See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/330701404

Laplace Transform and Systems of Ordinary Differential Equations

Chapter · January 2019

CITATIONS READS
0 8,566

1 author:

Carlos Enrique Frasser

World Life High School
13 PUBLICATIONS   8 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Is there a finite or an infinite number of biplanes? View project

Characterization, Construction, Enumeration, and Applications of Geodetic Graphs View project

All content following this page was uploaded by Carlos Enrique Frasser on 29 January 2019.

The user has requested enhancement of the downloaded file.

 Laplace Transform and Systems of Ordinary Differential Equations

Carlos E. Frasser

In this chapter, we describe a fundamental study of the Laplace transform, its use in the solution of initial
value problems and some techniques to solve systems of ordinary differential equations (DE) including
their solution with the help of the Laplace transform.

1. Laplace Transform

Improper Integrals. If 𝑡0 ∈ ℝ, we define for any function f (t)

∞ 𝐴
∫ 𝑓(𝑡)𝑑𝑡 = lim ∫ 𝑓(𝑡)𝑑𝑡
𝑡0 𝐴→∞ 𝑡
0

If this limit exists, we say that the improper integral converges, otherwise, diverges.

Example 1.
∞ 𝐴
𝑑𝑥 1 𝐴 1 1 1 1
∫ 5
= lim ∫ 𝑥 −5 𝑑𝑥 = lim − 4 | = lim − ( 4 − 4 ) =
1 𝑥 𝐴→∞ 1 𝐴→∞ 4𝑥 1 𝐴→∞ 4 𝐴 1 4

Initial Function and its Transform. Let f be a function of a single real variable t defined for any t ≥ 0
and let 𝑒 −𝑠𝑡 be a complex function of a real variable t, where s = a + bi, a > 0. Let us examine the product

𝑒 −𝑠𝑡 𝑓(𝑡), (1)

where (1) is also a complex function of a real variable t:

𝑒 −𝑠𝑡 𝑓(𝑡) = 𝑓(𝑡)𝑒 −𝑠𝑡 = 𝑓(𝑡)𝑒 −(𝑎+𝑏𝑖)𝑡 = 𝑓(𝑡)𝑒 −𝑎𝑡 𝑒 −𝑖𝑏𝑡 = 𝑓(𝑡)𝑒 −𝑎𝑡 (cos 𝑏𝑡 − 𝑖 sin 𝑏𝑡)
= 𝑓(𝑡)𝑒 −𝑎𝑡 cos 𝑏𝑡 − 𝑖 𝑓(𝑡)𝑒 −𝑎𝑡 sin 𝑏𝑡 . (2)

From (2), we get

∞ ∞ ∞
∫ 𝑒 −𝑠𝑡 𝑓(𝑡)𝑑𝑡 = ∫ 𝑓(𝑡)𝑒 −𝑎𝑡 cos 𝑏𝑡 𝑑𝑡 − 𝑖 ∫ 𝑓(𝑡)𝑒 −𝑎𝑡 sin 𝑏𝑡 𝑑𝑡 . (3)
0 0 0

The left side of (3) determines a function of s, which is called F(s). So

∞
𝐹(𝑠) = ∫ 𝑒 −𝑠𝑡 𝑓(𝑡) 𝑑𝑡 . (4)
0

Function (4) is called the Laplace transform or briefly, ℒ-transform, and function f (t) is called its initial
function. If F(s) is the ℒ-transform of function f (t), then we write

ℒ{𝑓(𝑡)} = 𝐹(𝑠). (5)

A function f is said to be of exponential order on the interval [0, +∞) if there exist constants C and  such
that

1
 |𝑓(𝑡)| ≤ 𝐶𝑒 𝛼𝑡 . (6)

Theorem 1. If, according to (6), f is of exponential order and a >, then the Laplace transform
∞
ℒ{𝑓(𝑡)} = 𝐹(𝑠) = ∫ 𝑒 −𝑠𝑡 𝑓(𝑡) 𝑑𝑡
0

does exist.

Proof. This theorem will be true if we are able to prove that each of the integrals on the right side of (3)
exists. Let us evaluate the first of these two integrals:
∞ ∞
∫ 𝑓(𝑡)𝑒 −𝑎𝑡 cos 𝑏𝑡 𝑑𝑡 ≤ |∫ 𝑓(𝑡)𝑒 −𝑎𝑡 cos 𝑏𝑡 𝑑𝑡|
0 0
∞ ∞
≤ ∫ |𝑓(𝑡)𝑒 −𝑎𝑡 cos 𝑏𝑡|𝑑𝑡 ≤ ∫ |𝑓(𝑡)𝑒 −𝑎𝑡 |𝑑𝑡
0 0
∞
≤ ∫ 𝐶𝑒 𝛼𝑡 𝑒 −𝑎𝑡 𝑑𝑡
0
∞ 𝐴
1 𝐴
= 𝐶 ∫ 𝑒 (𝛼−𝑎)𝑡 𝑑𝑡 = 𝐶 lim ∫ 𝑒 (𝛼−𝑎)𝑡 𝑑𝑡 = 𝐶 lim 𝑒 (𝛼−𝑎)𝑡 |
0 𝐴→∞ 0 𝐴→∞ 𝛼 − 𝑎 0
1 𝑐 𝐶
= 𝐶 lim [𝑒 (𝛼−𝑎)𝐴 − 1] = lim [𝑒 −(𝑎−𝛼)𝐴 − 1] = .
𝐴→∞ 𝛼 − 𝑎 𝛼 − 𝑎 𝐴→∞ 𝑎−𝛼
Therefore
∞
𝐶
∫ 𝑓(𝑡)𝑒 −𝑎𝑡 cos 𝑏𝑡 𝑑𝑡 ≤ . (7)
0 𝑎−𝛼

Let us evaluate the second integral:

∞ ∞ ∞ ∞
∫ 𝑓(𝑡)𝑒 −𝑎𝑡 sin 𝑏𝑡 𝑑𝑡 ≤ |∫ 𝑓(𝑡)𝑒 −𝑎𝑡 sin 𝑏𝑡 𝑑𝑡| ≤ ∫ |𝑓(𝑡)𝑒 −𝑎𝑡 sin 𝑏𝑡|𝑑𝑡 ≤ ∫ |𝑓(𝑡)𝑒 −𝑎𝑡 |𝑑𝑡
0 0 0 0
∞
𝐶
≤ ∫ 𝐶𝑒 𝛼𝑡 𝑒 −𝑎𝑡 𝑑𝑡 = . (8)
0 𝑎−𝛼

Thus, both integrals (7) and (8) do exist. We can then conclude that
∞
∫ 𝑒 −𝑠𝑡 𝑓(𝑡)𝑑𝑡
0

also exists and so Theorem 1 is proved. □

𝓛-transforms of functions 𝐬𝐢𝐧 𝒕 𝐚𝐧𝐝 𝐜𝐨𝐬 𝒕

I. If 𝑓(𝑡) = sin 𝑡, then

2
 ∞ 𝐴
1 cos 𝐴 𝑠 sin 𝐴
ℒ{sin 𝑡} = ∫ 𝑒 −𝑠𝑡 sin 𝑡 𝑑𝑡 = lim ∫ 𝑒 −𝑠𝑡 sin 𝑡 𝑑𝑡 = lim 2
(1 − 𝐴𝑠 − 𝐴𝑠 )
0 𝐴→∞ 0 𝐴→∞ 1+𝑠 𝑒 𝑒
1
= . (9)
𝑠2 +1

II. If 𝑓(𝑡) = cos 𝑡, then

∞ 𝐴
1 sin 𝐴 𝑠 cos 𝐴
ℒ{cos 𝑡} = ∫ 𝑒 −𝑠𝑡 cos 𝑡 𝑑𝑡 = lim ∫ 𝑒 −𝑠𝑡 cos 𝑡 𝑑𝑡 = lim 2
(𝑠 + 𝐴𝑠 − 𝐴𝑠 )
0 𝐴→∞ 0 𝐴→∞ 1+𝑠 𝑒 𝑒
𝑠
= . (10)
𝑠2 + 1

Change of Scale Property

Theorem 2. The ℒ-transform of f (at), a > 0, is

1 𝑠
ℒ{𝑓(𝑎𝑡)} = 𝐹 ( ). (11)
𝑎 𝑎
Proof. We know that

∞
ℒ{𝑓(𝑎𝑡)} = ∫ 𝑒 −𝑠𝑡 𝑓(𝑎𝑡)𝑑𝑡.
0

1
Making 𝑧 = 𝑎𝑡 ⇒ 𝑎
𝑑𝑧 = 𝑑𝑡, then

∞ 𝑠𝑧 𝑑𝑧 1 ∞ − 𝑠 𝑧 1 𝑠
ℒ{𝑓(𝑎𝑡)} = ∫ 𝑒 − 𝑎 𝑓(𝑧) = ∫ 𝑒 𝑎 𝑓(𝑧)𝑑𝑧 = 𝐹 ( ) . □
0 𝑎 𝑎 0 𝑎 𝑎

Example 2. Find the ℒ-transform of 𝑓(𝑎𝑡) = sin 𝑎𝑡.

