Scribd
Upload
64 views5 pages

1 Lecture 12 - 23/9: Henrik Jacobsson Notes From ODE-lectures September 26, 2011

1. The document summarizes notes from an ODE lecture on transforming systems of differential equations into canonical form through coordinate transformations. Key steps include finding eigenvalues and eigenvectors of the system matrix and using them to construct the transformation matrix T. 2. One example transforms a system with real eigenvalues into decoupled form, allowing the general solution and phase portrait to be described. Another example transforms a system with complex eigenvalues into canonical form. 3. The trace-determinant plane is introduced to classify systems based on the relationship between the trace and determinant of the system matrix, relating this to whether the eigenvalues are real/complex or repeated.

Uploaded by

jaikobi
Copyright
© Attribution Non-Commercial (BY-NC)
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
64 views5 pages

1 Lecture 12 - 23/9: Henrik Jacobsson Notes From ODE-lectures September 26, 2011

1. The document summarizes notes from an ODE lecture on transforming systems of differential equations into canonical form through coordinate transformations. Key steps include finding eigenvalues and eigenvectors of the system matrix and using them to construct the transformation matrix T. 2. One example transforms a system with real eigenvalues into decoupled form, allowing the general solution and phase portrait to be described. Another example transforms a system with complex eigenvalues into canonical form. 3. The trace-determinant plane is introduced to classify systems based on the relationship between the trace and determinant of the system matrix, relating this to whether the eigenvalues are real/complex or repeated.

Uploaded by

jaikobi
Copyright
© Attribution Non-Commercial (BY-NC)
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/ 5

Henrik Jacobsson Notes from ODE-lectures September 26, 2011

Lecture 12 - 23/9

1 0 X 1 2 such that the new system has a matrix on canonical form. Compare the phase portraits of the two systems. Need to nd a matrix T 2x2 such that the coordinate transformation X = T Y results in a system with system matrix T 1AT which is on canonical form. We have seen that if T = [V1 V2] the Example: Change the coordinates of the system X =
eigenvectors

1 0 So we need to nd the eigenvectors and eigen02 values. Eigenvalues are obtained from the characteristic equation. det(A I) = 0 1 0 0 det =0 1 2 0 1 0 det = 0 (1 )(2 ) 1 0 = 1 2 0 ( + 1)( + 2) = 0 1 = 1, 2 = 2 Eigenvector corresponding to 1 = 1 (Av = v) 1 0 v1 v = 1 1 1 2 v2 v2 First row: v1 = v1 1 Second row: v1 2v2 = v2 v1 = v2 for example 1 Eigenvector corresponding to 2 = 2 1 0 v1 v = 2 1 v2 1 2 v2 T 1AT =
1

First row: v1 = 2v1 v1 = 0 Second row: v1 2v2 = 2v2 v1 = 0 So the transformation matrix becomes 1 0 1 0 T = T 1 = 1 1 1 1
eigenvectors 1 & 2

0 1

Hence, we make the change of coordinates 1 0 X Y = 1 1


T 1

Then the new system becomes Y =

1 0

0 Y 2

eigenvalue 1 & 2

1 0 +c2e2t 0 1 The dynamics along the two axis are decoupled. The dynamics along 0 1 the direction is quicker than in the direction . So the gen1 0 eral solution of the original system is: 1 0 1 0 1 X(t) = T Y (t) = c1et + c2e2t = c1et + 1 0 0 1 0 0 c2e2t 1 What does the phase portrait look like? 0 limt X(t) = 0 General solution of the canonical system Y (t) = c1et
dx2 dx1

dx limt dx2 = 1 1

dx2 dt dx1 dt

c1 et 2c2 e2t c1 et

= 1 + 2 c2 e 2 c1

limt dx2 = dependent on signs of c1 and c2 dx1


1.1 Transformation of a system with complex eigenvalues

Consider X = AX where A has complex eigenvalues i ( = 0) with V1 and V2 are real vectors such that V1 iV2 are eigenvectors corresponding to i. Since V1 + iV2 eigenvectors and + i the corresponding eigenvalue. A (V1 + iV2) = ( + i) (V1 + iV2)
V V

Equate real and imaginary parts: AV1 = V1 V2 AV2 = V1 + V2 How dene transformation matrix T such that T 1AT is on canonical form? T = [V1 V2] hence T E1 = V1 T E2 = V2
unit vector

So (T 1AT )E1 = T 1AV1 = T 1(V1 V2) = E1 +E2 = which is on canonical form. Example: (Complex eigenvalues) 0 1 X Consider the system X = 4 0 Thus T 1AT =
A

Find a canonical form and describe the phase portraits of both systems. Eigenvalues are given by 2 + 4 = 0 = 2i
det(AI)=0

A complex eigenvector corresponding to = 2i is


3

1 2i

So there is a complex solution. 1 e2it Breakitintorealandimaginaryparts, givesthegeneralsolution : 2i cos2t sin2t X(t) = c1 + c2 2sin2t 2cos2t 1 Break the eigenvector into real and imaginary parts: 2i 1 0 + i 0 2
V1 V2

Construct a T by setting T = [V1 V2] = Then T 1AT = formula] Y (t) = c1 cos2t sin2t + c2 sin2t cos2t 0 2 2 0

1 0

0 2

so the solution on canonical form [use

Y = T 1AT Y X = AX looks the same but with ellipses instead of circles, more tall than they are wide.

1.2

The trace-determinant plane

For a matrix A =

a b c d (a + d) + (ad bc) = 0
det(A)=D

the eigenvalues are given by 2

trace(A)=T

So the characteristic equation is 2T + D = 0 and the eigenvalues are given by = T 2 4D) (1) Note: + + = T and + = D So if we know T and D we can nd + and (1) A ) T 2 4D < 0 Complex eigenvalues B ) T 2 4D > 0 Real distinct eigenvalues C ) T 2 4D = 0 Repeated eigenvalues

1 2 (T

You might also like