Scribd
Upload
86 views3 pages

Math 677. Fall 2009. Homework #4 Solutions

This document contains solutions to homework problems involving nonlinear dynamical systems and differential equations. It includes: 1) Determining the flow of a nonlinear system and showing an invariant set. 2) Analyzing equilibria and eigenvalues of the Lorenz system. 3) Showing a homeomorphism between a nonlinear system and its linearization. 4) Calculating successive approximations to find stable and unstable manifolds.

Uploaded by

Rodrigo Kosta
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
86 views3 pages

Math 677. Fall 2009. Homework #4 Solutions

This document contains solutions to homework problems involving nonlinear dynamical systems and differential equations. It includes: 1) Determining the flow of a nonlinear system and showing an invariant set. 2) Analyzing equilibria and eigenvalues of the Lorenz system. 3) Showing a homeomorphism between a nonlinear system and its linearization. 4) Calculating successive approximations to find stable and unstable manifolds.

Uploaded by

Rodrigo Kosta
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

Math 677. Fall 2009.

Homework #4 Solutions.
Part I. Exercises are taken from ”Diffential Equations and Dynamical Sys-
tems” by Perko, 3rd edition.

Problem Set 2.5: # 5


Determine the flow of the nonlinear system ẋ = f (x), f (x) = (−x1 , 2x2 +x21 )T
1
and show that the set S = {x2 = − x21 } is invariant with respect to the flow.
4
Solution to the IVP:
x1 (t) = c1 e−t
1 1
x2 (t) = (c2 + c21 )e2t − c21 e−2t
4 4
where x(0) = (c1 , c2 )T . The flow is defined by

x1 e−t
" #
φ(x, t) = 1 1
(x2 + x21 )e2t − x21 e−2t
4 4
1
To check invariance, start with y ∈ S, i.e. y2 = − y12 . Then φ(y, t) =
4
1 2 −2t
(y1 e , − y1 e ), so that φ(y, t) ∈ S as well.
−t
4
Problem Set 2.6: # 2
For the Lorenz equation ẋ = f (x), f (x) = (x2 −x1 , µx1 −x2 −x1 x3 , x1 x2 −x3 )T ,
µ > 0, the equilibrium point satisfies

~x0 = (a, a, a2 ), whereµa − a − a3 = 0

Obviously, x0 = 0 is one of the equilibria. Jacobian at the origin has eigen-


√
values λ1 = −1, λ2,3 = −1 ± µ, which is a sink when µ < 1 and a saddle
when µ > 1.
Other equilibria satisfy µ − 1 = a2 . It means that when µ > 1, there
2
are
√ two more points bifurcating from the origin at x0 = (a, a, a ) with a =
± µ − 1.

1
Jacobian here can be computed as
   
−1 1 0 −1 1 0
Df =  µ − x3 −1 −x1  , i.e.Df (x0 ) =  1 −1 −a 
x2 x1 −1 a a −1
1 p
It follows that λ1 = −2, λ2,3 = (−1 ± 5 − 4µ) are the eigenvalues. Here
2
if 1 < µ ≤ 5/4, we have λ2,3 < 0 i.e. x0 is a sink. Likewise for µ > 5/4, we
have a focus with negative real part of λ2,3 , which also can be classified as a
sink.
Hence µ = 1 is the bifurcation point. When µ < 1, there is a sink at the
origin. When µ = 1, the sink becomes √ degenerate and further for µ > 1 it
splits into sinks (a, a, a2 ) with a = ± µ − 1, forming a saddle point at the
location of the original equilibrium ((0, 0)).

Problem Set 2.6: # 3


1
For H(x) = (x1 , x2 + x21 , x3 + x21 )T , we compute H −1 (x) = (x1 , x2 − x21 , x3 −
3
1 2 T
x ) . Both maps are clearly continuous. Take the nonlinear system ẋ =
3 1
f (x) with f (x) = (−x1 , −x2 + x21 , x3 + x21 )T and let y = Hx. Then

ẋ1 −x1
   
2 
ẏ = Ḣ(x) =  ẋ2 + 2x1 ẋ1  =  −x2 − x1  = Df (0)y
  
2 1
ẋ3 + x1 ẋ1 x3 + x21
3 3
Hence H is a homeomorphism between the nonlinear system and its lineariza-
tion at the origin.

Problem set 2.7: #4.


Find the first 4 successive approximations for the system

ẋ1 = −x1
ẋ2 = −x2 + x21
ẋ3 = x3 + x22

The system has a hyperbolic saddle at the origin, since Df (0) = diag{−1, −1, 1}.
To find successive approximations, we do not need to translate anything, or
diagonalize the matrix A = Df (0), so A = B, F = G, C = I.

2
To find the stable manifold, we fix a = (a1 , a2 , 0) and we are going to look
for x3 = ψ(x1 , x2 ). We have U = diag{e−t , e−t , 0} and V = diag{0, 0, et },
F = G = (0, x21 , x22 )T .
Applying iterative formula
Z t Z ∞
(j+1) (j)
u = U (t)a + U (t − s)G(u (s, a))ds − V (t − s)G(u(j) (s, a))ds
0 t

with u(0) (t, a) = 0, we obtain:

u(1) (t, a) = (a1 e−t , a2 e−t , 0)T


u(2) (t, a) = (a1 e−t , (a2 + a21 )e−t − a21 e−2t , − 13 a22 e−2t )T
u(3) (t, a) = (a1 e−t , (a2 + a21 )e−t − a21 e−2t , − 13 (a2 + a21 )2 e−2t + 21 a21 (a2 + a21 )e−3t − 15 a41 e−4t )T

Since the first two components of the solution stabilize, u(3) = u(4) = . . ..
Hence we obtain
1 1 1
S = ψ3 (x1 , x2 ) = {x3 + x22 + x2 x21 − x41 }
3 6 30
To calculate the unstable manifold U , we put a = (0, 0, a3 ), replace t by
−t and express x1 = ψ1 (x3 ), x2 = ψ2 (x3 ). This yields u(t, a) = u(1) (t, a) =
(a3 e−t , 0, 0)T . It follows that a3 = 0, so that U = {x1 = x2 = 0}.

Part II. (c)


The system ẋ1 = −x31 , ẋ2 = −x2 has equilibrium at the origin, with E s = {y-
axis} and E c = {x−axis}. Center manifold should be tangent to the x-axis
then. Dividing one equation by the other, we get
dx2 x2
= 3
dx1 x1
2
which results in the solution x2 = Ce−1/(2x ) . Clearly, patching together
different branches of this solution leads to different center manifolds, all of
which are tangent to each other and to E c at the origin.
 −1/(2x2 )
 ae , x>0
Ma,b = 0, x=0
 −1/(2x2 )
be , x<0

defines a two-parameter continuum family of center manifolds.

You might also like