How Do You Solve a System of Second Order ODEs with Matrix Methods?

Click For Summary

Homework Help Overview

The discussion revolves around solving a system of second-order ordinary differential equations (ODEs) using matrix methods. The equations involve a matrix A and require transforming the system into a first-order format.

Discussion Character

  • Exploratory, Mathematical reasoning, Assumption checking

Approaches and Questions Raised

  • Participants discuss reducing the second-order system to a first-order system and the implications of eigenvalues and eigenvectors on the general solution. There is also a mention of verifying the solution by substituting it back into the original ODEs.

Discussion Status

Some participants have provided insights on the nature of the solutions, particularly regarding repeated roots and the inclusion of constants in the general solution. There is an acknowledgment of the need for boundary conditions, but no explicit consensus has been reached on the correctness of the proposed solutions.

Contextual Notes

Participants note the presence of four boundary conditions, which may influence the form of the solution. There is also a reference to the solution of a simpler ODE, suggesting that assumptions about the nature of the roots are being examined.

c0der
Messages
52
Reaction score
0

Homework Statement



Solve:

[ d^2y1/dx^2 ] = [ a -a ] [ y1 ]
[ d^2y2/dx^2 ] [ -a a ] [ y2 ]

A = [ a -a ]
[ -a a ]

Homework Equations



Everything required is in (1) above

The Attempt at a Solution



Reduce to 1st order system

M = [ 0 I ]
[ A 0 ]

Hence, M =
[ 0 0 1 0 ]
[ 0 0 0 1 ]
[ a -a 0 0 ]
[ -a a 0 0 ]

The eigenvalues of M are 0, 0, √2a and -√2a
The eigenvectors are [ 1 1 ], [ 1 1 ], [ 1 -1] and [ 1 -1]

Hence the general solution is (for y only):

y(x) = A*e^(0x)*[ 1 1 ]T + B*e^(0x)*[ 1 1 ]T + C*e^(√2a*x) [ 1 -1 ]T +
D*e^(-√2a*x) [ 1 -1 ]T

So:

y1 = A + B + C*cosh(√(2a)*x) + D*sinh(√(2a)*x)
y2 = A + B - C*cosh(√(2a)*x) - D*sinh(√(2a)*x)

Is this correct?
 
Physics news on Phys.org
The way to check if that's correct is to plug your answer back into the ODE system.
 
c0der said:

Homework Statement



Solve:

[ d^2y1/dx^2 ] = [ a -a ] [ y1 ]
[ d^2y2/dx^2 ] [ -a a ] [ y2 ]

A = [ a -a ]
[ -a a ]

Homework Equations



Everything required is in (1) above

The Attempt at a Solution



Reduce to 1st order system

M = [ 0 I ]
[ A 0 ]

Hence, M =
[ 0 0 1 0 ]
[ 0 0 0 1 ]
[ a -a 0 0 ]
[ -a a 0 0 ]

The eigenvalues of M are 0, 0, √2a and -√2a
The eigenvectors are [ 1 1 ], [ 1 1 ], [ 1 -1] and [ 1 -1]

Hence the general solution is (for y only):

y(x) = A*e^(0x)*[ 1 1 ]T + B*e^(0x)*[ 1 1 ]T + C*e^(√2a*x) [ 1 -1 ]T +
D*e^(-√2a*x) [ 1 -1 ]T

Do you not recall that the solution of u'' = 0 is u(x) = Ax + B?


So:

y1 = A + B + C*cosh(√(2a)*x) + D*sinh(√(2a)*x)
y2 = A + B - C*cosh(√(2a)*x) - D*sinh(√(2a)*x)

Is this correct?
 
Yes, ok. I see you mean because I have repeated roots, it's Ae^0(x) + Bxe^(0x). Even the above without A and B constants is a solution to the system. However, since I have 4 boundary conditions, I thought I'd keep them in there
 

Similar threads

Replies
4
Views
2K
Replies
2
Views
1K
  • · Replies 18 ·
Replies
18
Views
3K
  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 16 ·
Replies
16
Views
2K
  • · Replies 5 ·
Replies
5
Views
1K
  • · Replies 7 ·
Replies
7
Views
2K
  • · Replies 3 ·
Replies
3
Views
2K
Replies
33
Views
3K
  • · Replies 6 ·
Replies
6
Views
1K