Using eigenvalues and eigenvectors to solve system of ODEs

In summary, to find the general solution of the system of ODEs, you can use eigenvalues and eigenvectors. First, convert the system into a matrix and find the characteristic polynomial by using the equation 0=|A-λI|. Then, solve for the eigenvalues by setting the characteristic polynomial equal to 0. Next, find the corresponding eigenvectors by plugging in each eigenvalue into the equation (A-λI)x=0. Finally, verify that the eigenvectors are correct by multiplying them with the original matrix and seeing if the result is a multiple of the eigenvector.
  • #1
joseph_18_18
14
0

Homework Statement


Use eigenvalues and eigenvectors to find the general solution of the system of ODEs..

x1 = 3x1 - x2
x2 = -x1 + 2x2 - x3
x3 = -x2 + 3x3

Homework Equations





The Attempt at a Solution


I converted that into the matrix
(3-λ -1 0)
(-1 2-λ -1) using 0=[itex]|A-λI[/itex][itex]|[/itex]
(0 -1 3-λ)

Sorry, I am new and don't know how else to write out the matrix... but I hope you get the gist..


I then solved through to get (3-λ)(λ-3)(λ-2)

λ=2 got me an eigenvector (5,1,5)T

λ=3 got me an eigenvecotr (6,1,6)T


Not really sure if I have done this correctly, and if it's correct I'm not sure how I would get this into a general solution eλt(eigenvector)T seeing as there is a repeated root.


Thanks
 
Physics news on Phys.org
  • #2
joseph_18_18 said:

Homework Statement


Use eigenvalues and eigenvectors to find the general solution of the system of ODEs..

x1 = 3x1 - x2
x2 = -x1 + 2x2 - x3
x3 = -x2 + 3x3
You're missing the primes in the column vector at the left. These are supposed to be derivatives.
joseph_18_18 said:

Homework Equations





The Attempt at a Solution


I converted that into the matrix
(3-λ -1 0)
(-1 2-λ -1) using 0=[itex]|A-λI[/itex][itex]|[/itex]
(0 -1 3-λ)

Sorry, I am new and don't know how else to write out the matrix... but I hope you get the gist..


I then solved through to get (3-λ)(λ-3)(λ-2)
I'm pretty sure you made a mistake. I get a different characteristic polynomial with three distinct roots. Show us what you did to get the char. polynomial.
joseph_18_18 said:
λ=2 got me an eigenvector (5,1,5)T
When you reach the point that you have found an eigenvector, a sanity test is a good idea. For your matrix (call it A) and the eigenvector you found (call it x), verify that Ax = 2x.
joseph_18_18 said:
λ=3 got me an eigenvecotr (6,1,6)T


Not really sure if I have done this correctly, and if it's correct I'm not sure how I would get this into a general solution eλt(eigenvector)T seeing as there is a repeated root.


Thanks
 
  • #3
Yeah, sorry they were meant to have primes on them, my bad.

Okay, from

(3-λ -1 0)
(-1 2-λ -1)
(0 -1 3-λ)

I did Column 1 - C3 to get

(3-λ -1 0)
(0 2-λ -1)
(-(3-λ) -1 3-λ)

I then took out (3-λ) as it's a common factor of C1 to get: (This is the step i wasn't too confident on)

(3-λ) (1 -1 0)
(0 2-λ -1)
(-1 -1 3-λ)

I then did C2 - C1:

(3-λ) (1 0 0)
(0 2-λ -1)
(-1 0 3-λ)

Sorry if this is quite hard to read! The second brackets is the matrix and the (3-λ) is the common factor i have extracted.

From here i simplified to:

(3-λ) (2-λ -1)
( 0 3-λ) which gave this answer:

(3-λ)(6 - 5λ + λ2)

=

(3-λ)(λ-3)(λ-2) = 0

therefore λ = 3,3,2

I think you can check to see if you're on the right step by seeing if the product of the eigenvalues is the same as the determinant of the original matrix, and they both come out to be 18... But I'm not too sure if this is correct.


''When you reach the point that you have found an eigenvector, a sanity test is a good idea. For your matrix (call it A) and the eigenvector you found (call it x), verify that Ax = 2x.''

For Ax i ended up with (14,8,14)T and 2x was (10,2,10)T, i presume this means I have made a mistake somewhere..

Thanks Mark.
 
  • #4
joseph_18_18 said:
Yeah, sorry they were meant to have primes on them, my bad.

Okay, from

(3-λ -1 0)
(-1 2-λ -1)
(0 -1 3-λ)
I expanded the determinant across the first row to get
(3-λ)[(2 - λ)(3 - λ) - 1] + 1(λ - 3 - 0). I am confident of this work.
joseph_18_18 said:
I did Column 1 - C3 to get

(3-λ -1 0)
(0 2-λ -1)
(-(3-λ) -1 3-λ)

I then took out (3-λ) as it's a common factor of C1 to get: (This is the step i wasn't too confident on)

(3-λ) (1 -1 0)
(0 2-λ -1)
(-1 -1 3-λ)

I then did C2 - C1:

(3-λ) (1 0 0)
(0 2-λ -1)
(-1 0 3-λ)

Sorry if this is quite hard to read! The second brackets is the matrix and the (3-λ) is the common factor i have extracted.