If 𝑓(𝑧) = sin 𝑧, then, according to (9),

∞
𝑠 𝑠 1 𝑎2
𝐹 ( ) = ∫ 𝑒 − 𝑎𝑧 sin 𝑧 𝑑𝑧 = =
𝑎 0 𝑠 2 𝑠 2 + 𝑎2
(𝑎) + 1

and according to (11), the ℒ-transform of 𝑓(𝑎𝑡) = sin 𝑎𝑡 is

1 𝑠 1 𝑎2 𝑎
ℒ{sin 𝑎𝑡} = 𝐹( ) = 2 2
= 2 . (12)
𝑎 𝑎 𝑎 𝑠 +𝑎 𝑠 + 𝑎2

Example 3. Find the ℒ-transform of 𝑓(𝑎𝑡) = cos 𝑎𝑡.

If 𝑓(𝑧) = cos 𝑧, then, according to (10),

3
 ∞
𝑠
𝑠 𝑠
𝑎 𝑎𝑠
𝐹 ( ) = ∫ 𝑒 − 𝑎𝑧 cos 𝑧 𝑑𝑧 = 2 =
𝑎 0 𝑠 𝑠2 + 𝑎2
(𝑎) + 1

and according to (11), the ℒ-transform of 𝑓(𝑎𝑡) = cos 𝑎𝑡 is

1 𝑠 1 𝑎𝑠 𝑠
ℒ{cos 𝑎𝑡} = 𝐹( ) = 2 2
= 2 . (13)
𝑎 𝑎 𝑎 𝑠 +𝑎 𝑠 + 𝑎2

𝓛-transforms of functions 𝒇(𝒕) = 𝟏, 𝒇(𝒕) = 𝒕𝒏

I. If f (t) = 1, then
∞ ∞ 𝐴
1 𝐴 1
ℒ{1} = ∫ 𝑒 −𝑠𝑡 (1)𝑑𝑡 = ∫ 𝑒 −𝑠𝑡 𝑑𝑡 = lim ∫ 𝑒 −𝑠𝑡 𝑑𝑡 = lim − 𝑒 −𝑠𝑡 | = lim − (𝑒 −𝐴𝑡 − 𝑒 0 )
0 0 𝐴→∞ 0 𝐴→∞ 𝑠 0 𝐴→∞ 𝑠
1 1 1
= lim (1 − 𝐴𝑡 ) = . (14)
𝐴→∞ 𝑠 𝑒 𝑠

II. If f (t) = tn, then, by using integration by parts successively, we have

∞ 𝐴
𝑛! 𝐴 −𝑠𝑡 𝑛! 1 𝐴
ℒ{𝑡 𝑛 } = ∫ 𝑒 −𝑠𝑡 𝑡 𝑛 𝑑𝑡 = lim ∫ 𝑒 −𝑠𝑡 𝑡 𝑛 𝑑𝑡 = lim 𝑛
∫ 𝑒 𝑑𝑡 = lim 𝑛 (− 𝑒 −𝑠𝑡 ) |
0 𝐴→∞ 0 𝐴→∞ 𝑠 0 𝐴→∞ 𝑠 𝑠 0
𝑛! 𝑛! 1 𝑛!
= lim − 𝑛 (𝑒 −𝐴𝑠 − 𝑒 0 ) = 𝑛+1 lim (1 − 𝐴𝑠 ) = 𝑛+1 . (15)
𝐴→∞ 𝑠 (𝑠) 𝑠 𝐴→∞ 𝑒 𝑠

Linearity Property

Theorem 3. If a and b are constants, then

ℒ{𝑎𝑓 + 𝑏𝑔} = 𝑎ℒ{𝑓} + 𝑏ℒ{𝑔}. (16)

Proof. Applying the definition of the Laplace transform, we get

∞ ∞ ∞ ∞ ∞
ℒ{𝑎𝑓 + 𝑏𝑔} = ∫ 𝑒 −𝑠𝑡 (𝑎𝑓 + 𝑏𝑔)𝑑𝑡 = ∫ 𝑒 −𝑠𝑡 𝑎𝑓𝑑𝑡 + ∫ 𝑒 −𝑠𝑡 𝑏𝑔𝑑𝑡 = 𝑎 ∫ 𝑒 −𝑠𝑡 𝑓𝑑𝑡 + 𝑏 ∫ 𝑒 −𝑠𝑡 𝑔𝑑𝑡
0 0 0 0 0
= 𝑎ℒ{𝑓} + 𝑏ℒ{𝑔}. □

Example 4. Find the ℒ-transform of 𝑓(𝑡) = 𝑡 2 + 8𝑡 − 16.

2! 1! 1 2 8 16
ℒ{𝑡 2 + 8𝑡 − 16} = ℒ{𝑡 2 } + 8ℒ{𝑡} − 16ℒ{1} = + 8 ( 1+1 ) − 16 ( ) = 3 + 2 − .
𝑠 2+1 𝑠 𝑠 𝑠 𝑠 𝑠

Shift Theorem (or Shifting Property)

Theorem 4. If F(s) is the ℒ-transform of 𝑓(𝑡), then F(s + ) is the ℒ-transform of 𝑒 −𝛼𝑡 𝑓(𝑡), that is,

if ℒ{𝑓(𝑡)} = 𝐹(𝑠), then

ℒ{𝑒 −𝛼𝑡 𝑓(𝑡)} = 𝐹(𝑠 + ). (17)

4
Proof.
∞ ∞
𝓛{𝑒 −𝛼𝑡 𝑓(𝑡)} = ∫ 𝑒 −𝑠𝑡 𝑒 −𝛼𝑡 𝑓(𝑡)𝑑𝑡 = ∫ 𝑒 −(𝑠+)𝑡 𝑓(𝑡)𝑑𝑡 = 𝐹(𝑠 + ). □
0 0

𝓛-transforms of functions 𝒆−𝜶𝒕 , 𝒆−𝜶𝒕 𝐬𝐢𝐧 𝒂𝒕, 𝒆−𝜶𝒕 𝐜𝐨𝐬 𝒂𝒕

I. Find the ℒ-transform of 𝑒 −𝛼𝑡 .

1
From (17), making f (t) = 1 and keeping in mind that ℒ{𝑓(𝑡)} = ℒ{1} = = 𝐹(𝑠) , we get
𝑠

1
ℒ{𝑒 −𝛼𝑡 } = ℒ{𝑒 −𝛼𝑡 (1)} = ℒ{𝑒 −𝛼𝑡 𝑓(𝑡)} = 𝐹(𝑠 + ) = . (18)
𝑠+

II. Find the ℒ-transform of 𝑒 −𝛼𝑡 sin 𝑎𝑡.

𝑎
From (17), making f (t) = sin at and keeping in mind that ℒ{𝑓(𝑡)} = ℒ{sin 𝑎𝑡} = = 𝐹(𝑠) , we get
𝑠 2 + 𝑎2

𝑎
ℒ{𝑒 −𝛼𝑡 sin 𝑎𝑡} = ℒ{𝑒 −𝛼𝑡 𝑓(𝑡)} = 𝐹(𝑠 + ) = . (19)
(𝑠 + )2 + 𝑎2

III. Find the ℒ-transform of 𝑒 −𝛼𝑡 cos 𝑎𝑡.

𝑠
From (17), making f (t) = cos at and keeping in mind that ℒ{𝑓(𝑡)} = ℒ{cos 𝑎𝑡} = = 𝐹(𝑠) , we get
𝑠2 + 𝑎 2

𝑠+
ℒ{𝑒 −𝛼𝑡 cos 𝑎𝑡} = ℒ{𝑒 −𝛼𝑡 𝑓(𝑡)} = 𝐹(𝑠 + ) = . (20)
(𝑠 + )2 + 𝑎2

N-th Order Derivative Property

Theorem 5. If f (t) is of exponential order and F(s) is the ℒ-transform of f, then

𝑑𝑛
(−1)𝑛 ℒ{𝑓(𝑡)} = ℒ{𝑡 𝑛 𝑓(𝑡)}. (21)
𝑑𝑠 𝑛
Proof. Take the derivative of both sides of the equation
∞
ℒ{𝑓(𝑡)} = ∫ 𝑒 −𝑠𝑡 𝑓(𝑡) 𝑑𝑡
0

n times with respect to s.

i)

𝑑 𝑑 ∞ −𝑠𝑡 ∞
𝜕 −𝑠𝑡 ∞
ℒ{𝑓(𝑡)} = ∫ 𝑒 𝑓(𝑡) 𝑑𝑡 = ∫ [𝑒 𝑓(𝑡)]𝑑𝑡 = − ∫ 𝑒 −𝑠𝑡 𝑡𝑓(𝑡)𝑑𝑡,
𝑑𝑠 𝑑𝑠 0 0 𝜕𝑠 0

which means,

5
 𝑑
− ℒ{𝑓(𝑡)} = ℒ{𝑡𝑓(𝑡)}
𝑑𝑠
ii)

