Methods of Applied Mathematics

Definition The following system of differential equations:

is called a first-order system of 𝒏-linear differential equations of 𝑛 unknown functions

𝑥𝑖 (𝑡), 𝑖 = 1,2, … , 𝑛. The coefficients 𝑎𝑖𝑗 (𝑡), 𝑖, 𝑗 = 1,2, … , 𝑛 and 𝑓𝑖 (𝑡), 𝑖 = 1,2, … , 𝑛 are
known functions of the independent variable 𝑡.

Remarks
(1) If 𝑓𝑖 (𝑡) = 0, for all 𝑖 = 1,2, … , 𝑛, then system (1) is called homogeneous; otherwise,
it is nonhomogeneous.
(2) If 𝑎𝑖𝑗 are constant, for all 𝑖, 𝑗 = 1,2, … , 𝑛, then system (1) is called of constant
coefficients; otherwise, it is of variable coefficients.
(3) System in form (1) is called in normal form.
(4) System (1) can be expressed in a matrix form as follows:

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
or simply

where

where 𝐀(𝑡) is called the coefficient matrix.

(5) If the system (1) is homogeneous, its matrix form is then

Example Express the following system in matrix form:

(1)

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
 (2)

Definition (Initial-Value Problem)

Let 𝑡0 denote a point on an interval 𝐼 and

where the 𝛾𝑖 ,𝑖 1, 2, . . . , 𝑛 are given constants. Then the problem

is an initial-value problem on the interval.

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
Example Determine the largest interval for which the following initial value problem
ensures the existence and uniqueness of a solution:

Recall from Matrix Theory I

𝑎 𝑏
 Let 𝐴 = ( ). Then determinant of 𝐴: = |𝐴| = 𝑎𝑏 − 𝑐𝑑.
𝑐 𝑑
1 2
Example Let 𝐴 = ( ). Then |𝐴| =
3 4
𝑎11 𝑎12 𝑎13
 Let 𝐴 = (𝑎21 𝑎22 𝑎23 ).
𝑎31 𝑎32 𝑎33
𝑎22 𝑎23 𝑎21 𝑎23 𝑎21 𝑎22
Then |𝐴|=𝑎11 |𝑎 𝑎33 | − 𝑎12 |𝑎31 𝑎33 | + 𝑎13 |𝑎31 𝑎32 |.
32

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
 3 sin 𝑡 −cos 𝑡
Example Let 𝐴 = (4 cos 𝑡 sin 𝑡 ). Then |𝐴| =
0 − sin 𝑡 cos 𝑡

1, 𝑖 = 𝑗
 A matrix 𝐼𝑛×𝑛 = 𝐼𝑛 = [𝑎𝑖𝑗 ], 𝑖, 𝑗 = 1,2, … , 𝑛 is called identity matrix if 𝑎𝑖𝑗 = { .
0, 𝑖 ≠ 𝑗

1 0 0
1 0
Example 𝐼3 = (0 1 0) and 𝐼2 = ( ) are identity matrices.
0 1
0 0 1

 Note that 𝐴 × 𝐼 = 𝐼 × 𝐴 = 𝐴.

 An 𝑛 × 𝑛 matrix 𝐵 is called the inverse of 𝐴, |𝐴| ≠ 0 iff 𝐴𝐵 = 𝐵𝐴 = 𝐼𝑛 .

 Note that the inverse of 𝐴 is unique and denoted by 𝐴−1 .

 Note that |𝐴| ≠ 0 iff 𝐴 is invertible (has inverse).

Example Using the elementary row operations to find the inverse of the following
matrices:
1 5 −9
(1) 𝐴 = (0 1 0), (2) 𝐵 = (1 −1).
4 −2
0 −1 1

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
 |𝐴| ≠ 0 iff the algebraic system 𝐴𝑋⃗ = 𝑏⃗⃗ has a unique solution, where the solution is

𝑋⃗ = 𝐴−1 𝑏⃗⃗. Moreover, if 𝑏⃗⃗ = ⃗0⃗, the unique solution is the trivial one (𝑋⃗ = ⃗0⃗).

 Eigenvalues and Eigenvectors:

o Let 𝐴 = [𝑎𝑖𝑗 ] be an 𝑛 × 𝑛 matrix. Consider the following vector equation:

𝐴𝑋⃗ = 𝜆𝑋⃗ (6)

where 𝜆 is a scalar (real or complex) to be determined and 𝑋⃗ is a vector also to be

determined.
o 𝑋⃗ = 0 (trivial solution) is a solution for the equation (6) for every 𝜆.

o A scalar 𝜆 that satisfies the equation (6) for some 𝑋⃗ ≠ 0 is called an eigenvalue

of 𝐴, and the vector is called an eigenvector of 𝐴 corresponding to 𝜆.

 To find 𝜆, we write (6) as:

𝐴𝑋⃗ − 𝜆𝑋⃗ = ⃗0⃗

or
(𝐴 − 𝜆𝐼)𝑋⃗ = ⃗⃗
0

o If |𝐴 − 𝜆𝐼| ≠ 0, then 𝑋⃗ = ⃗⃗
0 ( not required).

o So, to obtain 𝑋⃗ ≠ ⃗0⃗, we assume

|𝐴 − 𝜆𝐼| = 0. (7)

o Equation (7) is called the auxiliary (characteristic) equation of 𝐴, and

𝑃(𝜆) = |𝐴 − 𝜆𝐼| is called the characteristic polynomial of 𝐴.

o The roots of |𝐴 − 𝜆𝐼| = 0 are the eigenvalues of 𝐴.

 To determine the eigenvector 𝑋⃗ ≠ 0 that corresponds to 𝜆, we need to solve the

vector equation (𝐴 − 𝜆𝐼)𝑋⃗ = ⃗0⃗.

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
Example Find the eigenvalues and corresponding eigenvectors to the following matrix:

−5 4
𝐴=( )
−6 6

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
Definition (Diagonal Matrix)

Let 𝐴 = [𝑎𝑖𝑗 ]𝑛×𝑛 .