From here i simplified to:

(3-λ) (2-λ -1)
( 0 3-λ) which gave this answer:

(3-λ)(6 - 5λ + λ2)

=

(3-λ)(λ-3)(λ-2) = 0

therefore λ = 3,3,2

I think you can check to see if you're on the right step by seeing if the product of the eigenvalues is the same as the determinant of the original matrix, and they both come out to be 18... But I'm not too sure if this is correct.




For Ax i ended up with (14,8,14)T and 2x was (10,2,10)T, i presume this means I have made a mistake somewhere..

Thanks Mark.
 
  • #5
I expanded the determinant across the first row to get
(3-λ)[(2 - λ)(3 - λ) - 1] + 1(λ - 3 - 0). I am confident of this work.

Does this mean I have the correct roots?

If so, how should I approach the eigenvectors with a repeated root?

Thanks Mark
 
  • #6
No, you don't have the correct roots. There are three distinct roots.

What I had in my post was part of the work needed to find the roots. You can start from where I left off to find these roots.
 
  • #7
multiplying out gave me

λ3 + 2λ2 -19λ -12

which factorised (using grouping) to:

(λ+2)(λ2-19)


Am I doing it right?

Thanks again.
 
  • #8
joseph_18_18 said:
multiplying out gave me

λ3 + 2λ2 -19λ -12

which factorised (using grouping) to:

(λ+2)(λ2-19)


Am I doing it right?

Thanks again.
No. Show me how you got this: λ3 + 2λ2 -19λ -12

The 3 roots are positive integers.
 
  • #9
I multiplied out the first brackets to get:

(3-λ)(5-5λ-λ2)+(-3) + λ

=

15-15λ-3λ2-5λ+5λ23-3+λ

=

λ3+2λ2-19λ-12...

Thanks
 
  • #10
joseph_18_18 said:
I multiplied out the first brackets to get:

(3-λ)(5-5λ-λ2)+(-3) + λ
That's still not quite right.
Starting from here -- (3-λ)[(2 - λ)(3 - λ) - 1] + 1(λ - 3 - 0) --
you should have (3-λ)[5 -5λ + λ2] + λ - 3.

Instead of multiplying things out, it's quicker to notice that 3 - λ is a factor in the first product and the second expression. This gives you
(3-λ)[5 -5λ + λ2 - 1]
Can you work it out from here?
joseph_18_18 said:
=

15-15λ-3λ2-5λ+5λ23-3+λ

=

λ3+2λ2-19λ-12...

Thanks
 
  • #11
joseph_18_18 said:
I multiplied out the first brackets to get:

(3-λ)(5-5λ-λ2)+(-3) + λ
There's a sign error in the line above. You didn't multiply (2 - λ)(3 - λ) correctly.
joseph_18_18 said:
=

15-15λ-3λ2-5λ+5λ23-3+λ

=

λ3+2λ2-19λ-12...

Thanks
 
Last edited:

1. What are eigenvalues and eigenvectors?

Eigenvalues and eigenvectors are mathematical concepts used to describe the behavior of linear transformations. Eigenvalues represent the scaling factor of the eigenvectors, which are vectors that do not change direction when multiplied by a transformation matrix.

2. How can eigenvalues and eigenvectors be used to solve systems of ordinary differential equations (ODEs)?

Eigenvalues and eigenvectors can be used to solve systems of ODEs by transforming the system into a diagonal form, where each equation is decoupled from the others. This makes it easier to solve the system by finding the eigenvalues and eigenvectors of the transformation matrix.

3. What is the significance of the eigenvalues and eigenvectors in solving ODEs?

The eigenvalues and eigenvectors provide important information about the stability and behavior of the system. The eigenvalues determine the rate of change of each variable, while the eigenvectors represent the direction of change. This allows us to make predictions about the long-term behavior of the system.

4. Can eigenvalues and eigenvectors be used to solve any type of ODE system?

Yes, eigenvalues and eigenvectors can be used to solve any linear homogeneous system of ODEs. However, they may not always provide the most efficient or accurate solution, and other methods may be more suitable for certain types of systems.

5. Are there any limitations to using eigenvalues and eigenvectors to solve ODE systems?

One limitation is that eigenvalues and eigenvectors can only be used for linear systems. Nonlinear systems require different methods for solving. Additionally, the complexity and size of the system may make it difficult to find the eigenvalues and eigenvectors, and other methods may be more practical in these cases.

Similar threads

  • Calculus and Beyond Homework Help
Replies
5
Views
524
  • Calculus and Beyond Homework Help
Replies
2
Views
334
  • Calculus and Beyond Homework Help
Replies
2
Views
521
  • Calculus and Beyond Homework Help
Replies
7
Views
1K
  • Linear and Abstract Algebra
Replies
12
Views
1K
  • Calculus and Beyond Homework Help
Replies
5
Views
1K
  • Calculus and Beyond Homework Help
Replies
12
Views
1K
  • Calculus and Beyond Homework Help
Replies
1
Views
1K
  • Calculus and Beyond Homework Help
Replies
3
Views
1K
  • Calculus and Beyond Homework Help
Replies
3
Views
1K
Back
Top