Scribd
Upload
24 views

CH 6

This document discusses nonlinear systems of the form x'=F(x,y) and y'=G(x,y). It defines critical points as points where F(x,y)=G(x,y)=0. The linearization of a nonlinear system around a critical point is presented. Three theorems are stated: Poincare-Lyapunov relates stability of the critical point to eigenvalues of the linearization; Grobman-Hartman relates phase portraits of the linearization and nonlinear system if the critical point is hyperbolic; an unstable theorem. Two examples find critical points and sketch phase portraits. Homework problems are given to find and analyze critical points.

Uploaded by

Brij Mohan Singh
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
24 views

CH 6

This document discusses nonlinear systems of the form x'=F(x,y) and y'=G(x,y). It defines critical points as points where F(x,y)=G(x,y)=0. The linearization of a nonlinear system around a critical point is presented. Three theorems are stated: Poincare-Lyapunov relates stability of the critical point to eigenvalues of the linearization; Grobman-Hartman relates phase portraits of the linearization and nonlinear system if the critical point is hyperbolic; an unstable theorem. Two examples find critical points and sketch phase portraits. Homework problems are given to find and analyze critical points.

Uploaded by

Brij Mohan Singh
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 6

Chapter 6

Nonlinear Systems
We consider the nonlinear system
(6.1)
x

= F(x, y),
y

= G(x, y),
where F and G are smooth functions. We assume the solution exists for all t 0
and is unique when initial data is provided. Typically, explicit solutions to (6.1)
cannot found. However, as we will see, we can construct a phase portrait for the
nonlinear system without nding the solutions.
Denition 6.1. A critical point, (x

, y

), (also called an equilibrium, xed, or


stationary point) satises
F(x

, y

) = 0 = G(x

, y

).
Denition 6.2. A trajectory starting at x
0
, y
0
is the set

(x(t), y(t)) | t 0, (x(0), y(0)) = (x


0
, y
0
)

.
Here are some facts about the solutions to the nonlinear system, (6.1):
If the initial data is a critical point, then the solution remains at the critical
point for all time.
Trajectories cannot intersect.
In particular, trajectories not starting at a critical point can never intersect
(in nite time) a critical point.
The interesting dynamics occur near the critical points.
It turns out that we can usually gure out the behavior of the nonlinear system
near a critical point. The rest of the phase portrait can usually be deduced from
this information. To determine the behavior near a critical point, we will linearize
the nonlinear system around the critical point and use our knowledge of linear
systems. We hope the full nonlinear system inherits the behavior of the linearized
system. As we shall see, this is frequently the case.
75
76 6. Nonlinear Systems
The Linearization
To nd the linear approximation to the function f(x) near a point x

, we of
course use Taylors theorem:
f(x) f(x

) + f

(x

)(x x

).
There is a similar version of this theorem for functions depending on two variables.
Indeed,
F(x, y) F(x

, y

) +
F
x
(x

, y

)(x x

) +
F
y
(x

, y

)(y y

).
At a critical point, F(x

, y

) = 0 = G(x

, y

). So the linear approximation to (6.1)


is
(6.2)

x
y

F
x
(x

, y

)
F
y
(x

, y

)
G
x
(x

, y

)
G
y
(x

, y

x x

y y

.
We can clean this up a little if we set u = x x

and v = y y

. Then the
linearization near (x

, y

) becomes

u
v

F
x
(x

, y

)
F
y
(x

, y

)
G
x
(x

, y

)
G
y
(x

, y

u
v

:= J

u
v

.
The matrix J is called the Jacobian.
Before stating the theorems relating the behavior of the linearized system to
the full nonlinear system, we need rst summarize the behavior of linear systems.
We observe three types of behaviors in linear systems. They are (note the origin is
always a critical point in a linear system)
all solutions converge to the origin as t . This happens when the eigen-
values are negative or the have negative real part. In this case we call the
origin asymptotically stable.
solutions near the origin stay near the origin for all time. This happens when
the eigenvalues are purely complex or there is an eigenvalue which is zero while
the other is negative. In this case we call the origin stable..
if neither of the above two occur, we call the origin unstable. That is, at least
one trajectory leaves the vicinity of the origin.
Here are the relevant theorems.
Theorem 6.3. (Poincare-Lyapunov) Suppose x

, y

is a critical point of the non-


linear system (6.1), and suppose the Re(), the real part of the eigenvalues of J (the
linearization) are negative. Then the critical point is locally asymptotically stable.
Theorem 6.4. Suppose x

, y

is a critical point, and the real part of at least one


eigenvalue of J is positive. Then the critical point is unstable.
These theorems only describe the local behavior of solutions near a xed point.
They dont say what happens to the phase portrait. For that we have
6. Nonlinear Systems 77
Theorem 6.5. (Grobman-Hartman) Suppose x