(1) If 𝑎𝑖𝑗 = 0 for all 𝑖 ≠ 𝑗, then 𝐴 is called a diagonal matrix.

(2) If 𝑎𝑖𝑗 = 0 for all 𝑖 > 𝑗, then 𝐴 is called an upper diagonal matrix.
(3) If 𝑎𝑖𝑗 = 0 for all 𝑖 < 𝑗, then 𝐴 is called a lower diagonal matrix.

Example The following matrices are diagonal, upper diagonal, and lower diagonal,
respectively:
1 0 0 0
1 0 0 4 5 −9
𝐴 = (0 2 0 ), 𝐵 = (0 3 0 ), 𝐶 = (7 2 0 0)
0 4 −3 0
0 0 −1 0 0 1 3 5 2 0
 If

is a diagonal matrix, then

Example Let

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
  If 𝐴 = [𝑎𝑖𝑗 ]𝑛×𝑛 is a diagonal, upper diagonal, or lower diagonal matrix, then 𝑎𝑖𝑖 , 𝑖 =

1,2, … , 𝑛 are the eigenvalues of 𝐴.

Example Find the eigenvalues and corresponding eigenvectors to the following matrix:
1 0 0
𝐴 = (0 2 0)
0 0 −1

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
Solving System of Linear Differential Equations
Definition (Solution Vector)

Example Verify that on the interval ( −∞, ∞):

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
Example Find another solution to the system in the previous example.

Definition (Linear Dependence/Independence)

Example Show that the following two vectors are linearly independent on the
interval ( −∞, ∞):

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
Definition (Wronskian)
𝑥11 (𝑡) 𝑥12 (𝑡) 𝑥1𝑛 (𝑡)
𝑥 (𝑡) 𝑥 (𝑡) 𝑥 (𝑡)
Let 𝑋1 (𝑡) = ( 21 ) , 𝑋2 (𝑡) = ( 22 ) , … , 𝑋𝑛 (𝑡) = ( 2𝑛 )
⋮ ⋮ ⋮
𝑥21 (𝑡) 𝑥𝑛2 (𝑡) 𝑥𝑛𝑛 (𝑡)
be 𝑛 vector functions. The Wronskian of 𝑋1 (𝑡), 𝑋2 (𝑡), … , 𝑋𝑛 (𝑡) is defined to be a real-
valued function:
𝑥11 (𝑡) 𝑥12 (𝑡) ⋯ 𝑥1𝑛 (𝑡)
𝑥 (𝑡) 𝑥22 (𝑡) ⋯ 𝑥2𝑛 (𝑡)
𝑊[𝑋1 , 𝑋2 , … , 𝑋𝑛 ](𝑡) = | 21 ⋱ | .
⋮ ⋮ ⋮
𝑥21 (𝑡) 𝑥𝑛2 (𝑡) ⋯ 𝑥𝑛𝑛 (𝑡)

3 sin 𝑡 − cos 𝑡
Example Let 𝑋1 (𝑡) = (4) , 𝑋2 (𝑡) = ( cos 𝑡 ) , … , 𝑋𝑛 (𝑡) = ( sin 𝑡 ). Find
0 −sin 𝑡 cos 𝑡
𝑊[𝑋1 , 𝑋2 , 𝑋3 ](𝑡).

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
Example We verified in a previous example that

on the interval ( −∞, ∞). Show that they are linearly independent.

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
Definition (Fundamental Set of Solutions)

Example We verified in previous examples that

on the interval ( −∞, ∞), and are linearly independent on ( −∞, ∞). So,

is a fundamental set of system solutions on the interval ( −∞, ∞).

Example For the system in the previous example, the general solution on the interval
( −∞, ∞) is

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
Example The vectors

are solutions of the system

Find a general solution to the system.

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
  Consider the system of linear ODEs:
𝑋 ′ = 𝐴𝑋 (8)
where 𝐴 is a constant matrix.

 Assume that
𝑋 = 𝐾𝑒 𝜆𝑡 (9)
be a solution to the system (8), where 𝜆 is a scalar and 𝐾 ≠ 0 is a constant vector, which
they shall be determined.

Then
𝑋 ′ = 𝜆𝑒 𝜆𝑡 𝐾 (10)
 Substitute (9) and (10) into (8), then we have
𝜆𝑒 𝜆𝑡 𝐾 = 𝑒 𝜆𝑡 𝐴𝐾
or
𝑒 𝜆𝑡 𝐴𝐾 − 𝜆𝑒 𝜆𝑡 𝐾 = 0
𝑒 𝜆𝑡 (𝐴 − 𝜆𝐼)𝐾 = 0
(𝐴 − 𝜆𝐼)𝐾 = 0.
 But 𝐾 ≠ 0. So,
|𝐴 − 𝜆𝐼| = 0 (11)
 Hence, the values of 𝜆 are the eigenvalues of 𝐴 and 𝐾 are the corresponding
eigenvectors.

 The values of the roots of Equation (11) (the eigenvalues of 𝐴) has one of the following
cases:

 Distinct real eigenvalues

 Repeated real eigenvalues
 Complex eigenvalues

So, Then the general solution of (8) on the interval (−∞ , ∞ ) is given by

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
Example Find a general solution of

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
Example Solve the initial value problem

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
We discuss two situations here:
 𝐴𝑛×𝑛 has 𝑛 linearly independent eigenvectors.
 𝐴𝑛×𝑛 has 𝑘 linearly independent eigenvectors, where 𝑘 < 𝑛.

Case I 𝐴𝑛×𝑛 has 𝑛 linearly independent eigenvectors.

Example Find a general solution of
2 0 0
′
𝑋 = 𝐴𝑋 , where 𝐴 = (1 1 2)
1 −1 4

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
Recall from Matrix Theory II

 Let 𝐴 = [𝑎𝑖𝑗 ]. The transpose of 𝐴 is denoted and defined as 𝐴𝑇 = [𝑎𝑗𝑖 ].