𝑑2 ∞
𝜕 −𝑠𝑡 ∞
ℒ{𝑓(𝑡)} = − ∫ [𝑒 𝑡𝑓(𝑡)]𝑑𝑡 = ∫ 𝑒 −𝑠𝑡 𝑡 2 𝑓(𝑡) 𝑑𝑡 = ℒ{𝑡 2 𝑓(𝑡)}
𝑑𝑠 2 0 𝜕𝑠 0

iii)

𝑑3 ∞
𝜕 −𝑠𝑡 2 ∞
ℒ{𝑓(𝑡)} = ∫ [𝑒 𝑡 𝑓(𝑡)]𝑑𝑡 = − ∫ 𝑒 −𝑠𝑡 𝑡 3 𝑓(𝑡) 𝑑𝑡,
𝑑𝑠 3 0 𝜕𝑠 0

which means,

𝑑3
− ℒ{𝑓(𝑡)} = ℒ{𝑡 3 𝑓(𝑡)}
𝑑𝑠 3
and so on. Thus, it follows that

𝑑𝑛
(−1)𝑛 ℒ{𝑓(𝑡)} = ℒ{𝑡 𝑛 𝑓(𝑡)}. □
𝑑𝑠 𝑛

𝓛-transforms of functions 𝒕 𝐬𝐢𝐧 𝒂𝒕 𝐚𝐧𝐝 𝒕 𝐜𝐨𝐬 𝒂𝒕

I. Find the ℒ-transform of 𝑡 sin 𝑎𝑡.

From (21), making f (t) = sin at and keeping in mind (12), we obtain

𝑑 𝑑 𝑑 𝑎 2𝑎𝑠
ℒ{𝑡 sin 𝑎𝑡} = ℒ{𝑡𝑓(𝑡)} = − ℒ{𝑓(𝑡)} = − ℒ{sin 𝑎𝑡} = − ( 2 2
) = − [− 2 ]
𝑑𝑠 𝑑𝑠 𝑑𝑠 𝑠 + 𝑎 (𝑠 + 𝑎2 )2
2𝑎𝑠
= . (22)
(𝑠 2 + 𝑎2 )2

II. Find the ℒ-transform of 𝑡 cos 𝑎𝑡.

From (21), making f (t) = cos at and keeping in mind (13), we obtain

𝑑 𝑑 𝑑 𝑠 𝑎2 − 𝑠 2
ℒ{𝑡 cos 𝑎𝑡} = ℒ{𝑡𝑓(𝑡)} = − ℒ{𝑓(𝑡)} = − ℒ{cos 𝑎𝑡} = − ( 2 ) = −[ 2 ]
𝑑𝑠 𝑑𝑠 𝑑𝑠 𝑠 + 𝑎2 (𝑠 + 𝑎2 )2
𝑠 2 − 𝑎2
= 2 . (23)
(𝑠 + 𝑎2 )2

Laplace Transform of derivatives

Theorem 6. If f (t) is a function of exponential order, then

ℒ{𝑓′(𝑡)} = 𝑠ℒ{𝑓(𝑡)} − 𝑓(0). (24)

6
Proof.
∞
ℒ{𝑓′(𝑡)} = ∫ 𝑒 −𝑠𝑡 𝑓 ′ (𝑡)𝑑𝑡
0
𝐴
= lim ∫ 𝑒 −𝑠𝑡 𝑓 ′ (𝑡)𝑑𝑡
𝐴→∞ 0
𝐴
𝐴
= lim [𝑒 −𝑠𝑡 𝑓(𝑡) | + ∫ 𝑠𝑒 −𝑠𝑡 𝑓(𝑡)𝑑𝑡]
𝐴→∞ 0 0
𝐴 ∞
𝑓(𝐴)
= lim [ 𝐴𝑠 − 𝑓(0)] + 𝑠 lim ∫ 𝑒 −𝑠𝑡 𝑓(𝑡)𝑑𝑡 = 𝑠 ∫ 𝑒 −𝑠𝑡 𝑓(𝑡)𝑑𝑡 − 𝑓(0)
𝐴→∞ 𝑒 𝐴→∞ 0 0
= 𝑠ℒ{𝑓(𝑡)} − 𝑓(0). □

Remark 1.

a)

ℒ{𝑓′′(𝑡)} = 𝑠ℒ{𝑓′(𝑡)} − 𝑓 ′ (0) = 𝑠[𝑠ℒ{𝑓(𝑡)} − 𝑓(0)] − 𝑓 ′ (0) = 𝑠 2 ℒ{𝑓(𝑡)} − 𝑠𝑓(0) − 𝑓 ′ (0).

b)

ℒ{𝑓′′′(𝑡)} = 𝑠ℒ{𝑓′′(𝑡)} − 𝑓 ′′ (0) = 𝑠[𝑠 2 ℒ{𝑓(𝑡)} − 𝑠𝑓(0) − 𝑓 ′ (0)] − 𝑓 ′ ′(0)

= 𝑠 3 ℒ{𝑓(𝑡)} − 𝑠 2 𝑓(0) − 𝑠𝑓 ′ (0) − 𝑓 ′ ′(0).

c)

ℒ{𝑓 (4) (𝑡)} = 𝑠ℒ{𝑓′′′(𝑡)} − 𝑓 ′′′ (0) = 𝑠[𝑠 3 ℒ{𝑓(𝑡)} − 𝑠 2 𝑓(0) − 𝑠𝑓 ′ (0) − 𝑓 ′ ′(0)] − 𝑓 ′′′ (0)
= 𝑠 4 ℒ{𝑓(𝑡)} − 𝑠 3 𝑓(0) − 𝑠 2 𝑓 ′ (0) − 𝑠𝑓 ′′ (0) − 𝑓 ′′′ (0).

And so on. In general,

d)

ℒ{𝑓 (𝑛) (𝑡)} = 𝑠 𝑛 ℒ{𝑓(𝑡)} − 𝑠 𝑛−1 𝑓(0) − 𝑠 𝑛−2 𝑓 ′ (0) − 𝑠 𝑛−3 𝑓 ′′ (0) − ⋯ − 𝑓 (𝑛−1) (0). (25)

Inverse Laplace Transform

The original function f (t) is called the inverse ℒ-transform of F(s) = ℒ{f(t)} and it is designated by

𝑓(𝑡) = ℒ −1 {ℒ{𝑓(𝑡)}}

2
Example 5. Find ℒ −1 {ℒ{𝑓(𝑡)}} = ℒ −1 {𝑠2 }

Keeping in mind (15), we obtain

2 1! 1!
𝑓(𝑡) = ℒ −1 { 2 } = ℒ −1 {2 ( 1+1 )} = 2 ℒ −1 { 1+1 } = 2𝑡1 = 2𝑡
𝑠 𝑠 𝑠

7
 Table 1. Some ℒ-transforms

∞
 f (t)
ℒ{𝑓(𝑡)} = 𝐹(𝑠) = ∫ 𝑒 −𝑠𝑡 𝑓(𝑡) 𝑑𝑡
0

1. 1 1
𝑠

2. 𝑎 sin at
𝑠2 + 𝑎2

3. 𝑠 cos at
𝑠2 + 𝑎2

4. 2𝑎𝑠 𝑡 sin 𝑎𝑡
(𝑠 2 + 𝑎2 )2

5. 𝑠 2 − 𝑎2 𝑡 cos 𝑎𝑡
(𝑠 2 + 𝑎2 )2

6. 1 𝑒 ±𝛼𝑡
𝑠∓𝛼

7. 1 𝑡𝑒 ±𝛼𝑡
(𝑠 ∓ 𝛼)2

8. 𝑎 𝑒 ±𝛼𝑡 sin 𝑎𝑡
(𝑠 ∓ 𝛼)2 + 𝑎2

9. 𝑠∓𝛼 𝑒 ±𝛼𝑡 cos 𝑎𝑡

(𝑠 ∓ 𝛼)2 + 𝑎2

10. 𝑎 𝑒 ±𝛼𝑡 sinh 𝑎𝑡

(𝑠 ∓ 𝛼)2 − 𝑎2

11. 𝑠∓𝛼 𝑒 ±𝛼𝑡 cosh 𝑎𝑡

(𝑠 ∓ 𝛼)2 − 𝑎2

12. 𝑛! 𝑡𝑛
𝑠 𝑛+1

13. 𝑑𝑛 𝑡 𝑛 𝑓(𝑡)
(−1)𝑛 𝐹(𝑠)
𝑑𝑠 𝑛

8
 1
Example 6. Find ℒ −1 { }.
𝑠2 +4

Keeping in mind (12), we get

1 1 2 1 2 1
𝑓(𝑡) = ℒ −1 { } = ℒ −1 { ( 2 )} = ℒ −1 { 2 } = sin 2𝑡
𝑠2 +4 2 𝑠 +2 2 2 𝑠 +2 2 2

Solution of Initial Value Problems Using the Laplace Transform

Example 7. Solve the initial value problem

𝑦 ′′ + 2𝑦 ′ + 5𝑦 = 0, 𝑦(0) = 𝑦 ′ (0) = 1.

