How do you solve differential equations using matrices?

Click For Summary
The discussion focuses on solving differential equations using matrices, specifically the equation 2(d^2y/dx^2) - 5(dy/dx) + 2y = 0. The technique involves rewriting the equation in operator form and factoring it to find solutions. By using a vector formulation, the original ODE is transformed into a system of equations represented by a matrix. Eigenvalues and eigenvectors are then determined to construct the general solution of the differential equation. The final solution is expressed as a linear combination of exponential functions derived from the eigenvalues.
gruffins
Messages
1
Reaction score
0
What is the technique to solve differential equations using matrices?

I think it would help if an example were solved:

2(d^2y/dx^2)-5(dy/dx)+2y = 0

I don't even know how to start
 
Physics news on Phys.org
Welcome to PF!

Hi gruffins! Welcome to PF! :smile:

(try using the X2 tag just above the Reply box :wink:)
gruffins said:
What is the technique to solve differential equations using matrices?

I think it would help if an example were solved:

2(d^2y/dx^2)-5(dy/dx)+2y = 0

I don't even know how to start


Matrices I don't know :redface:

but if you mean using operators,

you write it (2D2 - 5D + 2)y = 0,

and then you factor that into (D - a)(D - b)y = 0,

so the solutions are linear combinations of the solutions to Dy = ay or Dy = by :smile:
 
It's kind of a neat technique I think. Since this is a pretty simple ODE I will use the operator form for solving it as well.

You have 2y'' - 5y' + 2y = 0. We seek forms of ex in the solutions.

Using operator form we can rewrite this as: (2D2 - 5D +2)y = 0
with auxiliary polynomial 2r2 - 5r +2 = 0
containing roots r = 1/2 and 2
and linearly independent solutions e2x and ex/2.
Therefore by the superposition principle we can say: y(x) = c1e2x + c2ex/2 where c1 and c2 are our constants.

Now if we want to rewrite this as a vector formulation, in our original ODE we can let z1 = y and z2 = y'
so our ODE becomes 2z2' - 5z2 + 2z1 = 0.

Now we can write a system:

z1' = z2
z2' = -z1 + (5/2)z2

And letting a vector z = (z1, z2) and a matrix A = [0, 1; -1, 5/2]
we have the vector differential equation: z' = Az. (Note: in A, the 0 and 1 are the first row and -1 and 5/2 are the second row. I am new to this forum and haven't had time to see if the forum has tex or anything.)

We must now examine A and find its eigenvalues, which happen to be: 1/2 and 2. (When you find the characteristic polynomial of A to find the eigenvalues, notice how similar it is to the auxiliary equation in solving the ODE in operator form.)

Using the eigenvalue of r1 = 2, we have an eigenvector v1 = (1, 2) and since we require solutions in the form of ex our first solution is er1xv1 = e2x(1, 2).

With eigenvalue r2 = 1/2, an eigenvector is v2 = (1, 1/2) and our second linearly independent solution is ex/2(1, 1/2).

And our solution to our vector DE is z(x) = c1e2x(1, 2) + c2ex/2(1, 1/2).

And since z = (z1, z2) and z1 = y, we only need the first component of this vector solution. The second component is a solution to y' since z2 = y'.

Therefore y(x) = z1 = c1e2x + c2ex/2.

Few comments:
The solution to the vector differential equation can vary depending on how you pick your eigenvectors.
Also, extra steps need to be taken if the separate solutions are not linearly independent (i.e. same eigenvectors, etc..), which goes into cycles of generalized eigenvectors.
 
Last edited:
Welcome to PF!

Hi pbandjay! Welcome to PF! :smile:

Just two comments …

i] for LaTeX, type [noparse]before and after[/noparse] (or use the ∑ tag just above the Reply box) :wink:

ii] "axillary" means "relating to the armpits"! :biggrin:
 


tiny-tim said:
ii] "axillary" means "relating to the armpits"! :biggrin:[/INDENT]

:eek: Oops heh.. thanks for noticing that. I think I just accidentally spelled auxiliary wrong and automatically clicked axillary in the spell check.
 
I am studying the mathematical formalism behind non-commutative geometry approach to quantum gravity. I was reading about Hopf algebras and their Drinfeld twist with a specific example of the Moyal-Weyl twist defined as F=exp(-iλ/2θ^(μν)∂_μ⊗∂_ν) where λ is a constant parametar and θ antisymmetric constant tensor. {∂_μ} is the basis of the tangent vector space over the underlying spacetime Now, from my understanding the enveloping algebra which appears in the definition of the Hopf algebra...

Similar threads

  • · Replies 7 ·
Replies
7
Views
4K
  • · Replies 2 ·
Replies
2
Views
1K
  • · Replies 7 ·
Replies
7
Views
4K
  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 3 ·
Replies
3
Views
1K
  • · Replies 19 ·
Replies
19
Views
4K
  • · Replies 8 ·
Replies
8
Views
1K
  • · Replies 6 ·
Replies
6
Views
2K
  • · Replies 5 ·
Replies
5
Views
4K
  • · Replies 10 ·
Replies
10
Views
2K