Systems Of Linear Equations

1 Matrix Functions
A matrix (or vector) whose elements are functions of a real variable is called a matrix function.
 
a11 (t) a12 (t) · · · a1n (t)
 a21 (t) a22 (t) · · · a2n (t) 
A matrix function A(t) =  .. ..  = [aij (t)] is said to be continuous at t = t0
 
..
 . . . 
an1 (t) an2 (t) · · · ann (t)
(or on an open interval I) if each aij (t) is continuous at t0 (or on I).
A(t) is differentiable on I if each aij (t) is differentiable on I, and its derivative is defined by
 
0 dA(t) daij (t)
A (t) = = .
dt dt

The integral of A(t) is defined as

Z b Z b 
A(t) dt = aij (t) dt .
a a

The following are some of the rules of differentiation for matrix functions:
d dA
(a) (kA) = k , where k is a scalar;
dt dt
d dA d dA
(b) (CA) = C , (AC) = C, where C is a constant matrix;
dt dt dt dt
d dA dB
(c) (A + B) = + ;
dt dt dt
d dB dA
(d) (AB) = A + B.
dt dt dt
    Z π  
sin t t 0 cos t 1 2 π 2 /2
Example 1. If A(t) = , then A (t) = , A(t) dt = .
1 cos t 0 − sin t 0 π 0

1
2 Matrices & Systems Of Linear First-Order Equations
Definition 2.1. General First-Order Linear Systems
A first-order system of linear equations is the set of simultaneous differential equations of the
form

x01 = a11 (t)x1 + a12 (t)x2 + · · · + a1n (t)xn + f1 (t)

x02 = a21 (t)x1 + a22 (t)x2 + · · · + a2n (t)xn + f2 (t) (1)
.. ..
. .
0
xn = an1 (t)x1 + an2 (t)x2 + · · · + ann (t)xn + fn (t)

where the coefficients aij (t) and the nonhomogeneous terms fi (t) are continuous functions on a given
open interval I.

Remark.

(a) If the coefficients aij (t) are constants, then the linear system is said to have constant coefficients.
(b) If all the functions fi (t) = 0∀x ∈ I, then the system is called homogeneous; otherwise, it is said
to be nonhomogeneous.
(c) Using matrix notation,
     
x1 (t) a11 (t) a12 (t) · · · a1n (t) f1 (t)
 x2 (t)   a21 (t) a22 (t) · · · a2n (t)   f2 (t) 
x(t) =  ..  , A(t) =  .. , f (t) =  .. 
     
..
 .   . .   . 
xn (t) an1 (t) an2 (t) · · · ann (t) fn (t)
we can write system (1) in matrix form as

x0 (t) = A(t)x(t) + f (t) (2)

(d) A vector x(t) is said to be a solution of (2) if its components satisfy the system of equations
(1) on the interval I.
(e) The problem of finding a solution to the linear system (2) that also satisfies the specified initial
conditions
x1 (t0 ) = α1 , x2 (t0 ) = α2 , . . . , xn (t0 ) = αn
is called the initial-value problem for the linear system.

Example 2. Verify that

2 2 −3t 1 2 −3t
x1 (t) = − e , x2 (t) = + e
3 3 3 3
is a solution of the system
dx1
= −2x1 + x2 + 1
dt .
dx2
= x1 − 2x2
dt

2
Example 3. One reason for studying systems of DEs lies in the fact that all higher order DEs can be
written as an equivalent system of first-order DEs.
Rewrite the IVP
y 00 + 3y 0 + 2y = et , y(0) = 4, y 0 (0) = 5
as a system of two first-order equations.

Remark. The order n linear DE

y (n) + an−1 (t)y (n−1) + · · · + a1 (t)y 0 + a0 (t)y = f (t)

with initial conditions

y(x0 ) = A0 , y 0 (x0 ) = A1 , . . . , y (n−1) (x0 ) = An−1
can be rewritten as a system of n first-order linear DEs by defining

x1 = y, x2 = y 0 , x3 = y 00 , . . . , xn = y (n−1) .

Theorem 2.1. Existence and Uniqueness of Solutions to Linear Systems

If the elements of the matrix A = [aij (t)] and the column vector f = [aij (t)] are continuous on an open
interval I that contains the point t0 , then for any given vector x0 , the IVP

x0 (t) = A(t)x(t) + f (t), x(t0 ) = x0

has a unique solution on I.

Theorem 2.2. Principle of Superposition for Linear Systems

If the vector functions x(1) and x(2) are solutions to the linear homogeneous system x0 = A(t)x, then
the linear combination c1 x(1) + c2 x(2) is also a solution for any constants c1 and c2 .