Let us apply the Laplace transform to both sides of the given DE

ℒ{𝑦 ′′ + 2𝑦 ′ + 5𝑦} = ℒ{0}

ℒ{𝑦 ′′ } + 2ℒ{𝑦 ′ } + 5ℒ{𝑦} = 0

𝑠 2 ℒ{𝑦} − 𝑠𝑦(0) − 𝑦 ′ (0) + 2[𝑠ℒ{𝑦} − 𝑦(0)] + 5ℒ{𝑦} = 0

𝑠 2 ℒ{𝑦} − 𝑠 − 1 + 2𝑠ℒ{𝑦} − 2 + 5ℒ{𝑦} = 0

(𝑠 2 + 2𝑠 + 5)ℒ{𝑦} = 𝑠 + 3

𝑠+3
ℒ{𝑦} =
𝑠2 + 2𝑠 + 5
𝑠+3 𝑠+1 2
ℒ{𝑦} = 2
= 2 2
+
(𝑠 + 1) + 4 (𝑠 + 1) + 2 (𝑠 + 1)2 + 22

𝑠+1 2
𝑦 = ℒ −1 {ℒ{𝑦}} = ℒ −1 { 2 2 } + ℒ −1 { }
(𝑠 + 1) + 2 (𝑠 + 1)2 + 22

𝑦 = 𝑒 −𝑡 cos 2𝑡 + 𝑒 −𝑡 sin 2𝑡

𝑦 = 𝑒 −𝑡 (cos 2𝑡 + sin 2𝑡)

Example 8. Solve the initial value problem

𝑦 ′′ − 2𝑦 ′ − 3𝑦 = 5, 𝑦(0) = 0, 𝑦 ′ (0) = 1.

Let us apply the Laplace transform to both sides of the given DE

ℒ{𝑦 ′′ − 2𝑦 ′ − 3𝑦} = ℒ{5}

ℒ{𝑦 ′′ } − 2ℒ{𝑦 ′ } − 3ℒ{𝑦} = ℒ{5(1)}

𝑠 2 ℒ{𝑦} − 𝑠𝑦(0) − 𝑦 ′ (0) − 2[𝑠ℒ{𝑦} − 𝑦(0)] − 3ℒ{𝑦} = 5ℒ{1}

9
 1
𝑠 2 ℒ{𝑦} − 1 − 2𝑠ℒ{𝑦} − 3ℒ{𝑦} = 5 ( )
𝑠
5
(𝑠 2 − 2𝑠 − 3)ℒ{𝑦} = +1
𝑠
𝑠+5
ℒ{𝑦} = .
𝑠(𝑠 2 − 2𝑠 − 3)

We need to use the partial fraction expansion of a rational expression

𝑠+5 𝐴 𝐵𝑠 + 𝐶
= + 2
𝑠(𝑠 2 − 2𝑠 − 3) 𝑠 𝑠 − 2𝑠 − 3

𝑠 + 5 = 𝐴(𝑠 2 − 2𝑠 − 3) + (𝐵𝑠 + 𝐶)𝑠

𝑠 + 5 = (𝐴 + 𝐵)𝑠 2 + (−2𝐴 + 𝐶)𝑠 − 3𝐴.

Let us solve the system of linear equations

𝐴+𝐵 =0
−2𝐴 +𝐶 =1}
−3𝐴 =5
whose solution is

5 5 7
𝐴= − ,𝐵 = ,𝐶 = − ,
3 3 3
then

𝑠+5 1 5 5𝑠 − 7
ℒ{𝑦} = = (− + 2 )
𝑠(𝑠 2 − 2𝑠 − 3) 3 𝑠 𝑠 − 2𝑠 − 3

1 1 5𝑠 − 5 2
𝑦 = ℒ −1 {ℒ{𝑦}} = [−5ℒ −1 { } + ℒ −1 { − }]
3 𝑠 (𝑠 − 1)2 − 22 (𝑠 − 1)2 − 22

1 𝑠−1 2
𝑦 = [−5 + 5ℒ −1 { 2 2 } − ℒ −1 { }]
3 (𝑠 − 1) − 2 (𝑠 − 1)2 − 22

1
𝑦 = (−5 + 5𝑒 𝑡 cosh 2𝑡 − 𝑒 𝑡 sinh 2𝑡)
3

2. Systems of Ordinary DE with Constant Coefficients

Given the set of numbers aij and functions fi (t), the expression
𝑛
𝑑𝑦𝑖
= ∑ 𝑎𝑖𝑗 𝑦𝑗 + 𝑓𝑖 (𝑡) (𝑖 = 1,2, … , 𝑛) (1)
𝑑𝑡
𝑗=1

10
which is equivalent to

𝑑𝑦1
= 𝑎11 𝑦1 + 𝑎12 𝑦2 + ⋯ + 𝑎1𝑛 𝑦𝑛 + 𝑓1 (𝑡)
𝑑𝑡
𝑑𝑦2
= 𝑎21 𝑦1 + 𝑎22 𝑦2 + ⋯ + 𝑎2𝑛 𝑦𝑛 + 𝑓2 (𝑡)
𝑑𝑡
… (1A)

𝑑𝑦𝑛
= 𝑎𝑛1 𝑦1 + 𝑎𝑛2 𝑦2 + ⋯ + 𝑎𝑛𝑛 𝑦𝑛 + 𝑓𝑛 (𝑡)
𝑑𝑡
is called a system of linear nonhomogeneous DE with constant coefficients.

On the other hand, (1A) is called a system of linear homogeneous DE with constant coefficients if all of
the functions fi (t) are zero.

If the set of functions

𝑦1 = 𝜑1 (𝑡)
𝑦2 = 𝜑2 (𝑡)

… (2)

𝑦𝑛 = 𝜑𝑛 (𝑡)

with continuous derivatives on the interval (a, b) convert the equations of system (1A) into identities for
each value of t  (a, b), then (2) is called a solution of system (1A).

The simplest technique to solve system (1A) consists in reducing it to an n-th order DE. Let us illustrate
this method in the example of a system of two DE

𝑑𝑦1
= 𝑎𝑦1 + 𝑏𝑦2 + 𝑓(𝑡) (3𝐴)
𝑑𝑡 } (3)
𝑑𝑦2
= 𝑐𝑦1 + 𝑑𝑦2 + 𝑔(𝑡) (3𝐵)
𝑑𝑡
where a, b, c, d are constant coefficients, f(t), g(t) are given functions and y1(t), y2(t) are the unknown
functions. From (3A), we get

𝑑𝑦1
𝑏𝑦2 = − 𝑎𝑦1 − 𝑓(𝑡)
𝑑𝑡
1 𝑑𝑦1
𝑦2 = [ − 𝑎𝑦1 − 𝑓(𝑡)] (4)
𝑏 𝑑𝑡

Thus, we obtain from (4)

𝑑𝑦2 1 𝑑2 𝑦1 𝑑𝑦1
= [ 2 −𝑎 − 𝑓′(𝑡)] (5)
𝑑𝑡 𝑏 𝑑𝑡 𝑑𝑡

11
Substituting (4) into (3B), we get

𝑑𝑦2 𝑑 𝑑𝑦1
= 𝑐𝑦1 + [ − 𝑎𝑦1 − 𝑓(𝑡)] + 𝑔(𝑡) (6)
𝑑𝑡 𝑏 𝑑𝑡

Equating (5) and (6), we obtain

1 𝑑 2 𝑦1 𝑑𝑦1 𝑑 𝑑𝑦1
[ 2 −𝑎 − 𝑓′(𝑡)] = 𝑐𝑦1 + [ − 𝑎𝑦1 − 𝑓(𝑡)] + 𝑔(𝑡)
𝑏 𝑑𝑡 𝑑𝑡 𝑏 𝑑𝑡

1 𝑑2 𝑦1 𝑎 + 𝑑 𝑑𝑦1 𝑎𝑑 1 𝑑
2
−( ) + ( − 𝑐) 𝑦1 + [− 𝑓 ′ (𝑡) + 𝑓(𝑡) − 𝑔(𝑡)] = 0
𝑏 𝑑𝑡 𝑏 𝑑𝑡 𝑏 𝑏 𝑏

and so, we finally get

𝑑2 𝑦1 𝑑𝑦1
𝐴 2
+𝐵 + 𝐶𝑦1 + 𝑃(𝑡) = 0
𝑑𝑡 𝑑𝑡

where A, B, C are constants. From the last DE, we find 𝑦1 = 𝑦1 (𝑡, 𝐶1 , 𝐶2 ), and substituting y1 and dy1/dt
into (4), we get y2.

