Scribd
Upload
27 views

Linear, First-Order, Non-Homogeneous Systems

This document summarizes the derivation of the variation of parameters method for solving non-homogeneous linear systems of first-order differential equations. It shows that the solution involves finding the fundamental matrix Φ(t) of the associated homogeneous system, and using its inverse Φ-1(t) as an integrating factor. Differentiating Φ-1(t)Φ(t)=I gives an expression for the derivative of Φ-1(t) involving the system matrix P(t). This allows rewriting the original system in an integrated form involving an integral of Φ-1(t) times the non-homogeneous term f(t), showing that variation of parameters and the method of integrating factors are the same

Uploaded by

Cris Jackson
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
27 views

Linear, First-Order, Non-Homogeneous Systems

This document summarizes the derivation of the variation of parameters method for solving non-homogeneous linear systems of first-order differential equations. It shows that the solution involves finding the fundamental matrix Φ(t) of the associated homogeneous system, and using its inverse Φ-1(t) as an integrating factor. Differentiating Φ-1(t)Φ(t)=I gives an expression for the derivative of Φ-1(t) involving the system matrix P(t). This allows rewriting the original system in an integrated form involving an integral of Φ-1(t) times the non-homogeneous term f(t), showing that variation of parameters and the method of integrating factors are the same

Uploaded by

Cris Jackson
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

18.

03 at ESG

Independent Study

Fall 2003

Linear, First-Order, Non-homogeneous Systems


To try to keep things simple (heh-heh), Ill use boldface caps (e.g., A or P(t))
for matrices, overarrows for vectors, and time dierentiation will be shown explicitly.
The independent variable will be t, with the possibility of t as an integration
variable.
Anyway, lets start with
d
x = P(t) x + f(t)
dt

(A)

and the associated homogenous equation


d
x = P(t) x.
dt

(B)

At this point, we need to assume that we have found enough independent solutions to (B) to form a non-singular (i.e., the determinant, which is the Wronskian,
does not vanish. Ever.) square matrix (t). We have seen, or rather I insisted, and
it can be shown, that
d
(t) = P(t) (t).
(C)
dt
We use this to show a really neat thing. We know that (t) has an inverse (why?)
and so (t)1 (t) = 1 (t)(t) = I. Dierentiating and having some fun,


(t)1 (t) = I
d
d  1 
(t) 1 (t) + (t)
(t) = 0
dt
dt
d  1 
(t) = 0
P(t) + (t)
dt


d  1 
(t)
dt
d  1 
(t) ,
1 (t)P(t) =
dt
P(t) = (t)

where P(t) =

(r.n.t.)

 1
(t)
(t) has been snuck in while you may have been napping.
dt

d

The r.n.t. is perhaps not so strange. Recall that in Chapter 1, the integrating
d
= P (x)(x). Well, you can show that if this
factors were those (x) such that dx
is the case,

d 1
dx

= P (x)/, which is similar to the above.


1

Now that we have the r.n.t. at our disposal, lets mangle (A);
d
x P(t)x = f(t)
dt
d
x 1 (t)P(t)x = 1 (t)f(t)
dt

d  1 
d
1
(t) x = 1 (t)f(t).
(t) x +
dt
dt
1 (t)

Now we use the handy-dandy product rule for matrices (and hence vectors),
d
(M(t)x) =
dt
to get


d
d
M(t) x + M(t) x
dt
dt


d  1
(t)x = 1 (t)f.
dt

You can guess at the drill from here;

1 (t ) f(t ) dt and


a
 t
1 (t ) f(t ) dt .
x = (t)

(t)x =

Well, heres the punchline; The above result was presented in the boxed equation on Page 448 of E&P as the vector form of variation of parameters, while the
above method duplicated the method of integrating factors used for linear rst-order
equations in Chapter 1. But, as we have just shown,
THEY ARE THE SAME!
Still to come: In Chapter 1, the integrating factor was the exponential of
something. Does this mean that (t) is like the exponential of something? Funny
you should ask!

You might also like