For example;

 A real symmetric matrix 𝐴 is a matrix with real entries that satisfies 𝐴 = 𝐴𝑇 .

For example;

THEOREM 2.1 Eigenvectors of a real symmetric matrix

If 𝐴 is a real symmetric matrix, then it always has 𝑛 linearly independent eigenvectors.

Example Find a general solution of

Hint: 𝜆1 = 𝜆2 = 3, 𝜆3 = −3

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
Case II 𝐴𝑛×𝑛 has 𝑘 < 𝑛 linearly independent eigenvectors.

Example Find a general solution of

THEOREM 2.2 linearly independent solutions corresponding to an eigenvalue

of multiplicity 𝒎
If there is only one eigenvector 𝐾1 corresponding to the eigenvalue 𝜆1 of multiplicity 𝑚,
then there are 𝑚 linearly independent solutions to the system 𝑋 ′ = 𝐴𝑋 of the forms:
𝑋1 = 𝑒 𝜆1 𝑡 𝐾1 ,
𝑋2 = 𝑡𝑒 𝜆1 𝑡 𝐾1 + 𝑒 𝜆1 𝑡 𝐾2 ,
𝑡2 𝜆 𝑡
𝑋3 = 𝑒 1 𝐾1 + 𝑡𝑒 𝜆1 𝑡 𝐾2 + 𝑒 𝜆1 𝑡 𝐾3 ,
2!
⋮
𝑡 𝑚−1 𝑡 𝑚−2
𝑋𝑚 = 𝑒 𝜆1 𝑡 𝐾1 + 𝑒 𝜆1 𝑡 𝐾2 + ⋯ + 𝑒 𝜆1 𝑡 𝐾𝑚 ,
(𝑚 − 1)! (𝑚 − 2)!
where 𝐾𝑖 can be obtained from the vector equation:
(𝐴 − 𝜆1 𝐼)𝐾𝑖 = 𝐾𝑖−1 , 𝑖 = 2,3, … , 𝑚 .

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
Example Let 𝐾1 be the only eigenvector of a matrix 𝐴 corresponding to the eigenvalue 𝜆1 ,
which is of multiplicity greater than 1. Show that 𝑋 = 𝑡𝑒 𝜆1 𝑡 𝐾1 + 𝑒 𝜆1 𝑡 𝐾2 is a solution to the
system 𝑋 ′ = 𝐴𝑋, where (𝐴 − 𝜆1 𝐼)𝐾2 = 𝐾1 .

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
Example Solve

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
Example Find a general solution of

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
HOMOGENEOUS LINEAR SYSTEM OF THE FORM:
𝑡𝑋 ′ (𝑡) = 𝐴𝑋(𝑡), 𝑡 ≠ 0 , (12)
where 𝐴 is a constant matrix.

 Assume that
𝑋 = 𝑡𝜆𝐾 (13)
be a solution to the system (12), where 𝜆 is a scalar and 𝐾 ≠ 0 is a constant vector, which
they shall be determined.

Then
𝑋 ′ = 𝜆𝑡 𝜆−1 𝐾 (14)
 Substitute (13) and (14) into (12), then we have
𝜆𝑡 𝜆 𝐾 = 𝑡 𝜆 𝐴𝐾
or
𝑡 𝜆 𝐴𝐾 − 𝜆𝑡 𝜆 𝐾 = 0
𝑡 𝜆 (𝐴 − 𝜆𝐼)𝐾 = 0
(𝐴 − 𝜆𝐼)𝐾 = 0.
 But 𝐾 ≠ 0. So,
|𝐴 − 𝜆𝐼| = 0 (15)
 Hence, the values of 𝜆 are the eigenvalues of 𝐴 and 𝐾 are the corresponding
eigenvectors.

 The values of the roots of Equation (15) (the eigenvalues of 𝐴) has one of the following
cases:

 Distinct real eigenvalues

 Repeated real eigenvalues
 Complex eigenvalues

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
Example Find a general solution of
𝑑𝑥
𝑡 = 2𝑥 + 3𝑦
𝑑𝑡
𝑑𝑦
𝑡 = 2𝑥 + 𝑦
𝑑𝑡

Example Solve

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
Example Find a general solution of
2 0 0
′
𝑡𝑋 = 𝐴𝑋 , where 𝐴 = (1 1 2)
1 −1 4

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
Example Find a general solution of
1 −2 2
′
𝑡𝑋 = 𝐴𝑋 , where 𝐴 = ( −2 1 2)
2 2 1
Hint: 𝜆1 = 𝜆2 = 3, 𝜆3 = −3

Example Find a general solution of

THEOREM 2.3 linearly independent solutions corresponding to an eigenvalue

of multiplicity 𝒎
If there is only one eigenvector 𝐾1 corresponding to the eigenvalue 𝜆1 of multiplicity 𝑚,
then there are 𝑚 linearly independent solutions to the system 𝑡𝑋 ′ = 𝐴𝑋 of the forms:
𝑋1 = 𝑡 𝜆1 𝐾1 ,
𝑋2 = 𝑡 𝜆1 ln 𝑡 𝐾1 + 𝑡 𝜆1 𝐾2 ,
(ln 𝑡)2 𝜆
𝑋3 = 𝑡 1 𝐾1 + 𝑡 𝜆1 ln 𝑡 𝐾2 + 𝑡 𝜆1 𝐾3 ,
2!
⋮
(ln 𝑡)𝑚−1 𝜆 (ln 𝑡)𝑚−2 𝜆
𝑋𝑚 = 𝑡 1 𝐾1 + 𝑡 1 𝐾2 + ⋯ + 𝑡 𝜆1 𝐾𝑚 ,
(𝑚 − 1)! (𝑚 − 2)!
where 𝐾𝑖 can be obtained from the vector equation:
(𝐴 − 𝜆1 𝐼)𝐾𝑖 = 𝐾𝑖−1 , 𝑖 = 2,3, … , 𝑚 .