Example 1. Find the general solution, or briefly, integrate the following system

𝑑𝑦1
= 4𝑦1 − 𝑦2 + (𝑡 + 1) (7𝐴)
𝑑𝑡 } (7)
𝑑𝑦2
= 2𝑦1 + 𝑦2 + (𝑡 − 1) (7𝐵)
𝑑𝑡
From (7A), we get

𝑑𝑦1
𝑦2 = − + 4𝑦1 + (𝑡 + 1) (8)
𝑑𝑡

𝑑𝑦2 𝑑2 𝑦1 𝑑𝑦1
= − 2 +4 + 1. (9)
𝑑𝑡 𝑑𝑡 𝑑𝑡
Substituting (8) into (7B), we obtain

𝑑𝑦2 𝑑𝑦1
= 2𝑦1 + [− + 4𝑦1 + (𝑡 + 1)] + (𝑡 − 1)
𝑑𝑡 𝑑𝑡
𝑑𝑦2 𝑑𝑦1
= 6𝑦1 − + 2𝑡. (10)
𝑑𝑡 𝑑𝑡
By equating (9) and (10), we have

𝑑2 𝑦1 𝑑𝑦1 𝑑𝑦1
− 2 +4 + 1 = 6𝑦1 − + 2𝑡
𝑑𝑡 𝑑𝑡 𝑑𝑡

𝑦1′′ − 5𝑦1′ + 6𝑦1 = 1 − 2𝑡. (11)

12
Let us find the general solution of (11). See article [1]. But first, let us find the solution of the
corresponding linear homogeneous DE.

𝐷 2 𝑦1 − 5𝐷𝑦1 + 6𝑦1 = 0

(𝐷 2 − 5𝐷 + 6)𝑦1 = 0

𝑟 2 − 5𝑟 + 6 = 0 (12)

(𝑟 − 2)(𝑟 − 3) = 0

𝑟1 = 2, 𝑟2 = 3.

Therefore, the solution of the corresponding linear homogeneous DE is

𝑦1ℎ = 𝐶1 𝑒 2𝑡 + 𝐶2 𝑒 3𝑡 (13)

Now, let us find a particular solution of the linear nonhomogeneous DE (11) by using the method of
undetermined coefficients. See [1].  = 0 is not a root of the characteristic equation (12). Therefore,
1
0𝑥
𝑦1𝑝 = 𝑒 ∑ 𝐴𝑘 𝑡 𝑘 = 𝐴0 + 𝐴1 𝑡. (14)
𝑘=0

From (14), we get

′ ′′
𝑦1𝑝 = 𝐴1 , 𝑦1𝑝 = 0.

From (11), we obtain

0 − 5𝐴1 + 6(𝐴0 + 𝐴1 𝑡) = 1 − 2𝑡

6𝐴0 − 5𝐴1 + 6𝐴1 𝑡 = 1 − 2𝑡

6𝐴0 − 5𝐴1 = 1
}
6𝐴1 = −2

1 1
𝐴0 = − , 𝐴1 = − .
9 3
By substituting the values of A0 and A1 into (14), we have

𝑡 1
𝑦1𝑝 = − −
3 9
Consequently

𝑡 1
𝑦1 = 𝑦1ℎ + 𝑦1𝑝 = 𝐶1 𝑒 2𝑡 + 𝐶2 𝑒 3𝑡 − − . (15)
3 9
Let us get the derivative of (15)

13
 𝑑𝑦1 1
= 2𝐶1 𝑒 2𝑡 + 3𝐶2 𝑒 3𝑡 − . (16)
𝑑𝑡 3
By substituting (15) and (16) into (8), we have

1 𝑡 1
𝑦2 = − (2𝐶1 𝑒 2𝑡 + 3𝐶2 𝑒 3𝑡 − ) + 4 (𝐶1 𝑒 2𝑡 + 𝐶2 𝑒 3𝑡 − − ) + (𝑡 + 1)
3 3 9
𝑡 8
𝑦2 = 2𝐶1 𝑒 2𝑡 + 𝐶2 𝑒 3𝑡 − + .
3 9

Euler’s Method for Integration of a System of Three Linear Homogeneous DE with Constant
Coefficients

Let us consider the system

𝑑𝑦1
= 𝑎𝑦1 + 𝑏𝑦2 + 𝑐𝑦3
𝑑𝑡
𝑑𝑦2
= 𝑎1 𝑦1 + 𝑏1 𝑦2 + 𝑐1 𝑦3 (17)
𝑑𝑡
𝑑𝑦3
= 𝑎2 𝑦1 + 𝑏2 𝑦2 + 𝑐2 𝑦3
𝑑𝑡
Let us assume that the solution of system (17) can be written in the form

𝑦1 = 𝜆𝑒 𝑟𝑡 , 𝑦2 = 𝜇𝑒 𝑟𝑡 , 𝑦3 = 𝜈𝑒 𝑟𝑡 (18)

where 𝜆, 𝜇, 𝜈, 𝑟 are constants. From (18), we get

𝑑𝑦1 𝑑𝑦2 𝑑𝑦3

= 𝜆𝑟𝑒 𝑟𝑡 , = 𝜇𝑟𝑒 𝑟𝑡 , = 𝜈𝑟𝑒 𝑟𝑡 . (19)
𝑑𝑡 𝑑𝑡 𝑑𝑡
By substituting (18), (19) into (17), we have

𝜆𝑟𝑒 𝑟𝑡 = 𝑎𝜆𝑒 𝑟𝑡 + 𝑏𝜇𝑒 𝑟𝑡 + 𝑐𝜈𝑒 𝑟𝑡

𝜇𝑟𝑒 𝑟𝑡 = 𝑎1 𝜆𝑒 𝑟𝑡 + 𝑏1 𝜇𝑒 𝑟𝑡 + 𝑐1 𝜈𝑒 𝑟𝑡
𝜈𝑟𝑒 𝑟𝑡 = 𝑎2 𝜆𝑒 𝑟𝑡 + 𝑏2 𝜇𝑟𝑒 𝑟𝑡 + 𝑐2 𝜈𝑒 𝑟𝑡

which is equivalent to have

(𝑎 − 𝑟)𝜆 + 𝑏𝜇 + 𝑐𝜈 = 0
𝑎1 𝜆 + (𝑏1 − 𝑟)𝜇 + 𝑐1 𝜈 = 0 } (20)
𝑎2 𝜆 + 𝑏2 𝜇 + (𝑐2 − 𝑟)𝜈 = 0

System (20) has a nonzero solution only if its determinant Δ is zero, that is,

𝑎−𝑟 𝑏 𝑐
∆ = | 𝑎1 𝑏1 − 𝑟 𝑐1 | = 0 (21)
𝑎2 𝑏2 𝑐2 − 𝑟

(21) is called the characteristic equation associated with system (17).

14
Case I. Assume that the roots r1, r2, r3 of the characteristic equation (21) are real and distinct. Replacing
in (20) r by r1 and solving system (20), we obtain numbers 1, 1, 1. The same applies for r = r2
obtaining 2, 2, 2 and also for r = r3 obtaining now 3, 3, 3. Finally, for the three collections of
numbers , , , we have three systems of particular solutions given by
(1) (1) (1)
𝑦1 = 𝜆1 𝑒 𝑟1 𝑡 𝑦2 = 𝜇1 𝑒 𝑟1 𝑡 𝑦3 = 𝜈1 𝑒 𝑟1 𝑡
(2) (2) (2)
𝑦1 = 𝜆2 𝑒 𝑟2 𝑡 𝑦2 = 𝜇2 𝑒 𝑟2 𝑡 𝑦3 = 𝜈2 𝑒 𝑟2 𝑡 } (22)
(3) (3) (3)
𝑦1 = 𝜆3 𝑒 𝑟3 𝑡 𝑦2 = 𝜇3 𝑒 𝑟3 𝑡 𝑦3 = 𝜈3 𝑒 𝑟3 𝑡

and the solution of (17) takes the form

(1) (2) (3)
𝑦1 = 𝐶1 𝑦1 + 𝐶2 𝑦1 + 𝐶3 𝑦1
(1) (2) (3)
𝑦2 = 𝐶1 𝑦2 + 𝐶2 𝑦2 + 𝐶3 𝑦2 } (23)
(1) (2) (3)
𝑦3 = 𝐶1 𝑦3 + 𝐶2 𝑦3 + 𝐶3 𝑦3

Example 2. Integrate the following system

𝑑𝑦1
= 𝑦1 − 𝑦2 + 𝑦3
𝑑𝑡
𝑑𝑦2
= 𝑦1 + 𝑦2 − 𝑦3
𝑑𝑡
𝑑𝑦3
= 2𝑦1 − 𝑦2
𝑑𝑡
Let us form the system of type (20)

(1 − 𝑟)𝜆 − 𝜇 + 𝜈 = 0
𝜆 + (1 − 𝑟)𝜇 − 𝜈 = 0 } (24)
2𝜆 − 𝜇 − 𝑟𝜈 = 0

and write the characteristic equation

1−𝑟 −1 1
1 − 𝑟 −1 1 −1 1 1−𝑟
∆=| 1 1 − 𝑟 −1| = (1 − 𝑟) | |+| |+| |
−1 −𝑟 2 −𝑟 2 −1
2 −1 −𝑟
= (1 − 𝑟)[−𝑟(1 − 𝑟) − 1] + (−𝑟 + 2) + [−1 − 2(1 − 𝑟)]
= (1 − 𝑟)(𝑟 2 − 𝑟 − 1) + (2 − 𝑟) + (2𝑟 − 3) = −𝑟 3 + 2𝑟 2 + 𝑟 − 2 = 0
⇒ 𝑟 3 − 2𝑟 2 − 𝑟 + 2 = 0.

