Vous êtes sur la page 1sur 4

How to Eigenvector - Abridged

Nate
October 4, 2018

1 What is an eigenvector/what are eigenval-


ues?
Eigenvectors are relationships between the variables of a system of linear
equations that are invariant under transformation. Basically, no matter what
matrix operations you perform the eigenvectors will stay the same.

Eigenvalues are the adjustments made to the matrix to find the eigenvectors.

If the rows/columns of a matrix are linearly independent we can’t say any-


thing about their relationships (they’re independent for general values). So,
we have to introduce something (the eigenvalues) to find the conditions under
which they are related. A matrix has linearly dependent rows/columns if its
determinant is zero, so we make that happen by applying the same constant
to each variable:

det(A − λI) = 0 where λ is an eigenvalue; multiplying it it by I applies it


once to each variable as required.

Random Example
 
1 2
is linearly independent (det = -3)
2 1

1
     
1 2 λ 0 1−λ 2
− =
2 1 0 λ 2 1−λ
 
1−λ 2
det = (1 − λ)2 − 4 = 0
2 1−λ
λ = 3, −1
  
−2 2 x1
For λ = 3 we have =0
2 −2 x2
Pick a row, any row

−2x1 + 2x2 = 0 → x1 = x2
 
1
This gives us the eigenvector associated with λ = 3, which is
1
 
1
Similarly λ = −1 gives
−1
No matter how you adjust the original matrix (row algebra, constant multi-
plication, what have you) these relationships remain the same. Of course this
isn’t particularly exciting without the physics, so let’s throw in that spring
problem.

2 That spring problem

Shamelessly ripped straight from the homework.


All the F = ma boils down to the following:

2
−4kx1 + 3kx2 = mx¨1
3kx1 − 4kx2 = mx¨2
q
k
For simple harmonic motion we define ω0 = m , x = aeiωt which makes
ẍ = −ω 2 aeiωt and λ = ( ωω0 )2 . Yes, that’s the eigenvalue lambda, and you’ll
see why in just a second. If you’re curious about the rest just look up a
refresher on SHM.

Arranging things a little...

−4ω02 a1 eiωt + 3ω02 a2 eiωt = −ω 2 a1 eiωt

3ω02 a1 eiωt − 4ω02 a2 eiωt = −ω 2 a2 eiωt

Cleaning it up and dividing everything by ω02 ...

−4a1 + a2 = −λa1

3a1 − 4a2 = −λa2

Yep, now we can solve for λ and find our eigenvalues. Before we do, what
the heck do all these things mean?

Well, in SHM ω is the angular frequency, where ω0 is the fundamental


frequency of the system. Our eigenvalues will be the values for λ where
λ = ( ωω0 )2 gives us the maximum amplitude a. Remember resonant frequen-
cies? That’s what these are.
 
−4 + λ 3
det =0
3 −4 + λ
λ = 1, 7

This √
tells us first that the resonant frequencies of the system are ω = ω0 and
ω = 7ω0

We can then use the eigenvectors to find what kind of motion is happening
when each of these occurs.

3
Let’s just plug our eigenvalues into the first row.

(−4 + 1)x¨1 + 3x¨2 = 0

x¨1 = x¨2
 
1
Our eigenvector implies that the masses are moving the same amount
1
in the same direction for res. frequency ω0 .

For the other,

(−4 + 7)x¨1 + 3x¨2 = 0

x¨1 = −x¨2
 
1
Our eigenvector implies that the masses are moving the same amount
−1

in opposite directions for res. frequency 7ω0 .

Hope that helps, man!

Vous aimerez peut-être aussi