Definition 2.2. Linearly Dependent and Independent Vector Functions

The vector functions x(1) (t), x(2) (t), . . . , x(n) (t) are linearly dependent on an interval I if there exist
constants c1 , c2 , . . . , cn , not all zero, for which

c1 x(1) (t) + c2 x(2) (t) + · · · + cn x(n) (t) = 0 ∀t ∈ I.

If the vectors are not linearly dependent on I, they are said to be linearly independent on I.

Definition 2.3. The Wronskian

The Wronskian of the n (column) vector functions

x(1) (t), x(2) (t), . . . , x(n) (t)

is the determinant
  
W x(1) (t), x(2) (t), . . . , x(n) (t) = x(1) (t) x(2) (t) · · ·

x(n) (t) .

Theorem 2.3. Wronskian Test for Linearly Independence

A set x(1) (t), x(2) (t), . . . , x(n) (t) of solutions of the linear homogeneous system x0 = A(t)x is linearly


independent on an open interval I iff their Wronskian is never zero on I.

3
Definition 2.4. Fundamental Set of Solutions

A set x(1) (t), x(2) (t), . . . , x(n) (t) of n linearly independent solutions of the homogeneous system

x0 = A(t)x (3)

on an interval I is called a fundamental set of solutions on that interval.

Theorem 2.4. Existence of a Fundamental Set

There exists a fundamental set of solutions for the homogeneous system (3) on an interval I.

Theorem 2.5. General Solution of a Homogeneous System

If x(1) (t), x(2) (t), . . . , x(n) (t) are n linearly independent solutions of the n-dimensional linear homoge-
neous system (3) on an interval I where the elements of A(t) are continuous, then any solution x(t)
of this system can be written in the form

x(t) = c1 x(1) (t) + c2 x(2) (t) + · · · + cn x(n) (t) (4)

where c1 , c2 , . . . , cn are constants.

Remark. The linear combination (4) is called the general solution of the linear system.
 −2t   6t 
(1) e (2) 3e
Example 4. Show that x (t) = −2t and x (t) = 6t form a fundamental set of solutions of
  −e 5e
0 1 3
x = x on (−∞, ∞). Hence, write down the GS of this system on this interval.
5 3

3 Homogeneous Linear System with Constant Coefficients

Definition 3.1. Eigenvalues and Eigenvectors
Let A be a constant n × n matrix. The eigenvalues of A are the values of λ for which the system of
equations
(A − λI) v = 0
has at least one nonzero solution v. They can be found by solving the characteristic equation

p(λ) = |A − λI| .

The nonzero solutions v are called the eigenvectors of A corresponding to λ.

Theorem 3.1. General Solution of x0 = A(t)x

If A is an n × n constant matrix with n linearly independent eigenvectors v1 , v2 , . . . , vn corresponding
to the real eigenvalues λ1 , λ2 , . . . , λn , then {eλ1 t v1 , eλ2 t v2 , . . . , eλn t vn } constitutes a fundamental set
of the homogeneous system x0 = A(t)x. The general solution of x0 = A(t)x is

x(t) = c1 eλ1 t v1 + c2 eλ2 t v2 + · · · + cn eλn t vn

where c1 , c2 , . . . , cn are arbitrary constants.

4
Remark. It can be shown that if an n × n matrix A has n distinct real eigenvalues, then A has n
linearly independent eigenvectors.
 
0 2 3
Example 5. Find the general solution of x = Ax where A = .
2 1
Theorem 3.2. Real Solutions Corresponding to a Complex Eigenvalue
If α ± iβ are complex conjugate eigenvalues of the real matrix A with corresponding eigenvectors
a ± ib, then
x1 = eαt (a cos βt − b sin βt) and x2 = eαt (a sin βt + b cos βt)
are linearly independent real-valued solutions of x0 = Ax.
 
0 2 8
Example 6. Solve x = x.
−1 −1

4 Repeated Eigenvalues
Definition 4.1.

(a) An eigenvalue is of (algebraic) multiplicity k if it is a k-fold root of the characteristic equation

|A − λI| = 0.
(b) An eigenvalue of multiplicity k is said to be complete if it has k linearly independent associated
eigenvectors.
(c) An eigenvalue of multiplicity k is called defective if it is not complete.
Remark.

(a) If every eigenvalue of an n × n matrix A is complete, then A will have n linearly independent
eigenvectors associated with its eigenvalues.
(b) If some of the eigenvalues of A are defective, then our usual method will produce fewer than the
needed n linearly independent solutions of the system x0 = A(t)x.
Therefore to construct the general solution of the system, we need to find the “missing solutions”
using a different method.