Therefore

(𝑟 − 1)(𝑟 2 − 𝑟 − 2) = 0

(𝑟 − 1)(𝑟 − 2)(𝑟 + 1) = 0

𝑟1 = 1, 𝑟2 = 2, 𝑟3 = −1.

Thus, from (24) and knowing that r = r1 = 1, we get

15
 −𝜇1 + 𝜈1 = 0
𝜆1 − 𝜈1 = 0
2𝜆1 − 𝜇1 − 𝜈1 = 0

By solving the last system of equations, we have

𝜆1 1
(𝜇1 ) = 𝐴 (1) , 𝐴 = 𝜈1 , 𝐴 ∈ ℝ.
𝜈1 1

If A = 1, then

𝜆1 = 1, 𝜇1 = 1, 𝜈1 = 1
(1) (1) (1)
𝑦1 = 𝑒𝑡, 𝑦2 = 𝑒𝑡, 𝑦3 = 𝑒𝑡

Now, from (24) and knowing that r = r2 = 2, we get

−𝜆2 − 𝜇2 + 𝜈2 = 0
𝜆2 − 𝜇2 − 𝜈2 = 0
2𝜆2 − 𝜇2 − 2𝜈2 = 0

The solution of this system of equations is

𝜆2 1
(𝜇2 ) = 𝐴1 (0) , 𝐴1 = 𝜈2 , 𝐴1 ∈ ℝ.
𝜈2 1

If A1 = 1, then

𝜆2 = 1, 𝜇2 = 0, 𝜈2 = 1
(2) (2) (2)
𝑦1 = 𝑒 2𝑡 , 𝑦2 = 0, 𝑦3 = 𝑒 2𝑡

Lastly, from (24) and knowing that r = r3 = -1, we obtain

2𝜆3 − 𝜇3 + 𝜈3 = 0
𝜆3 + 2𝜇3 − 𝜈3 = 0
2𝜆3 − 𝜇3 + 𝜈3 = 0

The solution of this system of equations is

𝜆3 −1/5
(𝜇3 ) = 𝐴2 ( 3/5 ) , 𝐴2 = 𝜈3 , 𝐴2 ∈ ℝ.
𝜈3 1

If A2 = 1, then

𝜆3 = −1/5, 𝜇3 = 3/5, 𝜈3 = 1
(3) (3) (3)
𝑦1 = (−1/5)𝑒 −𝑡 , 𝑦2 = (3/5)𝑒 −𝑡 , 𝑦3 = 𝑒 −𝑡

16
Consequently, the general solution of the given system is
(1) (2) (3)
𝑦1 = 𝐶1 𝑦1 + 𝐶2 𝑦1 + 𝐶3 𝑦1 = 𝐶1 𝑒 𝑡 + 𝐶2 𝑒 2𝑡 − (1/5)𝐶3 𝑒 −𝑡
(1) (2) (3)
𝑦2 = 𝐶1 𝑦2 + 𝐶2 𝑦2 + 𝐶3 𝑦2 = 𝐶1 𝑒 𝑡 + (3/5)𝐶3 𝑒 −𝑡 }
(1) (2) (3)
𝑦3 = 𝐶1 𝑦3 + 𝐶2 𝑦3 + 𝐶3 𝑦3 = 𝐶1 𝑒 𝑡 + 𝐶2 𝑒 2𝑡 + 𝐶3 𝑒 −𝑡

Case II. Assume that r1 is a root of the characteristic equation and that r2 = r3 = r is a double root of it.
Replacing in (20) r by r1 and solving system (20), we obtain numbers λ, μ, ν. Thus, the following
particular solutions are obtained
(1) (1) (1)
𝑦1 = 𝐶1 𝜆𝑒 𝑟1 𝑡 𝑦2 = 𝐶1 𝜇𝑒 𝑟1 𝑡 𝑦3 = 𝐶1 𝜈𝑒 𝑟1 𝑡

On the other hand, since r2 = r3 = r is a double root of the characteristic equation, then we can assign to r
a system of particular solutions of the form
(2) (2) (2)
𝑦1 = (𝜆1 𝑡 + 𝜇1 )𝑒 𝑟𝑡 𝑦2 = (𝜆2 𝑡 + 𝜇2 )𝑒 𝑟𝑡 𝑦3 = (𝜆3 𝑡 + 𝜇3 )𝑒 𝑟𝑡 (25)

From (25), we obtain

(2)
𝑑𝑦1
= 𝑟(𝜆1 𝑡 + 𝜇1 )𝑒 𝑟𝑡 + 𝜆1 𝑒 𝑟𝑡 = [𝑟(𝜆1 𝑡 + 𝜇1 ) + 𝜆1 ]𝑒 𝑟𝑡
𝑑𝑡
(2)
𝑑𝑦2 (26)
= 𝑟(𝜆2 𝑡 + 𝜇2 )𝑒 𝑟𝑡 + 𝜆2 𝑒 𝑟𝑡 = [𝑟(𝜆2 𝑡 + 𝜇2 ) + 𝜆2 ]𝑒 𝑟𝑡
𝑑𝑡
(2)
𝑑𝑦3
= 𝑟(𝜆3 𝑡 + 𝜇3 )𝑒 𝑟𝑡 + 𝜆3 𝑒 𝑟𝑡 = [𝑟(𝜆3 𝑡 + 𝜇3 ) + 𝜆3 ]𝑒 𝑟𝑡 }
𝑑𝑡

By substituting (25) and (26) into (17), we have

𝑟(𝜆1 𝑡 + 𝜇1 ) + 𝜆1 = 𝑎(𝜆1 𝑡 + 𝜇1 ) + 𝑏(𝜆2 𝑡 + 𝜇2 ) + 𝑐(𝜆3 𝑡 + 𝜇3 )

𝑟(𝜆2 𝑡 + 𝜇2 ) + 𝜆2 = 𝑎1 (𝜆1 𝑡 + 𝜇1 ) + 𝑏1 (𝜆2 𝑡 + 𝜇2 ) + 𝑐1 (𝜆3 𝑡 + 𝜇3 )
𝑟(𝜆3 𝑡 + 𝜇3 ) + 𝜆3 = 𝑎2 (𝜆1 𝑡 + 𝜇1 ) + 𝑏2 (𝜆2 𝑡 + 𝜇2 ) + 𝑐2 (𝜆3 𝑡 + 𝜇3 )

which is equivalent to

𝑟𝜆1 𝑡 + (𝑟𝜇1 + 𝜆1 ) = (𝑎𝜆1 + 𝑏𝜆2 + 𝑐𝜆3 )𝑡 + (𝑎𝜇1 + 𝑏𝜇2 + 𝑐𝜇3 )

𝑟𝜆2 𝑡 + (𝑟𝜇2 + 𝜆2 ) = (𝑎1 𝜆1 + 𝑏1 𝜆2 + 𝑐1 𝜆3 )𝑡 + (𝑎1 𝜇1 + 𝑏1 𝜇2 + 𝑐1 𝜇3 ) } (27)
𝑟𝜆3 𝑡 + (𝑟𝜇3 + 𝜆3 ) = (𝑎2 𝜆1 + 𝑏2 𝜆2 + 𝑐2 𝜆3 )𝑡 + (𝑎2 𝜇1 + 𝑏2 𝜇2 + 𝑐3 𝜇3 )

From (27), we obtain two systems of equations; the first one is given by

𝑎𝜆1 + 𝑏𝜆2 + 𝑐𝜆3 = 𝑟𝜆1

𝑎1 𝜆1 + 𝑏1 𝜆2 + 𝑐1 𝜆3 = 𝑟𝜆2 }
𝑎2 𝜆1 + 𝑏2 𝜆2 + 𝑐2 𝜆3 = 𝑟𝜆3

and the second one is given by

17
 𝑎𝜇1 + 𝑏𝜇2 + 𝑐𝜇3 = 𝑟𝜇1 + 𝜆1
𝑎1 𝜇1 + 𝑏1 𝜇2 + 𝑐1 𝜇3 = 𝑟𝜇2 + 𝜆2 }
𝑎2 𝜇1 + 𝑏2 𝜇2 + 𝑐2 𝜇3 = 𝑟𝜇3 + 𝜆3

with which we can find the values of 𝜆1 , 𝜆2 , 𝜆3 , 𝜇1 , 𝜇2 , 𝜇3.

