Professional Documents
Culture Documents
We now present an application of matrix methods to linear systems of differential equations. We shall
follow the development given in Chapter 9 of Fundamentals of Differential Equations and Boundary
Value Problems by Nagle, Saff, Snider, third edition.
Calculus of Matrices
dA t A t a t
dt 0 0 ij 0 nn
Also
b b
a Atdt a a ij tdt
nn
d CA C dA , C a constant matrix
dt dt
d A B dA dB
dt dt dt
d AB A dB B dA
dt dt dt
In the last formula the order in which the matrices are written is important, since matrix multiplication
need not be commutative.
Linear Systems in Normal Form
x t Atxt ft 1
1
elements of A are constants, the system is said to have constant coefficients.
y n t p n1 ty n1 p 0 ty gt 2
can be rewritten as a first order system in normal form using the substitution
x 1 t yt, x 2 t y t, . . . . , x n t y n1 t 2. 5
Then
x 1 t y t x 2 t
x 2 t y t x 3 t
x n1 t y n1 t x n t
x n t y n t p n1 ty n1 p 0 ty gt
From 2. 5 we can write this last equation as
x n t p 0 tx 1 t p n1 tx n t gt
Thus the differential equation 2 can be put in the form 1 with
x1 0
x2 0
xt , ft
xn gt
and
0 1 0 0 0
0 0 1 0 0
A
0 0 0 0 1
p 0 t p 1 t p 2 t p n2 t p n1 t
The initial value problem for the normal system 1 is the problem of finding a differential vector
function xt that satisfies the system on an interval I and also satisfies the initial condition xt 0 x 0 ,
where t 0 is a given point of I and x 0 is a given constant vector.
2
Example:
Convert the initial value problem
y 3y 2y 0
y0 1
y 0 3
x 1 x 2
x 2 3x 2 2x 1
Thus
0 1 x1 0
x t
2 3 x2 0
x 1 0 1
We also have the initial condition x0 x0.
x 2 0 3
x t Atxt ft, xt 0 x 0
Remark: Just as in Ma 221 we may introduce the Wronskian of n vectors functions and use it to test for
linear independence. We have
3
x 11 t x 12 t x 1n t
x 21 t x 22 t x 2n t
Wx 1 , . . . . , x n t
x n1 t x n2 t x nn t
One can show that the Wronskian of solutions x 1 , . . . . , x n to x Ax is either identically zero or never
zero on and interval I. Also, a set of n solutions x 1 , . . . . . , x n to x Ax on I is linearly independent if
and only if their Wronskian is never zero on I. Thus the Wronskian provides us with an easy test for
linear independence for solutions of x Ax.
Theorem 2 (Representation of Solutions - Homogeneous Case)
Let x 1 , x 2 , . . . . , x n be n linearly independent solutions to the homogeneous system
x t Atxt 3
xt c 1 x 1 t c n x n t 4
A set of solutions x 1 , . . . . , x n that are linearly independent on I is called a fundamental solution set
for 3. The linear combination 4 is referred to as the general solution of 3.
Exercise:
e 2t e t e t
Verify that e 2t , 0 , e t is a fundamental solution set for the system
e 2t e t 0
0 1 1
x t 1 0 1 xt 5
1 1 0
0 1 1 1 1 1
1 0 1 , eigenvectors: 1 , 0 1, 1 2
1 1 0 0 1 1
4
e t
Consider x 3 t e t . Then
0
0 1 1 e t e t
Ax 3 t 1 0 1 e t e t x 3 t
1 1 0 0 0
e 2t e t e t
Remark: The matrix Xt e 2t 0 e t is a fundamental matrix for the DE 5. The general
e 2t e t 0
solution of 5 can be written as
e 2t e t e t
xt Xtc c 1 e 2t c2 0 c3 e t
e 2t e t 0
Remark: If we define an operator L by
Lx x Ax
x t Atxt ft 6
on the interval I, and let x 1 , x 2 , . . . . , x n be a fundamental solution set on I for the corresponding
homogeneous system x t Atxt. Then every solution to 6 on I can be expressed in the form
xt x p t c 1 x 1 t c n x n t 7
5
Remark: The linear combination of x p , x 1 , . . . . , x n written in 7 with arbitrary constants c 1 , . . . . . , c n is
called the general solution of 6. We may express this solution as x x p Xc, where X is a
fundamental matrix for the homogeneous system and c is an arbitrary constant vector.
x t Axt 8
Theorem 4
Suppose the n n constant matrix A has n linearly independent eigenvectors u 1 , u 2 , . . . , u n . Let r i be
the eigenvalue corresponding to the u i . Then
e r 1 t u 1 , e r 2 t u 2 , . . . . , e r n t u n 9
is a fundamental solution set on , for the homogeneous system x Ax. Hence the general
solution of x Ax is
xt c 1 e r 1 t u 1 c n e r n t u n
6
Remark: The eigenvalues may be real or complex and need not be distinct.