4.1 Defective Eigenvalues

The following method is used to find the above mentioned “missing solutions”.
Let λ be a defective eigenvalue of multiplicity m.
(a) (i) Eigenvalue of Multiplicity Two (m = 2)
If A has only one eigenvector associated with an eigenvalue λ of multiplicity two, then we
can find two linearly independent solutions of the form
x1 = v1 eλt
x2 = (v1 t + v2 ) eλt
where the columns vectors v1 and v2 must satisfy
(A − λI)v1 = 0
(A − λI)v2 = v1

5
(ii) Eigenvalue of multiplicity Three (m = 3)
If A has only one eigenvector associated with an eigenvalue λ of multiplicity three, then we
can find three linearly independent solutions of the form
x1 = v1 eλt
x2 =  (v1 t + v2 ) eλt 
1
x3 = v1 t + v2 t + v3 eλt
2
2!
where the columns vectors v1 , v2 and v3 must satisfy
(A − λI)v1 = 0
(A − λI)v2 = v1
(A − λI)v3 = v2
(iii) The General Case
If A has only one eigenvector associated with an eigenvalue λ of multiplicity m, then we can
find m linearly independent solutions of the form
x1 = v1 eλt
x2 = (v1 t + v2 ) eλt
..
.  
1 m−1 1
xm = v1 t 2
+ · · · + vm−2 t + vm−1 t + vm eλt
(m − 1)! 2!
where the column vectors vi must satisfy
(A − λI)v1 = 0
(A − λI)v2 = v1
.. .
.
(A − λI)vm−1 = vm

6
  
0 0 −2
Example 7. Solve x0 = 1 2 1  x.
1 0 3
Example 8. Defective Multiplicity 3 Eigenvalues
 
2 1 6
Solve x0 = 0 2 5 x.
0 0 2
Answer.

(a) |A − λI| = 0 ⇒ λ = 2, 2, 2
(b) (i) [A − λI]v1 = 0 ⇒ v1 = [1, 0, 0]T
(ii) [A − λI]v2 = v1 ⇒ v2 = [0, 1, 0]T
(iii) [A − λI]v3 = v2 ⇒ v3 = [0, −6, 1]T
(c) Therefore the general solution is
x(t) = c1 v1 e2t + c2 [v1 te2t + v2 e2t ] + c3 [v1 (t2 /2)e2t + v2 te2t + v3 e2t ] where v1 , v2 , and v3 are as
given above.

7
Exercise 1. Defective Multiplicity 3 Eigenvalues
 
−3 0 −4
Solve x0 = −1 −1 −1 x.
1 0 1

Answer. x(t) = c1 v1 e−t +c2 [v1 te−t + v2 e−t ]+c3 [v1 (t2 /2)e−t + v2 te−t + v3 e−t ] where v1 = [0, 1, 0]T , v2 =
[−2, −1, 1]T , and v3 = [1, 0, 0]T .

5 Fundamental Matrices
Definition 5.1. Fundamental Matrix
If the column vectors x1 , x2 , . . . , xn form a fundamental set of solutions for the homogeneous system
x0 = Ax on an open interval I, the the n × n matrix
 
Ψ(t) = x1 x2 · · · xn

is called a fundamental matrix of the system.

Remark. The general solution of the above homogeneous system is

x = c1 x1 + c2 x2 + · · · + cn xn
or, in terms of Ψ(t),
x = Ψ(t)c,
where c is a constant vector with arbitrary components c1 , . . . , cn .
 −2t   6t 
(1) e (2) 3e
Example 9. We showed in Example 4 that x (t) = −2t and x (t) = form a fundamental
  −e 5e6t
1 3
set of solutions of x0 = x on (−∞, ∞). Hence, a fundamental matrix of the system on this
5 3
interval is  −2t 
e 3e6t
Ψ(t) = .
−e−2t 5e6t
Furthermore, the general solution of the system is
 −2t  
e 3e6t c1
x = Ψ(t)c = .
−e−2t 5e6t c2

8
Theorem 5.1. Fundamental Matrix Solutions
Let Ψ(t) be a fundamental matrix for the homogeneous system x0 = Ax. Then the solution of the
IVP
x0 = Ax, x(t0 ) = x0
is given by
x = Ψ(t)Ψ−1 (t0 )x0 .

Example 10. A fundamental matrix for the system

 
0 1 3
x = x
5 3

in Example 9, was found to be

e−2t 3e6t
 
Ψ(t) = .
−e−2t 5e6t
 
1
Use it to find the solution of this system that satisfies the initial conditions x(0) = .
−1

6 Nonhomogeneous Systems
Theorem 6.1. General Solution of a Nonhomogeneous System
Let xp (t) be a particular solution of the nonhomogeneous linear system

x0 (t) = A(t)x(t) + f (t) (5)

on an interval I, and let xh (t) = c1 x(1) (t) + c2 x(2) (t) + · · · + cn x(n) (t) denote the general solution on
I of the corresponding n-dimensional homogeneous system x0 (t) = A(t)x(t). Then the the general
solution of the nonhomogeneous system (5) can be written in the form

x(t) = xh (t) + xp (t).

