Chapter 10

Systems of Linear
Differential Equations
Outline
10.1 Theory of Linear Systems
10.2 Homogenous Linear Systems
10.3 Solution by Diagonalization
10.4 Nonhomogenous Linear Systems
10.5 Matrix Exponential
Theory of Linear Systems
Matrix form of a linear system
If the system is homogenous,
The solution vector is


Example


'
= + X AX F
'
= X AX
3 4
5 7
dx
x y
dt
dy
x y
dt
= +
=
6 1 1
' 8 7 1 10
2 9 1 6
t
t
t
| | | |
| |
= +
| |
| |

\ . \ .
X X
( )
( )
( )
1
2
4
x t
x t
X
x t
| |
|
|
=
|
|
|
\ .
Homogenous Linear Systems
To find a nontrivial solution X of
we must find a nontrivial solution of

We must have
The solutions of the characteristic equation are
the eigenvalues
n
with corresponding
eigenvectors K
n



'
= X AX
( )
0 = A I K
( )
det 0 = A I
Homogenous Linear Systems (contd.)
Example: Solve
Find the eigenvalues and eigenvectors
From the characteristic equation


we see that and
For , and
Thus and when the related
eigenvector is


2 3
dx
x y
dt
= + 2
dy
x y
dt
= +
( ) ( )( )
2
2 3
det 3 4 1 4 0
2 1

= = = + =

A I
1
1 =
2
4 =
1
1 =
1 2
k k =
2
1 k =
1
1
1
| |
=
|

\ .
K
1 2
3 3 0 k k + =
1 2
2 2 0 k k + =
Homogenous Linear Systems (contd.)
Example (contd.):
For , and
Thus and when the related
eigenvector is

We have found 2 linearly independent solutions
and where the general
solution is




2
4 =
( )
1 2
3/ 2 k k =
2
2 k =
2
3
2
| |
=
|
\ .
K
1 2
2 3 0 k k + =
1 2
2 3 0 k k =
1
1
1
t
e

| |
=
|

\ .
X
4
2
3
2
t
e
| |
=
|
\ .
X
4
1 1 2 2 1 2
1 3
1 2
t t
c c c e c e

| | | |
= + = +
| |

\ . \ .
X X X
Homogenous Linear Systems (contd.)
If is a factor of the characteristic
equation while is not a factor,
is said to be an eigenvalue of multiplicity m
The solutions corresponding to complex
eigenvalues are and




1
1
t
e

K
1
1
t
e

K
( )
1
m

( )
1
1
m

+

Solution by Diagonalization
If A is diagonalizable, consider the following
alternative method for solution
A homogenous system




is said to be coupled if each is expressed as a
linear combination of





1 11 12 1 1
2 21 22 2 2
1 11
n
n
n n nn n
x a a a x
x a a a x
x a a a x
'
| | | || |
| | |
'
| | |
=
| | |
| | |
'
\ . \ .\ .
i
x
'
1 2
, , ,
n
x x x
Solution by Diagonalization (contd.)
If A is diagonalizable the system can be
uncoupled so each is expressed solely in terms
of
If A has n linearly independent eigenvectors, we
can find a matrix P such that where D
is a diagonal matrix






i
x
'
i
x
1
= P AP D
Solution by Diagonalization (contd.)
Making the substitution in ,

is the same as









'
= Y DY
1 1 1
2 2 2
0 0
0 0
0 0
n n n
y y
y y
y y

'
| | | || |
| | |
'
| | |
=
| | |
| | |
'
\ . \ .\ .
= X PY
'
= X AX
1
' ' '
= = = PY APY Y P APY Y DY
Solution by Diagonalization (contd.)
The general solution of can be written as




Since we now know Y and the matrix P can be
constructed from eigenvectors of A, the general
solution of is obtained from








= X PY '
= X AX
1
2
n
c e
y
y
| |
|
|
=
|
|
|
\ .
Y
'
Y
Nonhomogenous Linear Systems
Solution methods for nonhomogenous linear
systems are similar to those for nonhomogenous
linear ODEs
Undetermined coefficients
Sometimes quicker
Variation of parameter
More powerful