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
Example Solve 𝑡

Example Find a general solution of

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
  There are two methods to find a particular solution to the system (16):
 The undetermined coefficients method.
 The variation of parameters method.

 To apply the method of undetermined coefficients to find a particular solution 𝐗 𝑝 for

the system (16), the following conditions must be met:

1. A must be a constant matrix.

2. 𝐟(𝑡) should take the form:

𝐟⃗(𝑡) = 𝐏 ⃗⃗⃗𝑚 (𝑡)𝑒 𝛼𝑡 sin 𝛽𝑡,

⃗⃗𝑛 (𝑡)𝑒 𝛼𝑡 cos 𝛽𝑡 + 𝐐

⃗⃗𝑛 (𝑡) and ⃗⃗⃗

where 𝐏 𝐐𝑚 (𝑡) are polynomial vectors of degree 𝑛, 𝑚, respectively. That is

⃗⃗⃗⃗⃗
𝐏𝒏 (𝑡) = 𝐚⃗⃗𝑛 𝑡 𝑛 + 𝐚⃗⃗𝑛−1 𝑡 𝑛−1 + ⋯ + 𝐚
⃗⃗1 𝑡+𝐚
⃗⃗0 .

 Then 𝐗 𝑝 is of the form:

⃗⃗⃗𝑘 (𝑡)𝑒 𝛼𝑡 cos 𝛽𝑡 + 𝑡 𝑠 𝐇

𝐗 𝑝 (𝑡) = 𝑡 𝑠 𝐇 ⃗⃗⃗𝑘∗ (𝑡)𝑒 𝛼𝑡 sin 𝛽𝑡,

⃗⃗⃗𝑘 (𝑡) and 𝐇

where𝐇 ⃗⃗⃗𝑘∗ (𝑡) are polynomial vectors of degree 𝑘 = max{𝑛, 𝑚}, and 𝑠 is the

multiplicity of 𝜆0 among the eigenvalues of A.

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
Example Find a general solution of

Solution: Recall from a previous example that

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
Example Find a general solution of

Solution: Hint

Example (H.W) Find a general solution of

Example (H.W) Find a general solution of

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
Definition (Fundamental Matrix)
Let 𝐗1 (𝑡), 𝐗2 (𝑡), … , 𝐗𝑛 (𝑡) be 𝑛 linearly independent solutions to the homogeneous system
𝑋 ′ = 𝐴𝑋 on an interval 𝐼. Then the following matrix:

is called a fundamental matrix for 𝑋 ′ = 𝐴𝑋 on the interval 𝐼.

Example We find, in a previous example, that a general solution of the system

is

Determine a fundamental matrix for the system.

Remark
The general solution of the homogeneous system 𝑋 ′ = 𝐴𝑋 can be expressed as:
𝑋 = 𝚽(𝑡)𝐶
where 𝚽(𝑡) is a fundamental matrix for the system and 𝐶 is an arbitrary constant vector.

Example The general solution of the homogeneous system

can be expressed as:

Remark
 A fundamental matrix 𝚽(𝑡) is nonsingular.
 If 𝚽(𝑡) is a fundamental matrix of the system 𝐗 ′ (𝑡) = 𝐀(𝑡)𝐗(𝑡), then
𝚽 ′ (𝑡) = 𝐀(𝑡)𝚽(𝑡) (18)

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
VARIATION OF PARAMETERS

 Assume that
𝐗 𝑝 = 𝚽(𝑡)U(𝑡) (19)
be a particular solution to the nonhomogeneous system:

𝐗 ′ = 𝐀(𝑡)𝐗(𝑡) + 𝐟(𝑡) (20)

where 𝚽(𝑡) is a fundamental matrix for the homogeneous system 𝐗 ′ = 𝐀(𝑡)𝐗(𝑡), and

is a variable vector that will be determined.

 Derivation of the expression in (18) gives

𝐗 ′𝑝 = 𝚽(𝑡)U ′ (𝑡) + 𝚽 ′ (𝑡)U(𝑡) (21)

Substitute (19) and (21) into (20) to have

𝚽(𝑡)U ′ (𝑡) + 𝚽 ′ (𝑡)U(𝑡) = 𝐀(𝑡)𝚽(𝑡)U(𝑡) + 𝐟(𝑡) (22)

Substitute (18) into (22) to obtain

𝚽(𝑡)U ′ (𝑡) + 𝐀(𝑡)𝚽(𝑡)U(𝑡) = 𝐀(𝑡)𝚽(𝑡)U(𝑡) + 𝐟(𝑡)

and so,

𝚽(𝑡)U ′ (𝑡) = 𝐟(𝑡) (23)

Since 𝚽(𝑡) is nonsingular, multiplying (23) by 𝚽 −𝟏 (𝑡) gives

U ′ (𝑡) = 𝚽 −𝟏 (𝑡)𝐟(𝑡) (24)

and so,

U(𝑡) = ∫ 𝚽 −𝟏 (𝑡)𝐟(𝑡) 𝑑𝑡.

Hence,

𝐗 𝑝 = 𝚽(𝑡) ∫ 𝚽 −𝟏 (𝑡)𝐟(𝑡) 𝑑𝑡 (25)

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
is a particular solution to the nonhomogeneous system, and

𝐗(𝑡) = 𝚽(𝑡)𝐶 + 𝚽(𝑡) ∫ 𝚽 −𝟏 (𝑡)𝐟(𝑡) 𝑑𝑡 (26)

is a general solution to the nonhomogeneous system.