6.1 Undetermined Coefficients

Example 11. Use the method of undetermined coefficients to solve the system
     
0 1 3 12 −11
x = x+ t+
5 3 0 −3

on (−∞, ∞). Hence, write down the GS of this system on this interval.
 −2t   6t 
e 3e
[Hint: You may assume the result of Example 4 that xh (t) = c1 −2t + c2 .]
−e 5e6t
 −2t   6t     
e 3e 3 −4
Answer. The GS is x(t) = c1 −2t + c2 6t + t+ .
−e 5e −5 6

9
Example 12. Determined the form of the particular solution xp for

dx
= 5x + 3y − 2e−t + 1
dt
dy
= −x + y + e−t − 5t + 7
dt
given that the general solution of the associated homogeneous system is
   
1 2t 3
xh (t) = c1 e + c2 e4t .
−1 −1

Note : Do not actually determine the values of the coefficients.

10
7 Variation Of Parameters
Theorem 7.1. Variation Of Parameters
If Ψ(t) is a fundamental matrix for x0 (t) = A(t)x(t) on some interval where A(t) and f (t) are contin-
uous, then a particular solution of the nonhomogeneous linear system
x0 (t) = A(t)x(t) + f (t)
is Z
xp (t) = Ψ(t) Ψ−1 (t)f (t)dt.

Remark.
(a) The general solution of the nonhomogeneous system is
Z
x = Ψ(t)c + Ψ(t) Ψ−1 (s)f (s)ds

where c is a vector of arbitrary constants.

(b) The solution of the IVP
x0 (t) = A(t)x(t) + f (t), x(t0 ) = x0
is Z t
−1
x = Ψ(t)Ψ (t0 )x0 + Ψ(t) Ψ−1 (s)f (s)ds.
t0
Example 13. Use variation of parameters to find the general solution of
   3t 
0 1 1 2e
x = x+ .
4 1 0

e−t
 
e3t
[Hint: You may assume the fundamental matrix for this system is Ψ(t) = .]
2e3t −2e−t
Reading Assignment 1. Use variation of parameters to solve
   3t   
0 1 1 2e 1
x = x+ , x(0) = .
4 1 0 0
Answer. You can check that the fundamental matrix and its inverse are
e−t −2e−t −e−t
 3t   
e −1 1
Ψ(t) = , Ψ (t) = .
2e3t −2e−t −4e2t −2e3t e3t
−2e−s −e−s 2e3s
    
−1 1 1
Hence Ψ (s)f (s) = −4e2s 3s 3s = 4s and
−2e e 0 e
Z t Z t   
1 t
Ψ−1 (s)f (s)ds = 4s ds = e4t −1 .
0 0 e 4
Therefore, the solution of the IVP is
Z t
−1
x = Ψ(t)Ψ (t0 )x0 + Ψ(t) Ψ−1 (s)f (s)ds
t0
−t 1 (4t + 1)e3t − e−t 1 (4t + 3)e3t + e−t
 3t
      
1 e e −2 −1 1
= + = .
−4 2e3t −2e−t −2 1 0 4 (8t − 2)e3t + 2e−t 4 (8t + 2)e3t − 2e−t

11
8 Solution of x0(t) = A(t)x(t) + f (t) by Diagonalizing A
If A is a constant , diagonalizable n×n matrix, then A has n linearly independent eigenvectors. These
eigenvectors form columns of an invertible matrix P such that
 
λ1 0 · · · 0
 0 λ2 · · · 0 
P−1 AP = D =  ..
 
.. .. 
. . ··· . 
0 0 · · · λn

with eigenvalues down the main diagonal in the order corresponding to the eigenvector columns of P.
Using the substitution x = Pz, the system x0 (t) = A(t)x(t) + f (t) is transformed into

z0 = Dz(t) + P−1 f .

This is an uncoupled system of the form

z10 = λ1 z1 + g1 (t)
z20 = λ2 z2 + g2 (t)
..
.
zn0 = λn zn + gn (t)

where  
g1 (t)
 g2 (t) 
P−1 f =  ..  .
 
 . 
gn (t)
Solve these n first-order DEs independently to get
 
z1 (t)
 z2 (t) 
z(t) =  .. 
 
 . 
zn (t)

and the solution to the original problem

x(t) = Pz(t).
 
−2 1
Example 14. Find the general solution of the system x0 = x by diagonalizing the coefficient
−4 3
matrix.

12