Matrix Inversion for Variation of Parameters

5 replies · 4K views · last reply
gkirkland
Messages
11
Reaction score
0
I am working on the following problem:

Handwritten derivation of a 2x2 linear system x' = Ax + F(t): eigenvalues found from det(A - λI) = 0 giving λ = -1 and -2, eigenvectors, fundamental matrix Φ(t) with entries e^(-t), e^(-2t), -e^(-t), -2e^(-2t), and the variation of parameters formula for the particular solution.


Can someone please show or explain the steps to invert the phi matrix? I've given it a few tries, but I can't reach what the book has for the answer.

Please help! Thanks
 
Phi(t)*Phi^-1(t) = I, where I is the 2x2 identity matrix.
You can use Cramer's rule to solve for the inverse by breaking the identity matrix into two columns.

If you post some of your work, we may help you spot what is going wrong with you calculations.
 
I didn't even think about using Cramer's rule to find the inverse! That was a great idea, got it on the first try.

Thanks!
 
So I calculated the inverted matrix, but for some reason I'm still not getting the correct solution. Can someone please spot my mistake?

Handwritten-style worked solution of a 2x2 linear ODE system by variation of parameters: eigenvalues -1 and -2 with eigenvectors, fundamental matrix Φ(t), its inverse, and the particular solution integral setup with forcing vector (0, e^t).

Handwritten variation of parameters working: X_p equals a 2x2 matrix of e^(-t), e^(-2t), -e^(-t), -2e^(-2t) times the integral of a column vector of -e^(2t) and 2e^(2t) dt; the next line shows the integrated column vector, and the final result column vectors are marked INCORRECT with the correct answer shown beside them.
 
I apologize for my lack of clarity in my earlier post.

In order to find the inverse of the Phi(t) matrix using Cramer's Rule, it must be used twice.

Code:
This is how the inverse is set up:
| e^-t    e^-2t|  |a11  a12|     |1  0|
|-e^-t  -2e^-2t|  |a21  a22|  =  |0  1|

where the a values represent the inverse of Phi(t). In order to calculate the a values,
the matrix equation above is split into two parts, like so:
Code:
This is how the inverse calculations are set up:
| e^-t   e^-2t|  |a11|     |1|
|-e^-t -2e^-2t|  |a21|  =  |0|
and
| e^-t   e^-2t|  |a12|     |0|
|-e^-t -2e^-2t|  |a22|  =  |1|

Cramer's Rule can be used to solve both systems. Remember, always check the calculations by multiplying Phi(t) by Phi^-1(t) to see if the Identity matrix is obtained.

I should point out that Phi^-1(t) is the inverse matrix of Phi(t) rather than 1/Phi(t)
 
isn't the inverse of the matrix 1/det(phi) * adj(phi) not just phi?
You have to find the adjoint of phi then multiply by 1/det(phi), the adjoint matrix being the transpose of the C expansion matrix, for a 2x2 it always works out to be that you swap the signs on entry 2 and 3 and swap the placement of entry 1 and 4.

but I am not sure, trying to work it out now also not getting the books answer..