Example Find a general solution of

Solution: Hint

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
Remark (1) If the solution given in (26) is defined for 𝑡 ≥ 𝑡0 , then the solution becomes
as:
𝑡

𝐗(𝑡) = 𝚽(𝑡)𝐶 + 𝚽(𝑡) ∫ 𝚽 −𝟏 (𝜏)𝐟(𝜏) 𝑑𝜏 (27)

𝑡0

(2) If the solution given in (27) satisfies the initial condition 𝐗(𝑡0 ) = 𝐗 𝟎 , then the
solution becomes as follows:
𝑡

𝐗(𝑡) = 𝚽(𝑡)𝚽 −𝟏 (𝑡0 )𝐗 𝟎 + 𝚽(𝑡) ∫ 𝚽 −𝟏 (𝜏)𝐟(𝜏) 𝑑𝜏 (28)

𝑡0

Example Solve the IVP:

Solution: Hint

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
Example Find a general solution of

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
Recall that the expansion of the exponential function 𝑒 𝑎𝑡 is expressed as follows:
∞
𝑎𝑡 2
𝑡2 3
𝑡3 𝑘
𝑡𝑘 𝑘
𝑡𝑘
𝑒 = 1 + 𝑎𝑡 + 𝑎 +𝑎 + ⋯+ 𝑎 +⋯ = ∑𝑎 .
2! 3! 𝑘! 𝑘!
𝑘=0

Example Find 𝑒 𝐴𝑡 where 𝐴 = (1 −1).

2 −2

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
 𝑑
Example: Show that 𝑑𝑡 (𝑒 𝐴𝑡 ) = 𝐴𝑒 𝐴𝑡 .

Example: Prove the previous Theorem.

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
 Corollary Let 𝐴 be an 𝑛 × 𝑛 constant matrix. Then, a general solution to the
nonhomogeneous system
𝐗 ′ (𝑡) = 𝐀(𝑡)𝐗(𝑡) + 𝐟(𝑡), 𝑡 ≥ 𝑡0
has the form
𝑡

𝐗(𝑡) = 𝑒 𝐴𝑡 𝐶 + ∫ 𝑒 𝐴(𝑡−𝜏) 𝐟(𝜏) 𝑑𝜏.

𝑡0
With the initial condition 𝐗(𝑡0 ) = 𝐗 0 , it becomes
𝑡

𝐗(𝑡) = 𝑒 𝐴(𝑡−𝑡0 ) + ∫ 𝑒 𝐴(𝑡−𝜏) 𝐟(𝜏) 𝑑𝜏.

𝑡0
Example: Solve the IVP:
1 −1 1 1
𝐗 ′ (𝑡) = ( ) 𝐗(𝑡) + ( ) 𝑒 −𝑡 , 𝐗(0) = ( ).
2 −2 −1 0

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
 Using Laplace Transform
Theorem Let 𝐴 be an 𝑛 × 𝑛 constant matrix. Then

Proof.

Example Find 𝑒 𝐴𝑡 where 𝐴 = (1 −1).

2 −2

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
 Exponential Diagonal Matrix
𝑑1 0 0 0 𝑒 𝑑1 𝑡 0 0 0
𝑑2 0 0 0 𝑒 𝑑2 𝑡 0 0
Theorem Let 𝐴 = ( 0 ⋱
𝐴𝑡
0 ). Then 𝑒 = ( 0 ⋱ 0 ).
0 0 0
0 0 0 𝑑𝑛 0 0 0 𝑒 𝑑𝑛𝑡

Example: Find 𝑒 𝐼3 𝑡 .

Example: Solve
2 0 0
′ (𝑡)
𝐗 = 𝐴𝐗(𝑡), 𝐴 = (0 3 0 ).
0 0 −1

Definition A matrix 𝐴 is said to be nilpotent if there exists some positive integer 𝑚 such
that 𝐴𝑚 = 0.

Example: Verify the following matrix is nilpotent

−1 1 1
𝐴 = (−1 0 1).
−1 1 1

Remark: If 𝐴𝑚 = 0 for some positive integer 𝑚. Then

𝑡2 𝑡 𝑚−1
𝑒 𝐴𝑡 = 𝐼 + 𝐴𝑡 + 𝐴2 + ⋯ + 𝐴𝑚−1 .
2! (𝑚 − 1)!

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
Example: Find a general solution to
−1 1 1
𝐗 ′ (𝑡) = 𝐴𝐗(𝑡), 𝐴 = (−1 0 1).
−1 1 1

Theorem Let 𝚽(𝑡) be a fundamental matrix for the system 𝑋 ′ = 𝐴𝑋. Then 𝚿(𝑡) = 𝚽(𝑡)𝑪
is also a fundamental matrix for the system, where 𝑪 is a constant matrix. In particular,
𝑒 𝐴𝑡 = 𝚽(𝑡)𝚽(0)−𝟏 . (𝟑𝟎)

Proof.

Example: Let 𝚽(𝑡) = (1 𝑒 −𝑡 ) be a fundamental matrix for the system 𝑋 ′ = (1 −1

) 𝑋. Find
1 2𝑒 −𝑡 2 −2
𝑒 𝐴𝑡 and a general solution to the system.

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
Theorem If |𝐴 − 𝜆𝐼| = (𝜆 − 𝜆1 )𝑛 = 0 is the characteristic equation of the matrix 𝐴𝑛×𝑛 , then
(𝐴 − 𝜆1 𝐼)𝑛 = 0 (i.e. 𝐴 − 𝜆1 𝐼 is nilpotent).

Remark: According to the previous theorem,

(𝐴−𝜆1 𝐼)𝑡 𝑡2 𝑡𝑛−1

𝑒 = 𝐼 + (𝐴 − 𝜆1 𝐼)𝑡 + (𝐴 − 𝜆1 𝐼 )2 + ⋯ + ( 𝐴 − 𝜆1 𝐼 )𝑛−1
2! ( 𝑛 − 1) !
Thus,

𝐴𝑡 𝜆1 𝐼𝑡
𝑡2 𝑡𝑛−1
𝑒 =𝑒 {𝐼 + (𝐴 − 𝜆1 𝐼)𝑡 + (𝐴 − 𝜆1 𝐼 )2 + ⋯ + ( 𝐴 − 𝜆1 𝐼 )𝑛−1 }
2! ( 𝑛 − 1) !

