0:

kλ − λ0 k < δ ⇒ kg(t, z)k ≤ µ

Therefore, using Theorem 3.4 we conclude Theorem 3.5 by noticing that ky0 − z0 k
and µ can be chosen arbitrarily small.

14
3. Fundamental properties

Comparison Principle

Quite often when we study the state equation ẋ = f (t, x) we need to compute bounds
on the solution x(t). For that we have
• Gronwall-Bellman inequality
• The comparison Lemma → Compares the solution of the differential inequality
v̇(t) ≤ f (t, v(t)) with the solution of u̇(t) = f (t, u). Moreover, v(t) is not needed
to be differentiable.

Definition
Upper right-hand derivative

v(t + h) − v(t)
D+ v(t) = lim sup
h→0+ h

The following properties hold:

• if v(t) is differentiable at t then D + v(t) = v̇(t)
• if 1 |v(t + h) − v(t)| ≤ g(t, h) and limh→0+ g(t, h) = g0 (t), then
h
D+ v(t) ≤ g0 (t)

15
3. Fundamental properties

Comparison Principle

Lemma 3.4 - Comparison Lemma

Let
u̇ = f (t, u), u(t0 ) = u0 , µ ∈ R
where f (t, u) is continuous in t and locally Lipschitz in u, for all t ≥ 0 and u ∈ J ⊂ R.
Let [t0 , T ) (T can be ∞) be the maximal interval of existence of the solution
u(t) ∈ J. Let v(t) be a continuous function that satisfies

D+ v(t) ≤ f (t, v), v(t0 ) ≤ u0

with v(t) ∈ J for all t ∈ [t0 , T ). Then, v(t) ≤ u(t), ∀t ∈ [t0 , T )

16
3. Fundamental properties

Example

Show that the solution of

ẋ = f (x) = −(1 + x2 )x, x(0) = a

is unique and defined for all t ≥ 0.

Because f (x) is locally Lipschitz it has a unique solution on [0, t1 ] for some t1 > 0.
Let v(t) = x2 (t). Then

v̇(t) ≤ −2v(t), v(0) = a2

Let u(t) be the solution of

u̇ = −2u, u(0) = a2 −→ u(t) = a2 e−2t

Then, by comparison lemma the solution x(t) is defined ∀t ≥ 0 and satisfies

|x(t)| = v(t) ≤ |a| e−t , ∀t ≥ 0

p

By Theorem 3.3 it follows that the solution is unique and defined for all t ≥ 0.

17