Proof
Since Au i r i u i we have
d e r i t u r e r i t u e r i t Au Ae r i t u
i i i i i
dt
so each element of the set 9 is a solution of the system 8. Also the Wronskian of these solutions is
Wt dete r 1 t u 1 , . . . . . , e r n t u n e r 1 r 2 r n t detu 1 , . . . . . . , u n 0
Example
Find a general solution of
5 4
x x
1 0
5 4 1 4
, eigenvectors: 1, 4
1 0 1 1
1 4
Thus xt c 1 e t c 2 e 4t
1 1
Thus the solution is
x 1 t c 1 e t 4c 2 e 4t
x 2 t c 1 e t c 2 e 4t
x 1 5x 1 4x 2 x 1 t 13 C 1 e t 43 C 1 e 4t 43 C 2 e 4t 43 C 2 e t
, Exact solution is:
x 2 x 1 x 2 t 13 C 1 e 4t 13 C 1 e t 43 C 2 e t 13 C 2 e 4t
7
Exercise:
Nagle and Saff page 535 #23. Find a fundamental matrix for the system
2 1 1 1
0 1 0 1
x t xt
0 0 3 1
0 0 0 7
Solution:
2 1 1 1 1 1 1
0 1 0 1 3 0 1
, eigenvectors: 1, 2, 7,
0 0 3 1 0 0 2
0 0 0 7 0 0 8
1
0
3
1
0
Hence the four linearly independent solutions are
1 1 1 1
3 0 1 0
e t , e 2t , e 7t , e 3t
0 0 2 1
0 0 8 0
Therefore a fundamental matrix is
e t e 2t e 7t e 3t
3e t 0 e 7t 0
0 0 2e 7t e 3t
0 0 8e 7t 0
We know that if a matrix has n distinct eigenvalues, then the eigenvectors associated with these
eigenvalues are linearly independent. Hence
Corollary
If the n n constant matrix A has n distinct eigenvalues r 1 , . . . . , r n and u i is an eigenvector associated
with r i then e r 1 t u 1 , . . . . , e r n t u n is a fundamental solution set for the homogeneous system x Ax.
8
Example
Solve the initial value problem
1 2 1 1
x t 1 0 1 xt x0 0
4 4 5 0
1 2 1
Solution: 1 0 1 , eigenvectors:
4 4 5
1 2 1
1 1, 1 2, 1 3
2 4 4
1 2 1 c 1 e t 2c 2 e 2t c 3 e 3t
Thus xt c 1 e t 1 c 2 e 2t 1 c 3 e 3t 1 c 1 e t c 2 e 2t c 3 e 3t
2 4 4 2c 1 e t 4c 2 e 2t 4c 3 e 3t
c 1 e t 2c 2 e 2t c 3 e 3t
We define xt via this. xt c 1 e t c 2 e 2t c 3 e 3t so that
2c 1 et 4c 2 e 2t 4c 3 e 3t
c 1 2c 2 c 3 1
x0 c1 c2 c3 0
2c 1 4c 2 4c 3 0
1 2 1 1 1 0 0 0
Thus we form 1 1 1 0 , row echelon form: 0 1 0 1 . Hence
2 4 4 0 0 0 1 1
c 1 0, c 2 1, c 3 1. Then the solution is
2 1
xt e 2t 1 e 3t 1
4 4
Complex Eigenvalues
9
We now discuss how one solves the system
x t Axt
in the case where A is a real matrix and the eigenvalues are complex. We shall show how to obtain two
real vector solutions of the system . Recall that if r 1 i is a solution of the equation that
determines the eigenvalues, namely,
p detA rI 0
then r 2 i is also a solution of this equation, and hence is an eigenvalue. Recall that r 2 is called
the complex conjugate of r 1 and r 1 r 2 .