𝜆1 𝑡
𝑡2 𝑡𝑛−1
=𝑒 𝐼 {𝐼 + (𝐴 − 𝜆1 𝐼)𝑡 + (𝐴 − 𝜆1 𝐼 )2 + ⋯ + ( 𝐴 − 𝜆1 𝐼 )𝑛−1 }
2! ( 𝑛 − 1) !

𝜆1 𝑡
𝑡2 𝑡𝑛−1
=𝑒 {𝐼 + (𝐴 − 𝜆1 𝐼)𝑡 + (𝐴 − 𝜆1 𝐼)2 + ⋯ + (𝐴 − 𝜆1 𝐼)𝑛−1 }
2! ( 𝑛 − 1) !
Example: Find a general solution to the system
2 1 1
𝑋 ′ = 𝐴𝑋 where 𝐴=( 1 2 1 ).
−2 −2 −1
(Hint: the characteristic equation of the matrix 𝐴 is (𝜆 − 1)3 = 0)

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
Definition: Let A be a square matrix. A nonzero vector u that satisfies
(A−𝜆I)𝒎 u= 0
for some eigenvalue 𝜆 of A and some positive integer 𝑚, a vector u is called a generalized
eigenvector associated with 𝜆.

Remark:
(1) Note that 𝜆 must be an eigenvalue of A since (A−𝜆I)𝒎−𝟏u is a “regular” eigenvector.
(2) If the characteristic polynomial of A is
𝑝(𝜆) = (𝜆 − 𝜆1 )𝑚1 (𝜆 − 𝜆1 )𝑚2 … (𝜆 − 𝜆1 )𝑚𝑘
where 𝜆𝑖 , 𝑖 = 1,2, … , 𝑘 are the distinct eigenvalues of A and 𝑚𝑖 is the multiplicity of
the eigenvalue 𝜆𝑖 , then for each 𝑖, there exist 𝑚𝑖 linearly independent generalized
eigenvectors satisfying
(A−𝜆𝑖 I)𝑚𝑖 u= 0.
(3) In total, there exist 𝑛 = 𝑚1 + 𝑚2 + ⋯ + 𝑚𝑘 linearly independent generalized
eigenvectors of A.
(4) We can compute 𝑒 𝐴𝑡 u in finite terms without knowing 𝑒 𝐴𝑡 because

𝑒 𝐴𝑡 𝐮 = 𝑒 𝐴𝑡−𝜆𝐈𝑡+𝜆𝐈𝑡 𝐮 = 𝑒 𝜆𝐈𝑡 𝑒 (𝐴−𝜆𝐈)𝑡 𝐮

𝑡2 𝑡 𝑚−1 𝑡𝑚
= 𝑒 𝜆𝐈𝑡 {𝐈 + (𝐴 − 𝜆𝐈)𝑡 + (𝐴 − 𝜆𝐈)2 + ⋯ + (𝐴 − 𝜆𝐈)𝑚−1 + (𝐴 − 𝜆𝐈)𝑚 + ⋯}𝐮
2! (𝑚 − 1)! 𝑚!
𝑡2 𝑡 𝑚−1
= 𝑒 𝜆𝑡 𝐈 {𝐈 + (𝐴 − 𝜆𝐈)𝑡 + (𝐴 − 𝜆𝐈)2 + ⋯ + (𝐴 − 𝜆𝐈)𝑚−1 + 𝟎 + ⋯}𝐮
2! (𝑚 − 1)!
𝑡2 𝑡 𝑚−1
= 𝑒 𝜆𝑡 {𝐮 + (𝐴 − 𝜆𝐈)𝐮𝑡 + (𝐴 − 𝜆𝐈)2 𝐮 + ⋯ + (𝐴 − 𝜆𝐈)𝑚−1 𝐮 }.
2! (𝑚 − 1)!

(5) Sure, 𝑒 𝐴𝑡 u is a solution to the system 𝑋 ′ = 𝐴𝑋.

(6) If there exists {𝐮1 , 𝐮2 , … , 𝐮𝑛 } as a linearly independent set of generalized
eigenvectors of an 𝑛 × 𝑛 matrix A, then

𝚽(𝑡) = [𝑒 𝐴𝑡 𝐮1 : 𝑒 𝐴𝑡 𝐮2 : … : 𝑒 𝐴𝑡 𝐮𝑛 ]
forms a fundamental matrix of the system 𝑋 ′ = 𝐴𝑋.
So,

𝑒 𝐴𝑡 = 𝚽(𝑡)𝚽(0)−𝟏 = [𝑒 𝐴𝑡 𝐮1 : 𝑒 𝐴𝑡 𝐮2 : … : 𝑒 𝐴𝑡 𝐮𝑛 ][𝐮1 : 𝐮2 : … : 𝐮𝑛 ]−1.

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
Example: Find the fundamental matrix 𝑒 𝐴𝑡 of the system
1 0 0
X ′ =AX where A= [1 3 0].
0 1 1

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
 In Problems 1–6, (a) Show that the given matrix A satisfies 21. Use the results of Problem 5 to find
(A−𝜆I)𝑘 u= 0 for some number 𝜆 and some positive integer 𝑘. the solution to the initial value problem:
(b) Use this fact to determine the matrix 𝑒 𝐴𝑡 .