Lastly, the solution of the system is

(1) (2)
𝑦1 = 𝑦1 + 𝑦1 = 𝐶1 𝜆𝑒 𝑟1 𝑡 + (𝜆1 𝑡 + 𝜇1 )𝑒 𝑟𝑡
(1) (2)
𝑦2 = 𝑦2 + 𝑦2 = 𝐶1 𝜇𝑒 𝑟1 𝑡 + (𝜆2 𝑡 + 𝜇2 )𝑒 𝑟𝑡 }
(1) (2)
𝑦3 = 𝑦3 + 𝑦3 = 𝐶1 𝜈𝑒 𝑟1 𝑡 + (𝜆3 𝑡 + 𝜇3 )𝑒 𝑟𝑡

Example 3. Integrate the system

𝑑𝑦1
= 2𝑦1 + 2𝑦2 − 𝑦3
𝑑𝑡
𝑑𝑦2
= −2𝑦1 + 4𝑦2 + 𝑦3
𝑑𝑡
𝑑𝑦3
= −3𝑦1 + 8𝑦2 + 2𝑦3
𝑑𝑡
The system of type (20) has the form

(2 − 𝑟)𝜆 + 2𝜇 − 𝜈 = 0
−2𝜆 + (4 − 𝑟)𝜇 + 𝜈 = 0 } (28)
−3𝜆 + 8𝜇 + (2 − 𝑟)𝜈 = 0

and the corresponding characteristic equation can be written in the form

2−𝑟 2 −1
4−𝑟 1 −2 1 −2 4 − 𝑟
∆ = | −2 4−𝑟 1 | = (2 − 𝑟) | |− 2| |−| |
8 2−𝑟 −3 2 − 𝑟 −3 8
−3 8 2−𝑟
= (2 − 𝑟)[(4 − 𝑟)(2 − 𝑟) − 8] − 2[−2(2 − 𝑟) + 3] − [−16 + 3(4 − 𝑟)]
= (2 − 𝑟)(𝑟 2 − 6𝑟) − 2(2𝑟 − 1) − (−3𝑟 − 4) = −𝑟 3 + 8𝑟 2 − 13𝑟 + 6 = 0
⇒ 𝑟 3 − 8𝑟 2 + 13𝑟 − 6 = 0.

Therefore

(𝑟 − 1)(𝑟 2 − 7𝑟 + 6) = 0

(𝑟 − 1)(𝑟 − 6)(𝑟 − 1) = 0

𝑟1 = 6, 𝑟2 = 𝑟3 = 𝑟 = 1.

Thus, from (28) and knowing that r = r1 = 6, we get

−4𝜆 + 2𝜇 − 𝜈 = 0
−2𝜆 − 2𝜇 + 𝜈 = 0 }
−3𝜆 + 8𝜇 − 4𝜈 = 0

18
whose solution is

𝜆 0
(𝜇) = 𝐴 (1⁄2) , 𝐴 = 𝜈, 𝐴 ∈ ℝ.
𝜈 1
If A = 2, then

𝜆 = 0, 𝜇 = 1, 𝜈=2
(1) (1) (1)
𝑦1 = 0, 𝑦2 = 𝐶1 𝑒 6𝑡 , 𝑦3 = 2𝐶1 𝑒 6𝑡 . (29)

Since r2 = r3 = r = 1 is a double root of Δ = 0, then

(2) (2) (2)
𝑦1 = (𝜆1 𝑡 + 𝜇1 )𝑒 𝑡 𝑦2 = (𝜆2 𝑡 + 𝜇2 )𝑒 𝑡 𝑦3 = (𝜆3 𝑡 + 𝜇3 )𝑒 𝑡 . (30)

It follows that
(2)
𝑑𝑦1
= (𝜆1 𝑡 + 𝜇1 )𝑒 𝑡 + 𝜆1 𝑒 𝑡 = [𝜆1 𝑡 + (𝜆1 + 𝜇1 )]𝑒 𝑡
𝑑𝑡
(2)
𝑑𝑦2 (31)
= (𝜆2 𝑡 + 𝜇2 )𝑒 𝑡 + 𝜆2 𝑒 𝑡 = [𝜆2 𝑡 + (𝜆2 + 𝜇2 )]𝑒 𝑡
𝑑𝑡
(2)
𝑑𝑦3
= (𝜆3 𝑡 + 𝜇3 )𝑒 𝑡 + 𝜆3 𝑒 𝑡 = [𝜆3 𝑡 + (𝜆3 + 𝜇3 )]𝑒 𝑡 }
𝑑𝑡

By substituting (30) and (31) into the original system, we have

𝜆1 𝑡 + (𝜆1 + 𝜇1 ) = (2𝜆1 + 2𝜆2 − 𝜆3 )𝑡 + (2𝜇1 + 2𝜇2 − 𝜇3 )

𝜆2 𝑡 + (𝜆2 + 𝜇2 ) = (−2𝜆1 + 4𝜆2 + 𝜆3 )𝑡 + (−2𝜇1 + 4𝜇2 + 𝜇3 ) }
𝜆3 𝑡 + (𝜆3 + 𝜇3 ) = (−3𝜆1 + 8𝜆2 + 2𝜆3 )𝑡 + (−3𝜇1 + 8𝜇2 + 2𝜇3 )

We can now form two systems of equations. The first one is given by

2𝜆1 + 2𝜆2 − 𝜆3 = 𝜆1 𝜆1 + 2𝜆2 − 𝜆3 = 0

−2𝜆1 + 4𝜆2 + 𝜆3 = 𝜆2 } ⟹ −2𝜆1 + 3𝜆2 + 𝜆3 = 0}
−3𝜆1 + 8𝜆2 + 2𝜆3 = 𝜆3 −3𝜆1 + 8𝜆2 + 𝜆3 = 0

whose solution is

𝜆1 5⁄7
(𝜆2 ) = 𝐶 (1⁄7) , 𝐶 = 𝜆3 , 𝐶 ∈ ℝ.
𝜆3 1

If we make C = 7C2, then

𝜆1 = 5𝐶2 , 𝜆2 = 𝐶2 , 𝜆3 = 7𝐶2 . (32)

The second system is given by

19
 2𝜇1 + 2𝜇2 − 𝜇3 = 𝜆1 + 𝜇1 𝜇1 + 2𝜇2 − 𝜇3 = 5𝐶2
−2𝜇1 + 4𝜇2 + 𝜇3 = 𝜆2 + 𝜇2 } ⟹ −2𝜇1 + 3𝜇2 + 𝜇3 = 𝐶2 }
−3𝜇1 + 8𝜇2 + 2𝜇3 = 𝜆3 + 𝜇3 −3𝜇1 + 8𝜇2 + 𝜇3 = 7𝐶2

whose solution is
𝜇1 13⁄7 5⁄7
𝜇
( 2 ) = 𝐶2 (11⁄7) + 𝐶3 (1⁄7) , 𝐶3 = 𝜇3 , 𝐶3 ∈ ℝ.
𝜇3 0 1
This gives

1 1
𝜇1 = (13𝐶2 + 5𝐶3 ), 𝜇2 = (11𝐶2 + 𝐶3 ), 𝜇3 = 𝐶3 . (33)
7 7
By substituting (32) and (33) into (30), we have

(2) 1 (2) 1 (2)

𝑦1 = [5𝐶2 𝑡 + (13𝐶2 + 5𝐶3 )] 𝑒 𝑡 , 𝑦2 = [𝐶2 𝑡 + (11𝐶2 + 𝐶3 )] 𝑒 𝑡 , 𝑦3 = (7𝐶2 𝑡 + 𝐶3 )𝑒 𝑡 .
7 7

Therefore, the general solution is

(1) (2) 1
𝑦1 = 𝑦1 + 𝑦1 [5𝐶2 𝑡 + (13𝐶2 + 5𝐶3 )] 𝑒 𝑡
=
7
(1) (2) 1
𝑦2 = 𝑦2 + 𝑦2 = 𝐶1 𝑒 6𝑡 + [𝐶2 𝑡 + (11𝐶2 + 𝐶3 )] 𝑒 𝑡
7
(1) (2)
𝑦3 = 𝑦3 + 𝑦3 = 2𝐶1 𝑒 + (7𝐶2 𝑡 + 𝐶3 )𝑒 𝑡
6𝑡
}

Case III. Assume that r1 is a real root of the characteristic equation (21) and the remaining two roots r2, r3
are complex conjugate.