Let z a ib, where a and b are real vectors, be an eigenvector corresponding to r 1 . Then it is not
hard to see that z a ib is an eigenvector corresponding to r 2 . Since
A r 1 Iz 0
then taking the conjugate of this equation and noting that since A and I are real matrices then A A and
II
A r 1 Iz A r 1 Iz A r 2 Iz 0
so z is an eigenvector corresponding to r 2 . Therefore the vectors
w 1 t e it a ib
and
w 2 t e it a ib
are two linearly independent vector solutions of . However, they are not real. To get real solutions
we proceed as follows: Since
e it e t cos t i sin t
then
10
so that x 1 t and x 2 t are real vector solutions of corresponding to the eigenvalues i.
Note that we can get the two expressions above for x 1 t and x 2 t by taking the real and imaginary
parts of w 1 t
Example
Find the general solution of the initial value problem
3 1 0
x t xt x
2 1 2 1
Solution:
We first find the eigenvalues and eigenvectors of the matrix. We want the roots of
3 r 1
3 r1 r 2 r 2 4r 5 0
2 1 r
Thus
4 16 415
r 2 i
2
The system of equations to determine the eigenvectors is
3 rx 1 x 2 0
2x 1 1 rx 2 0
or
3 rx 1 x 2 0
2x 1 1 rx 2 0
For r 2 i we have
1 ix 1 x 2 0
2x 1 1 ix 2 0
Multiplication of the first equation by i i yields the second equation, since 1 i1 i 2.
Thus
x 2 1 ix 1
1
Letting x 1 1 gives the eigenvector . Since the second eigenvector is the complex
1 i
conjugate of the first we have
3 1 1 1
, eigenvectors: 2 i, 2 i
2 1 1 i 1 i
1 1 0
i
1 i 1 1
1 0
Thus 2, 1, a ,b . The two linearly independent solutions are
1 1
11
1 1e 2t e it
wt e 2it
1 i 1 ie 2t e it
1 0 1 0
e 2t cos t sin t ie 2t sin t cos t
1 1 1 1
Therefore
1 0
x 1 t e t cos ta sin tb e 2t cos t sin t
1 1
1 0
x 2 t e t sin ta cos tb e 2t sin t cos t
1 1
Thus
1 0 1 0
xt c 1 e 2t cos t sin t c 2 e 2t sin t cos t
1 1 1 1
0 1 0
x 2 c 1 e c 2 e . Therefore
1 1 1
c 1 e and c 2 0
so
1 0
xt e 2t cos t sin t
1 1
Example Find a [real] general solution to
x 1 3 1 x1
.
x 2 2 1 x2
Solution: We first find the eigenvalues and eigenvectors of the matrix. We want the solutions to
3r 1
det 3 r1 r 2 r 2 4r 5 0
2 1r
Thus
4 16 415
r 2i
2
The system of equations to determine the eigenvectors is
12
3 rx 1 x 2 0
2x 1 1 rx 2 0
For r 2 i we have
1 ix 1 x 2 0
2x 1 1 ix 2 0
or
1 ix 1 x 2 0
2x 1 1 ix 2 0
One can see that the first and second equations are the same by multiplying the first equation by 1 i
and recalling that 1 i1 i 2. Thus
x2 1 x1
1i
We have
x1 1 x2 1 1i x 1i x
1i 1i 1i 2 2 2
Letting x 2 2 we have the eigenvector
1 i
2
3 1
Since the eigenvectors are complex conjugates we have that the eigenvalues of the matrix
2 1
1 i 1 i
are 2 i and 2 i and the corresponding eigenvectors are and .
2 2
3 1
The eigenvalues of the matrix are 2 i and 2 i and the corresponding eigenvectors are
2 1
1 i 1 i
and .
2 2
Let r 1 2 i, so 2, 1. Also
1 i 1 1
i
2 2 0
1 1
so a and b .
2 0
1 1
x 1 t e t cos ta sin tb e 2t cos t sin t
2 0
1 1
x 2 t e t sin ta cos tb e 2t sin t cos t
2 0
13
1 1 1 1
xt c 1 e 2t cos t sin t c 2 e 2t sin t cos t
2 0 2 0
Undetermined Coefficients
Consider the nonhomogeneous constant coefficient system
x t Axt ft
Before presenting the method for systems we recall the following result for second nonhomogeneous
order differential equations. For more on this see Linear Second Order DEs (Hold down the Shift key
and click.)