22. Use the results of Problem 12 to find

the solution to the initial value problem:

23. Use the results of Problem 3 and the

variation of parameters formula (16) to
find the solution to the initial value
problem:

24. Use your answer to Problem 9 and

the variation of parameters to find the
solution to the initial value problem:

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
Solving Systems Using Laplace Transforms

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
Example: Solve the IVP:
𝑑𝑦 𝑑𝑥
= 𝑦 − 𝑥 + 𝑒 −𝑡 , = 2𝑦 − 2𝑥 − 𝑒 −𝑡 , 𝑦(0) = 0, 𝑥(0) = 1.
𝑑𝑡 𝑑𝑡

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
Converting a 𝒏th-order differential equation into a system of
first-order differential equations:

The 𝑛th-order differential equation

𝑦 (𝑛) = 𝑓(𝑡, 𝑦, 𝑦 ′ , 𝑦 ′′ , … , 𝑦 (𝑛−1) )
can be converted to a system of first-order differential equations as follows:
Let 𝑥1 = 𝑦 Then, 𝑥1′ = 𝑦 ′ = 𝑥2
𝑥2 = 𝑦′ 𝑥′2 = 𝑦′′ = 𝑥3
𝑥3 = 𝑦′′ 𝑥′3 = 𝑦′′′ = 𝑥4
⋮ ⋮
𝑥𝑛 = 𝑦(𝑛−1) . 𝑥′𝑛 = 𝑦(𝑛) = 𝑓(𝑡, 𝑦, 𝑦′ , 𝑦′′ , … , 𝑦(𝑛−1) )
= 𝑓(𝑡, 𝑥1 , 𝑥2 , 𝑥3 , … , 𝑥𝑛 ).
Example: Convert the following 2nd-order ODE into a system of 1st-order ODEs:
4𝑦 ′′ + 8𝑦′ + 3𝑦 = 0.

Exercise: Convert the following 2nd-order ODE into a system of 1st-order ODEs:
(𝒊) 𝑦 ′′ − 2𝑦′ − 3𝑦 = 0.
(𝒊𝒊) 𝑦 ′′′ − 2𝑡𝑦′ + 3𝑦 = sin 𝑡.
This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
Chapter 3: Plane Autonomous Systems
★ Autonomous Systems

Definition A system of first-order DEs is said to be autonomous when the system can be
written in the form:

Note: The independent variable 𝑡 does not appear explicitly on the right-hand side of each
equation in the system.

Example: The following system

is not autonomous, whereas the following system is autonomous:

Remarks:

① If the variable 𝑡 in (2.1) is interpreted as time, we can refer to (2.1) as a dynamical

system.

② The linear system 𝐗 ′ = 𝐀𝐗 + 𝐹(𝑡) is autonomous if 𝐀 and 𝐹 are constants.

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
③ If 𝑛 = 2 or 3 in (2.4), we call the curve of the solution of (2.1), X(𝑡) a path or
trajectory, and 𝑥 = 𝑦1 (𝑡), 𝑦 = 𝑦2 (𝑡), 𝑧 = 𝑦3 (𝑡) are parametric equations of a curve.

Definition The autonomous system (2.1) is called a plane autonomous system if 𝑛 = 2,

and then we write the system as:

Remarks:

① The equation

is called the phase plane equation.

𝑑𝑦
② 𝑑𝑥 is the slope of the tangent of the trajectory (the curve of the solution of (2.2) ) at

𝑡 = 𝑡0 .

③ 𝑥𝑦-plane that contains the trajectories is called the phase plane.

The set of all tangent lines of the trajectories is called the direction field of the
solution.

Definition

 A point (𝑥0 , 𝑦0 ) where 𝑃(𝑥0 , 𝑦0 ) = 0 and 𝑄(𝑥0 , 𝑦0 ) = 0 is called a critical point

(or equilibrium point) of the system (2.2).
 The corresponding constant solution where 𝑥(𝑡) = 𝑥0 , 𝑦(𝑡) = 𝑦0 , is called an
equilibrium solution.
 The set of all critical points is called the critical point set.

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
Note: The trajectories of equilibrium solutions consist of just single points (the
equilibrium points).

Example: Consider the following system:

1. Find the critical point set.

2. Sketch the direction field in the phase plane.
3. Sketch the trajectories of the solutions.

Note: From the figure, we can see that all solutions "flow into" the critical point (0,0)
Such a critical point is called asymptotically stable.

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
★ Types of solutions for the plane autonomous system:

𝜕𝑃 𝜕𝑃 𝜕𝑄 𝜕𝑄
 If 𝑃(𝑥, 𝑦), 𝑄(𝑥, 𝑦), and 𝜕𝑥 , 𝜕𝑦 , 𝜕𝑥 , and 𝜕𝑦 are continuous in a region 𝑅 of the plane, then
𝑥(𝑡) 𝑥(0) 𝑥0
a solution of (2.4), X(𝑡) = ( ), that satisfies X(0) = ( ) =X0= (𝑦 ), is unique
𝑦(𝑡) 𝑦(0) 0
and one of three basic types:

(i) A constant solution (or equilibrium solution)

𝑥0
𝑥(𝑡) = 𝑥0 , 𝑦(𝑡) = 𝑦0 or X(𝑡) = (𝑦 )
0

where (𝑥0 , 𝑦0 ) is a critical point of (2.4).

𝑥(𝑡)
(ii) A solution X(𝑡) = ( ) whose trajectory is an arc (a plane curve that does not cross
𝑦(𝑡)
itself).

There would be two solutions that start from the

point P of intersection.

(iii) A periodic solution (or cyclic solution)

𝑥(𝑡 + 𝑃) = 𝑥(𝑡) and 𝑦(𝑡 + 𝑃) = 𝑦(𝑡) where 𝑃 is the period of the solution.

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
Example: Find all critical points of each of the following plane autonomous systems:

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
Example: Determine whether the given linear system possesses a periodic solution:

2
In each case, sketch the graph of the solution that satisfies 𝐗(0) = ( ).
0

Solution: (a) Check that:

be a general solution to the system in (a).

Thus, every solution is periodic with period 𝑃 = 𝜋.

2
At the initial condition 𝐗(0) = ( ), we have:
0

which are parametric equations of an ellipse.

(b) Check that:

be a general solution to the system in (b),

which is not a periodic solution.