, y

is a hyperbolic critical point


(i.e. the real part of the eigenvalues of J are not zero). Then the phase portrait of
the linearization and the nonlinear equations are locally homeomorphic.
This just says the phase portraits of the linearization and nonlinear equations
are similar provided none of the eigenvalues of the linearization have zero real part.
To summarize:
Robust Cases:
Sources or Repellers: both eigenvalues have positive real part.
Sinks or Attractors: both eigenvalues have negative real part.
Saddles: one eigenvalues is positive and the other negative.
Marginal Cases:
Focus or Center: eigenvalues are pure imaginary. Linearized system does
NOT describe the nonlinear system.
Zero Eigenvalue: usually results from non isolated critical points. If the other
eigenvalue is positive, the critical point is unstable. If the other eigenvalue is
negative, the linearization may NOT describe the nonlinear system.
Example 6.6. Consider
x

= y
y

= x + 2x
3
.
We rst nd the critical points. We must solve x

= y

= 0. This implies y = 0 and


x(1 + 2x
2
) = 0. That is, x = 0 = y is the only critical point.
Next we nd the linearization. Here
F
x
= 0,
F
y
= 1,
G
x
= 1 + 6x
2
, and
G
y
= 0. Thus
J(0, 0) =

0 1
1 0

.
By inspection = 1 and = 0. Therefore the linearization implies a saddle.
Theorem 6.4 implies (0, 0) is unstable. The two phase portraits are given below.
Note the two are similar in accordance with Theorem 6.5.
78 6. Nonlinear Systems
Figure 1. Phase Portrait for the linearization in Example 6.6(left), and the
nonlinear equations (right).
Example 6.7. Consider
x

= 2x + y + xy
3
y

= x 2y xy.
We rst nd the critical points. We must solve x

= y

= 0. The equation for y = 0


implies y =
x
2+x
. Plugging this back in the equation for x = 0, we nd
2x +
x
2 + x
+
x
4
(2 + x)
3
= 0,
or
x(3x
3
+ 13x
2
+ 28x + 20) = 0.
The roots are zero and
x

=
(287 + 18(2019)
2/3
83 13(287 + 18

2019))
1/3
9(287 + 18

2019)
1/3
.
Thus the critical points are (0, 0) and approximately, (1.1934, 1.4797).
The Jacobian is
J(x, y) =

2 + y
3
1 + 3xy
2
1 y 2 x

.
Thus
J(0, 0) =

2 1
1 2

.
Here = 0, = 5, and the origin is a saddle. For the other critical point we
nd the approximate eigenvalues r = 1.023 4.11i, and (1.1934, 1.4797) is an
asymptotically stable spiral.
6. Nonlinear Systems 79
Figure 2. Phase Portrait for the linearization in Example 6.7.
Homework 6.1
In each problem nd the critical points and the corresponding linear system. Find
the eigenvalues of the linearized system and sketch (when possible) a local phase
portrait.
1. x

= x y, y

= x
2
+ y
2
1
2. x

= 1 y, y

= x
2
y
2
3. x

= x + x
2
+ y
2
, y

= y xy
4. x

= (1+x) sin y, y

= 1xcos y
5. x

= 2x+2x
2
, y

= 3x+y +3x
2
6. x

= (yx)(1xy), y

= x(2+y)
7. Consider a pendulum with damping. The governing equations are
m
d
2

dt
2
+
d
dt
+ mgLsin = 0.
with > 0. If we set m = 1, gL = 1,
d
dt
=

and =

, we nd the system

=
= sin .
Find all of the critical points and the associated linearize systems. Using the
theorems in the chapter, sketch a phase portrait for the nonlinear system.
Interpret the phase portrait physically.
80 6. Nonlinear Systems
8. Consider the predator-prey model given by
x = x(1 ax cy)
y = y(1 by + dx).
a) Interpret the model and parameters biologically (what does small/large
a, b, c, d mean?
b) The xed points are
(0, 0), (1/a, 0), (0, 1/b),

b c
ab + cd
,
a + d
ab + cd

.
Check this and interpret the xed points biologically.
c) Linearize the system about the xed points (0, 0), (1/a, 0), (0, 1/b) and
describe the local phase portrait. You will discover that the xed point
(0, 1/b) has two dierent phase portraits depending on the sign of b
c. You can deduce the phase portrait near the remaining xed point
(coexistence) from other xed points.
d) sketch or use pplane (with appropriate values for a, b, c, d) to make phase
portraits when b > c and b < c.
Answers
1. (
1

2
,
1

2
), unstable spiral; (
1

2
,
1

2
), saddle
2. (1, 1), asymptotically stable spiral; (1, 1), saddle
3. (0, 0), unstable node; (1, 0), saddle
4. (0, 2n), indeterminate; (2, (2n 1)), saddle
5. (0, 0), saddle; (1, 0), unstable node
6. (0, 0), saddle; (0, 1), asymptotically stable spiral; (2, 2), asymptotically
stable node; (3, 2), unstable node

You might also like