A particular solution of
ay by cy Ke x
where a, b, c are constants is
x
y p Ke if p 0
p
x
y p Kxe
if p 0, p 0
p
y p K x 2 e x if p p 0
p
where
pr ar 2 br c
Example
y 5y 4y 2e x
14
Homogeneous solution: p 2 5 4 4 1 4, 1 y h c 1 e x c 2 e 4x
Now to find a particular solution for 2e x . 1 p1 0 Since p 2 5
p 1 2 5 3 0
x 2xe x
y p kxe
p 3
y y h y p c 1 e x c 2 e 4x 2 xe x
3
Example Problem 3, page 547 of text.
Find the general solution of
1 2 2 2e t
x t 2 1 2 xt 4e t
2 2 1 2e t
Solution:
We first find the homogeneous solution.
1 2 2 1 1 1
2 1 2 , eigenvectors: 1 3, 1 , 0 3
2 2 1 1 0 1
a 1 2a 2 2a 3 2
et 2a 1 a 2 2a 3 4
2a 1 2a 2 a 3 2
Thus
15
a 1 a 1 2a 2 2a 3 2
a 2 2a 1 a 2 2a 3 4
a 3 2a 1 2a 2 a 3 2
, Solution is: a 2 0, a 1 1, a 3 1
Therefore
1
x p t et 0
1
and
1 1 1 1
xt x h t x p t c 1 e 3t 1 c 2 e 3t 1 c 3 e 3t 0 et 0
1 0 1 1
Note: the Method of Undetermined Coefficients works only for constant coefficient systems.
16
1 1
x h t c 1 e 3t c 2 e t .
2 2
To find x p we let
a1t b1
x p t
a2t b2
since ft is a polynomial.
b 1 t b 2 ta 1 ta 2 1
4b 1 4t b 2 4ta 1 ta 2 2
Equating the coefficients of t on both sides we have
0 1 a 1 a 2
0 4 4a 1 a 2
or
a1 a2 1
4a 1 a 2 4
Therefore a 1 1, a 2 0.
Equating the constant terms on both sides we have
a1 b1 b2 1
a 2 4b 1 b 2 2
Using the values for a 1, a 2 we have
b1 b2 2
4b 1 b 2 2
Thus b 1 0, b 2 2. With these values for the constants we have that
t
x p t
2
Finally
1 1 t
xt x h t x p t c 1 e 3t c 2 e t
2 2 2
17
2 2 4 cos t
x t xt
2 2 sin t
Solution: We first find a homogeneous solution.
2r 2
2 r 2 4 4 4r r 2 r r 2 4r rr 4
2 2r
Thus the eigenvalues are r 0, 4.
The equations for the eigenvectors are
2 rx 1 2x 2 0
2x 1 2 rx 2 0
For r 0 we have
2x 1 2x 2 0
1
or x 1 x 2 . Thus we have the eigenvector . For r 4 we have
1
2x 1 2x 2 0
2x 1 2x 2 0
1
or x 1 x 2 Thus we have the eigenvector .
1
Hence
1 1
x h t c 1 e 0t c 2 e 4t
1 1
We assume
a 1 cos t b 1 sin t
x p t
a 2 cos t b 2 sin t
Hence
a 1 sin t b 1 cos t
x p
a 2 sin t b 2 cos t
Plugging into the DE yields
a 1 sin t b 1 cos t 2 2 a 1 cos t b 1 sin t 4 cos t
a 2 sin t b 2 cos t 2 2 a 2 cos t b 2 sin t sin t
or
a 1 sin t b 1 cos t 2a 1 cos t 2a 2 cos t 2b 1 sin t 2b 2 sin t 4 cos t
a 2 sin t b 2 cos t 2a 1 cos t 2a 2 cos t 2b 1 sin t 2b 2 sin t sin t
We equate the coefficients of the sin t and cos t terms. Thus
18
b 1 2a 1 2a 2 4
b 2 2a 1 2a 2
a 1 2b 1 2b 2
a 2 2b 1 2b 2 1
or
2a 1 2a 2 b 1 4
2a 1 2a 2 b 2 0
a 1 2b 1 2b 2 0
a 2 2b 1 2b 2 1
To solve this system we form
1 0 2 2 0
0 1 2 2 1
2 2 1 0 4
2 2 0 1 0
and row reduce.