2
At the initial condition 𝐗(0) = ( ), the solution is:
0

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
★ Changing to Polar Coordinates:

By the formulas:

we have:

Example: Find a solution of the nonlinear plane autonomous system:

Satisfying the initial condition:

3
X(0) = ( ).
3

Solution: From polar coordinate conversion, we get:

This is the spiral described by the system.

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
Example: A plane autonomous system in polar coordinates is given by:

𝑑𝑟
= 0.5(3 − 𝑟)
𝑑𝑡

𝑑𝜃
= 1.
𝑑𝑡

Find and sketch the solutions satisfying:

0
(1) X(0) = ( )
1
3
(2) X(0) = ( )
0

in rectangular coordinates.

Solution:

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
Theorem: Let 𝑥 = 𝑥(𝑡), 𝑦 = 𝑦(𝑡) be a solution in [0, ∞) to the autonomous system:

𝑑𝑥
= 𝑃(𝑥, 𝑦)
𝑑𝑡

𝑑𝑦
= 𝑄(𝑥, 𝑦)
𝑑𝑡

where 𝑃 and 𝑄 are continuous in the plane.

If the limit lim (𝑥(𝑡), 𝑦(𝑡)) = (𝑥₀, 𝑦₀) exists and is finite, then the point (𝑥₀, 𝑦₀) is a
𝑡→∞
critical point for the system.

Example: Let
2𝑡
𝑥(𝑡) = 𝑒 −𝑡 – 1, 𝑦(𝑡) = + 𝑒 −2𝑡 .
1+𝑡

be a solution to the system: 𝑥′ = 𝑃(𝑥, 𝑦), 𝑦′ = 𝑄(𝑥, 𝑦)

Find the critical point of this system.

★ Stability of Linear Systems

Definition: Consider the following autonomous system:

𝐗 ′ = 𝐹(𝑥, 𝑦) (2.5)

Let 𝐗 ∗ be a critical point to the system, and 𝐗(𝒕) be a solution to it.

(i) 𝐗 ∗ is said to be stable if for every 𝜀 > 0, there exists 𝛿 > 0 such that:

‖𝐗(𝑡) – 𝐗 ∗ ‖ < 𝜀 whenever ‖𝐗(0) – 𝐗 ∗ ‖ < 𝛿, for all 𝑡 ≥ 0.

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
(ii) 𝐗 ∗ is said to be asymptotically stable if it is stable, and there exists 𝛿0 with 𝛿 > 𝛿0 >
0 such that:
If ‖𝐗(0)– 𝐗 ∗ ‖ < 𝛿0 , then lim 𝐗(𝑡) = 𝐗 ∗ .
𝑡→∞

(iii) 𝐗 ∗ is unstable if it is not stable.

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
Definition: A critical point 𝐗 ∗ of an autonomous system is called isolated if there exists a
neighborhood 𝑈 of 𝐗 ∗ such that 𝐗 ∗ is the only critical point of the system in 𝑈.

Example: Show that the following autonomous system has no isolated critical point.

𝑑𝑥 𝑑𝑦
= 𝑦 − 𝑥, = 𝑥 − 𝑦.
𝑑𝑡 𝑑𝑡

Theorem: For the linear autonomous system

𝑎 𝑏
𝐗′ = 𝑨𝐗, where 𝑨 = ( ) is a constant matrix,
𝑐 𝑑
or written as:
𝑥′ = 𝑎𝑥 + 𝑏𝑦
𝑦 ′ = 𝑐𝑥 + 𝑑𝑦,

The following are equivalent:

1. The system has (0, 0) as an isolated critical point.

2. 𝑑𝑒𝑡(𝐴) ≠ 0.
3. The roots 𝜆₁, 𝜆₂ of |𝐴 – 𝜆𝐼| = 0 satisfy 𝜆₁ · 𝜆₂ ≠ 0.

★ Types of Isolated Critical Points:

1. Proper node

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
 2. Improper node

3. Saddle point

4. Spiral point

5. Center point

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
Theorem: Assume the origin (0,0) is an isolated critical point for the plane autonomous
linear system:
𝐗′ = 𝑨𝐗
Then the origin is classified based on the eigenvalues 𝜆1 , 𝜆2 of the constant matrix 𝑨 as
follows:

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
Example: The solution of the system:

𝐗′ = 𝑨𝐗

is given by:

1 −2
𝐗(𝑡) = 𝑐1 𝑒 −3𝑡 ( ) + 𝑐2 𝑒 −6𝑡 ( )
1 1

Classify the critical point at the origin and sketch the phase portrait (trajectories) according
to the eigenvalues.

Example: Classify the critical point at the origin and sketch a phase portrait for the system:

𝑑𝑥 𝑑𝑦
= 5𝑥 − 3𝑦 , = 4𝑥 − 3𝑦.
𝑑𝑡 𝑑𝑡

(Hint: The characteristic equation for this system is 𝜆2 − 2𝜆 − 3 = 0).

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
Example: Classify the critical point at the origin and sketch a phase portrait for the system:

𝑑𝑥 𝑑𝑦
= 𝑥 − 4𝑦 , = 4𝑥 + 𝑦.
𝑑𝑡 𝑑𝑡

Example: Classify the critical point at the origin and sketch a phase portrait for the system:

𝑑𝑥 𝑑𝑦
= 2𝑥 , = 2𝑦.
𝑑𝑡 𝑑𝑡

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
Example: For the following systems, classify the critical point at the origin and sketch the
solution for the system that satisfies X(0) = (1, 0)

𝐗 ′ (𝑡) =A 𝐗(𝑡),

where

3 −18 −1 2
(a) A= [ ] (b) A= [ ].
2 −9 −1 1

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.
Example: Find and classify the critical point of the linear system:

𝑑𝑥 𝑑𝑦
= 2𝑥 + 𝑦 + 3 , = −3𝑥 − 2𝑦 − 4.
𝑑𝑡 𝑑𝑡

This document was prepared by Professor Ahmad El-Ajou of the Mathematics Department at Al-Balqa Applied
University. Printing this document requires prior permission from the author.