Nonhomogenous Linear Systems (contd.)
The method of undetermined coefficients
consists of making an educated guess about the
form of a particular solution vector X
p
Guess is motivated by types of functions that
comprise the column matrix F(t)
Applicable only to where the
entries of A are constants and the entries of F(t)
are constants, polynomials, exponential functions,
sines and cosines, or finite sums and products of
these functions









( )
t
'
= + X AX F
Nonhomogenous Linear Systems (contd.)
Variation of parameters
If is a fundamental set of solutions
of , its general solution is
or






11 12 1 1 11 2 12 1
21 22 2 1 21 2 22 2
1 2
1 2 1 1 2 2
n n n
n n n
n
n n nn n n n nn
x x x c x c x c x
x x x c x c x c x
c c c
x x x c x c x c x
+ + +
| | | | | | | |
| | | |
+ + +
| | | |
= + + + =
| | | |
| | | |
+ + +
\ . \ . \ . \ .
X
1 2
, , ,
n
X X X
'
= X AX
1 1 2 2 n n
c c c = + + + X X X X
Nonhomogenous Linear Systems (contd.)
Variation of parameters (contd.)
This solution can be written as
where C is the column vector of constants and u is
the fundamental matrix



u is nonsingular


( )
11 12 1
21 22 2
1 2
n
n
n n nn
x x x
x x x
t
x x x
| |
|
|
=
|
|
\ .

( )
t = X C
( ) ( )
t t
'
= A
Nonhomogenous Linear Systems (contd.)
Variation of parameters (contd.)
We ask whether it is possible to replace C by


so that
is a particular solution of

( )
( )
( )
1
2
n
u t
u t
u t
| |
|
|
=
|
|
|
\ .
U
( ) ( )
p
t t = X U
( )
t
'
= + X AX F
Nonhomogenous Linear Systems (contd.)
Variation of parameters (contd.)
Use the Product Rule and substitution to find that


The general solution of the system is
or

( ) ( ) ( )
1
p
t t t dt

=
}
X F
c p
= + X X X
( ) ( ) ( ) ( )
1
t t t t dt

= +
}
X C F
Matrix Exponential
We can define a matrix exponential so
that has a solution

where A is a matrix of constants
Since , is solution of
for every vector C of constants
'
= X AX
t
e =
A
X C
t
e
A
2
2
0
2! ! !
k k
t k k
k
t t t
e t
k k

=
= + + + + + =

A
I A A A A
t t
d
e e
dt
=
A A
A
t
e =
A
X C
'
= X AX
( )
t t t
d
e e e
dt
'
= = = =
A A A
X C A C A C AX
Matrix Exponential (contd.)
, denoted as (t) , is a fundamental matrix
of
can be calculated:
Using the Laplace transform
Using powers A
m

Using computer software
'
= X AX
t
e
A
t
e
A
Matrix Exponential (contd.)
Example: Use the Laplace transform to
compute for

First compute the matrix and find its
inverse
1 1
2 2

| |
=
|

\ .
A
t
e
A
s I A
1 1
2 2
s
s
s

| |
=
|
+
\ .
I A
( )
( ) ( )
( ) ( )
1
1
2 1
1 1
1 1
2 2 2 1
1 1
s
s s s s
s
s
s s
s s s s

+
| |
|
+ +

| |
|
= =
|
|
+
\ .
|
+ +
\ .
I A
Matrix Exponential (contd.)
Example (contd.):
Decompose the entries into partial fractions



Take the inverse Laplace transform to give
( )
1
2 1 1 1
1 1
2 2 1 2
1 1
s s s s
s
s s s s

| |
+
|
+ +
= |
|
+
|
+ + \ .
I A
2 1
2 2 1 2
t t
t
t t
e e
e
e e


| | +
=
|
+
\ .
A