1 0 2 2 0 1 0 2 2 0 1 0 2 2 0
2R 1 R 3 2R 2 R 3
0 1 2 2 1 2R 1 R 4 0 1 2 2 1 2R 2 R 4 0 1 2 2 1
2 2 1 0 4 0 2 5 4 4 0 0 9 8 2
2 2 0 1 0 0 2 4 5 0 0 0 8 9 2
1 0 2 2 0 1 0 2 2 0 1 0 2 2 0
0 1 2 2 1 0 1 2 2 1 0 1 2 2 1
R 3 R 3 R 4 R 4 R 3
0 0 9 8 2 0 0 9 8 2 0 0 1 1 4
0 0 8 9 2 0 0 1 1 4 0 0 9 8 2
1 0 2 2 0 1 0 2 2 0 1 0 2 2 0
0 1 2 2 1 1 0 1 2 2 1 0 1 2 2 1
9R 3 R 4 17 R 4 R 4 R 3
0 0 1 1 4 0 0 1 1 4 0 0 1 0 2
0 0 0 17 34 0 0 0 1 2 0 0 0 1 2
1 0 0 2 4 1 0 0 0 0
2R 3 R 2 2R 4 R 2
2R 3 R 1 0 1 0 2 5 2R 4 R 1 0 1 0 0 1
0 0 1 0 2 0 0 1 0 2
0 0 0 1 2 0 0 0 1 2
19
Thus a 1 0, a 2 1, b 1 2, b 2 2 and
2 sin t
x p t
cos t 2 sin t
Finally a general solution is
1 1 2 sin t
xt x h t x p t c 1 c 2 e 4t
1 1 cos t 2 sin t
3 1
Example The eigenvalues of the matrix are 2 i and 2 i and the corresponding
2 1
1 i 1 i
eigenvectors are and .
2 2
Find a [real] general solution to
x 1 3 1 x1 25t
.
x 2 2 1 x2 0
x 1 3 1 x1
Solution: First we find a real general solution to .
x 2 2 1 x2
1 i 1 1
Here 2, 1 and i a ib.
2 2 0
Since
x 1 t e t cos ta sin tb
x 2 t e t sin ta cos tb
then
1 1
x 1 t e 2t cos t sin t
2 0
1 1
x 2 t e 2t sin t cos t
2 0
Hence
x h t c 1 x 1 t c 2 x 2 t
e 2t cos t sin t e 2t sin t cos t
c1 c2
2e 2t cos t 2e 2t sin t
Or we may expand one of the complex solutions and take the real and imaginary parts.
20
1 i 1 i
e 2it e 2t cos t i sin t
2 2
a 3 1 at b 25t
c 2 1 ct d 0
a 3a ct 3b d 25t
c 2a ct 2b d 0
We equate like terms.
0 3a c 25
0 2a c
a 3b d
c 2b d
Thus (from the first pair of equations) a 5, c 10 and then b 1 and d 8. Combining
homogeneous and particular solutions, we have a general solution.
21
x1 e 2t cos t sin t e 2t sin t cos t 5t 1
c1 c2
x2 2e 2t cos t 2e 2t sin t 10t 8
Solution: The solution is the general solution x h to the homogeneous equation plus one [particular]
solution x p to the full non-homogeneous equation. First well find x p . It is in the form
c 1 e 2t
xp
c 2 e 2t
22
Substituting into the D.E., we obtain
x 1 2c 1 e 2t 2 1 c 1 e 2t 0
x 2 2c 2 e 2t 1 2 c 2 e 2t 12e 2t
Hence
2c 1 e 2t 2c 1 e 2t c 2 e 2t 0 2c 1 e 2t c 2 e 2t
2c 2 e 2t c 1 e 2t 2c 2 e 2t 12e 2t c 1 e 2t 2c 2 e 2t 12e 2t
We can divide by e 2t (which is never zero) and move the unknowns to the left side to obtain
c 2 0
c 1 12
12e 2t
xp
0
To find a solution to the homogeneous solution we use the eigenvalue 2 i i and the
1 1 0
corresponding eigenvector i a ib.
i 0 1
Since
x 1 t e t cos ta sin tb
x 2 t e t sin ta cos tb
then
1 0
x 1 t e 2t cos t sin t
0 1
1 0
x 2 t e 2t sin t cos t
0 1
Hence
x h t c 1 x 1 t c 2 x 2 t
e 2t cos t e 2t sin t
c1 c2
e 2t sin t e 2t cos t
Or, for the solution to the homogeneous equation, we may use one of the eigenvalues and eigenvectors
found in 2a to write a complex solution and break it into real and imaginary parts. well use 2 i.
23
1 1
x e 2it e 2t cos t i sin t
i i
e 2t cos t ie 2t sin t
e 2t sin t ie 2t cos t
e 2t cos t e 2t sin t
xh c1 c2
e 2t sin t e 2t cos t
e 2t cos t e 2t sin t c1
e 2t sin t e 2t cos t c2
Finally, we add to obtain the desired solution.
e 2t cos t e 2t sin t c1 12e 2t
x
e 2t sin t e 2t cos t c2 0
where the entries in At may be any continuous functions of t. Let Xt be a fundamental matrix for
the homogeneous system
x t Atxt 2
The general solution of 2 is
Xtc
where c is an n 1 constant vector. To find a particular solution of 1, consider
x p t Xtvt
where vt is an n 1 vector function of t that we wish to determine. Then
x p t Xtv t X tvt
so that 1 yields
Xtv t X tvt Atx p t ft AtXtvt ft
Since Xt is a fundamental matrix for 2, then X t AXt, so the last equation becomes
Xtv t AtXtvt AtXtvt ft
or
Xtv t ft
24
Hence
vt X 1 tftdt
and
x p t Xtvt Xt X 1 tftdt
Solution:
We first find a fundamental matrix for the homogeneous solution.
2 3 1 3
, eigenvectors: 1, 1
1 2 1 1
3e t e t
Xt
et e t
1 1
3e t e t 1t
2e t
Hence 2e
X 1
et e t 1t 3
2e 2e t
Formula 3 yields
1 1t
3e t e t c1 3e t e t e 2t
xt 2e t 2e
dt
et e t c2 et e t 1t 3 1
2e 2e t
1 et 1
3e t e t c1 3e t e t 2 2e t
et e t c2 et e t 1 e 3t 3
6 2e t
3e t e t c1 3e t 12 e t 1 t e t 16 e 3t 3t
2e 2e
et e t c2 e t 12 e t 1 e t 16 e 3t 3t
2e t 2e
25
3e t e t c1 3e t 12 e t 1 t e t 16 e 3t 3t
2e 2e
xt
et e t c2 e t 12 e t 1 e t 16 e 3t 3t
2e t 2e
3c 1 c 2 13
3 1
x0
c1 c2 3 7 0
Thus
3c 1 c 2 13 1
3
c1 c2 7 0
3
, Solution is: c 2 56 , c 1 32 . Finally
3e t e t 32 3e t 12 e t 1 t e t 16 e 3t 3t
2e 2e
xt
et e t 56 e t 12 e t 1 e t 16 e 3t 3t
2e t 2e
16 e 2t 27e t 5e 3t 8 18e 2t
16 e 2t 9e t 5e 3t 2 12e 2t
x t Axt 1
2 n n
e At I At A2 t
2!
An t
n!
A n n!
t 2
n0
This is an n n matrix.
Example
1 0
Let D . Then
0 2
26
1 0 1 0 1 n 0
D2 , D3 ,.... Dn
0 4 0 8 0 2n
Therefore
n
n n0 1 n t 0 e t 0
e Dt Dn t
n!
n!
n
n0 0 n0 2n t 0 e 2t
n!
In general if D is an n n diagonal matrix with r 1 , r 2 , . . . . , r n down its main diagonal, then e Dt is the
diagonal matrix with e r 1 t , e r 2 t , . . . . , e r n t down its main diagonal.
It can be shown that the series 2 converges for all t and has many of the same properties as the scalar
exponential e at .
Remark: It can be shown that if a matrix A has n linearly independent eigenvectors, then P 1 AP is a
diagonal matrix, where P is formed from the n linearly independent eigenvectors of A. Thus
P 1 AP D 3
where D is a diagonal matrix. In fact, D has the eigenvalues of A along its diagonal.
1 2 1 1 1 2
Let A 1 0 1 . Then 1 3, 1 1, 1 2
4 4 5 4 2 4
1 0 1
1 1 2 2
Thus P 1 1 1 , inverse: 0 2 12 P 1
4 2 4 1 1 0
1 0 1
2 1 2 1 1 1 2 3 0 0
Hence P 1 AP 0 2 12 1 0 1 1 1 1 0 1 0
1 1 0 4 4 5 4 2 4 0 0 2
A PDP 1
so that
27
1
e At e PDP t I PDP 1 t 1 PDP 1 t PDP 1 t
2
I PDP 1 t 1 PDP 1 PDP 1 t 2
2
I PDP 1 t 1 PD 2 P 1 t 2
2
P I Dt Dt 2 P 1
1
2
Pe P
Dt 1
Example:
5 4 2 1
Let A , , eigenvectors: 3, 1
2 1 1 1
1 2 1 2 1 0
P , P 1 and D . Show that A P 1 DP and use this to
1 1 1 1 0 3
compute e At .
1 2 1 0 1 2 5 4
A PDP 1 as required.
1 1 0 3 1 1 2 1
5 4
t
2 1 e t 2e 3t 2e 3t 2e t
Thus e from SNB.
e 3t e t 2e t e 3t
1 0
t
1 2 0 3 1 2
Also Pe Dt P 1 e so
1 1 1 1
1 2 et 0 1 2 e t 2e 3t 2e t 2e 3t
e At Pe Dt P 1
1 1 0 e 3t 1 1 e t e 3t 2e t e 3t
This e At is a fundamental matrix for the system
5 4
x t xt
2 1
since if we let
28
5 4
t
2 1 c1 e t 2e 3t 2e t 2e 3t c1
x h t e
c2 et e 3t 2e t e 3t c2
c 2 2e 3t 2e t c 1 e t 2e 3t
c 1 e t e 3t c 2 e 3t 2e t
c 2 2e 3t 2e t c 1 e t 2e 3t
c 1 e t e 3t c 2 e 3t 2e t
then
c 2 6e 3t 2e t c 1 e t 6e 3t
x h t
c 1 e t 3e 3t c 2 3e 3t 2e t
and
5 4 c 2 2e 3t 2e t c 1 e t 2e 3t
Ax h t
2 1 c 1 e t e 3t c 2 e 3t 2e t
4c 1 e t e 3t 5c 1 e t 2e 3t 4c 2 e 3t 2e t 5c 2 2e 3t 2e t
2c 1 e t 2e 3t c 1 e t e 3t c 2 e 3t 2e t 2c 2 2e 3t 2e t
6c 1 e 3t 2c 2 e t c 1 e t 6c 2 e 3t c 2 6e 3t 2e t c 1 e t 6e 3t
x h t
c 1 e t 2c 2 e t 3c 1 e 3t 3c 2 e 3t c1 et 3e 3t c2 3e 3t 2e t
:
Proof of f.
29
d e At d I At A 2 t2 An tn
dt dt 2! n!
2 n1 2 n
A A2t A3 t An t A I At A 2 t A n t
2! n 1! 2! n!
Ae At
e At e r 1 t u 1, e r 2 t u 2 , , e r n t u 1 , u 2 , , u n 1
Since
2 n n
e At I At A2 t
2!
An t
n!
A n n!
t
n0
we see that if A is nilpotent, then the infinite series has only a finite number of terms since
A k A k1 0 and in this case
2 k1
e At I At A 2 t A k1 t
2! k 1!
This may be taken further. The Cayley-Hamilton Theorem says that a matrix satisfies its own
characteristic equation, that is, pA 0. Therefore, if the characteristic polynomial for A has the form
pr 1 n r r 1 n , that is A has only one multiple eigenvalue r 1 , then
pA 1 n A r 1 I n 0. Hence A r 1 I is nilpotent and
e At e r 1 IAr 1 It e r 1 It e Ar 1 It e r 1 t I A r 1 It A r 1 I n1 t n1
n 1!
Example Find the fundamental matrix e At for the system
30
2 1 1
x t Axt where A 1 2 1
2 2 1
Solution: The characteristic polynomial for A is
2r 1 1
pr det 1 2r 1 r 3 3r 2 3r 1 r 1 3
2 2 1 r
Hence r 1 is an eigenvalue of A with multiplicity 3. By the Cayley-Hamilton Theorem A I 3 0
and
2
e At e t e AIt e t I A It A I 2 t
2!
1 1 1 0 0 0
AI 1 1 1 and A I 2 0 0 0
2 2 2 0 0 0
Thus
1 0 0 1 1 1 e t te t te t te t
e At e t 0 1 0 te t 1 1 1 te t e t te t te t
0 0 1 2 2 2 2te t 2te t e t 2e t
31