𝑟2 = 𝛼 + 𝑖𝛽, 𝑟3 = 𝛼 − 𝑖𝛽

Replacing in (20) r by r1 and solving system (20), we obtain numbers 1, 1, 1. Thus, we have a system
of particular solutions given by
(1) (1) (1)
𝑦1 = 𝐶1 𝜆1 𝑒 𝑟1 𝑡 𝑦2 = 𝐶1 𝜇1 𝑒 𝑟1 𝑡 𝑦3 = 𝐶1 𝜈1 𝑒 𝑟1 𝑡

Now, replacing in (20) r by r2 =  + i and solving (20), we get 2, 2, 2. Therefore, we have a system of
particular solutions given by
(2) (2) (2)
𝑦1 = 𝜆2 𝑒 (𝛼+𝑖𝛽)𝑡 𝑦2 = 𝜇2 𝑒 (𝛼+𝑖𝛽)𝑡 𝑦3 = 𝜈2 𝑒 (𝛼+𝑖𝛽)𝑡

which is equivalent to
(2) (2) (2)
𝑦1 = 𝜆2 𝑒 𝛼𝑡 (cos 𝛽𝑡 + 𝑖 sin 𝛽𝑡), 𝑦2 = 𝜇2 𝑒 𝛼𝑡 (cos 𝛽𝑡 + 𝑖 sin 𝛽𝑡), 𝑦3 = 𝜈2 𝑒 𝛼𝑡 (cos 𝛽𝑡 + 𝑖 sin 𝛽𝑡).

Lastly, replacing in (20) r by r3 =  - i and solving (20), we get 3, 3, 3. Therefore, we have a system
of particular solutions given by

20
 (3) (3) (3)
𝑦1 = 𝜆3 𝑒 (𝛼−𝑖𝛽)𝑡 𝑦2 = 𝜇3 𝑒 (𝛼−𝑖𝛽)𝑡 𝑦3 = 𝜈3 𝑒 (𝛼−𝑖𝛽)𝑡

which is equivalent to
(3) (3) (3)
𝑦1 = 𝜆3 𝑒 𝛼𝑡 (cos 𝛽𝑡 − 𝑖 sin 𝛽𝑡), 𝑦2 = 𝜇3 𝑒 𝛼𝑡 (cos 𝛽𝑡 − 𝑖 sin 𝛽𝑡) , 𝑦3 = 𝜈3 𝑒 𝛼𝑡 (cos 𝛽𝑡 − 𝑖 sin 𝛽𝑡).

If z = a + bi is a complex number, then the real and imaginary parts of z denoted by Re z and Im z,
respectively, are defined as follows:

𝑎 = Re 𝑧, 𝑏 = Im 𝑧

̅̅̅̅̅
(2) ̅̅̅̅̅
(2) ̅̅̅̅̅
(2) (2) (3) (2) (3) (2) (3)
Assume that 𝑦1 , 𝑦2 , 𝑦3 represent the real parts of { 𝑦1 , 𝑦1 } , { 𝑦2 , 𝑦2 } , { 𝑦3 , 𝑦3 } ,
̅̅̅̅̅
(3) ̅̅̅̅̅
(3) ̅̅̅̅̅
(3) (2) (3) (2)
respectively. Also, assume that 𝑦1 , 𝑦2 , 𝑦3 represent the imaginary parts of { 𝑦1 , 𝑦1 } , { 𝑦2 ,
(3) (2) (3)
𝑦2 } , { 𝑦3 , 𝑦3 } .

Then the general solution of the system is

̅̅̅̅̅
(2) ̅̅̅̅̅
(3)
𝑦1 = 𝐶1 𝜆1 𝑒 𝑟1 𝑡 + 𝐶2 𝑦1 + 𝐶3 𝑦1
̅̅̅̅̅
(2) ̅̅̅̅̅
(3)
(34)
𝑦2 = 𝐶1 𝜇1 𝑒 𝑟1 𝑡 + 𝐶2 𝑦2 + 𝐶3 𝑦2
̅̅̅̅̅
(2) ̅̅̅̅̅
(3)
𝑦3 = 𝐶1 𝜈1 𝑒 𝑟1 𝑡 + 𝐶2 𝑦3 + 𝐶3 𝑦3 }

Example 4. Integrate the following system:

𝑑𝑦1
= 8𝑦2
𝑑𝑡
𝑑𝑦2
= −2𝑦3
𝑑𝑡
𝑑𝑦3
= 2𝑦1 + 8𝑦2 − 2𝑦3
𝑑𝑡

Let us form the system of type (20)

−𝑟𝜆 + 8𝜇 = 0
−𝑟𝜇 − 2𝜈 = 0 } (35)
2𝜆 + 8𝜇 + (−2 − 𝑟)𝜈 = 0

and write the characteristic equation

−𝑟 8 0
−𝑟 −2 0 −2
∆=| 0 −𝑟 −2 | = −𝑟 | | − 8| | = −𝑟[−𝑟(−2 − 𝑟) + 16] − 8(4)
8 −2 − 𝑟 2 −2 − 𝑟
2 8 −2 − 𝑟
= −𝑟(𝑟 2 + 2𝑟 + 16) − 32 = −𝑟 3 − 2𝑟 2 − 16𝑟 − 32 = 0 ⇒ 𝑟 3 + 2𝑟 2 + 16𝑟 + 32
= 0.

Thus, we get

(𝑟 + 2)(𝑟 2 + 16) = 0

21
 𝑟1 = −2, 𝑟2 = 4𝑖, 𝑟3 = −4𝑖

If r = r1 = -2, then from (35), we get the system

2𝜆1 + 8𝜇1 = 0
2𝜇1 − 2𝜈1 = 0 }
2𝜆1 + 8𝜇1 = 0

whose solution is

𝜆1 −4
(𝜇1 ) = 𝐴 ( 1 ) , 𝐴 = 𝜈1 , 𝐴 ∈ ℝ.
𝜈1 1

If A = 1, then

𝜆1 = −4, 𝜇1 = 1, 𝜈1 = 1
(1) (1) (1)
𝑦1 = −4𝑒 −2𝑡 , 𝑦2 = 𝑒 −2𝑡 , 𝑦3 = 𝑒 −2𝑡

If r = r2 = 4i, then from (35), we get the system

−4𝑖𝜆2 + 8𝜇2 = 0
−4𝑖𝜇2 − 2𝜈2 = 0}
2𝜆2 + 8𝜇2 + (−2 − 4𝑖) 𝜈2 = 0

whose solution is

𝜆2 1
(𝜇2 ) = 𝐴1 (−1⁄2𝑖 ) , 𝐴1 = 𝜈2 , 𝐴1 ∈ ℂ.
𝜈2 1

If A1 = 2i, then

𝜆2 = 2𝑖, 𝜇2 = −1, 𝜈2 = 2𝑖
(2) (2) (2)
𝑦1 = 2𝑖𝑒 4𝑖𝑡 , 𝑦2 = −𝑒 4𝑖𝑡 , 𝑦3 = 2𝑖𝑒 4𝑖𝑡

(2) (2) (2)

𝑦1 = 2𝑖(cos 4𝑡 + 𝑖 sin 4𝑡), 𝑦2 = − cos 4𝑡 − 𝑖 sin 4𝑡 , 𝑦3 = 2𝑖(cos 4𝑡 + 𝑖 sin 4𝑡)

(2) (2) (2)

𝑦1 = −2 sin 4𝑡 + 𝑖(2 cos 4𝑡), 𝑦2 = −cos 4𝑡 − 𝑖 sin(4𝑡) , 𝑦3 = −2 sin 4𝑡 + 𝑖(2 cos 4𝑡). (36)

Lastly, If r = r3 = -4i, then from (35), we get the system

4𝑖𝜆3 + 8𝜇3 = 0
4𝑖𝜇3 − 2𝜈3 = 0}
2𝜆3 + 8𝜇3 + (−2 + 4𝑖) 𝜈3 = 0

whose solution is

22
 𝜆3 1
(𝜇3 ) = 𝐴2 (1⁄2𝑖 ) , 𝐴2 = 𝜈3 , 𝐴2 ∈ ℂ.
𝜈3 1

If A2 = 2i, then

𝜆3 = 2𝑖, 𝜇3 = 1, 𝜈3 = 2𝑖
(3) (3) (3)
𝑦1 = 2𝑖𝑒 −4𝑖𝑡 , 𝑦2 = 𝑒 −4𝑖𝑡 , 𝑦3 = 2𝑖𝑒 −4𝑖𝑡

(3) (3) (3)

𝑦1 = 2𝑖(cos 4𝑡 − 𝑖 sin 4𝑡), 𝑦2 = cos 4𝑡 − 𝑖 sin 4𝑡 , 𝑦3 = 2𝑖(cos 4𝑡 − 𝑖 sin 4𝑡)

(3) (3) (3)

𝑦1 = 2 sin 4𝑡 + 𝑖(2 cos 4𝑡), 𝑦2 = cos 4𝑡 − 𝑖 sin(4𝑡) , 𝑦3 = 2 sin 4𝑡 + 𝑖(2 cos 4𝑡). (37)

Note that the corresponding real parts of (36) and (37) have opposite signs. The solutions of the system
̅̅̅̅̅
(2) ̅̅̅̅̅
(2) ̅̅̅̅̅
(2)
will be equivalent once 𝑦1 , 𝑦2 , 𝑦3 are replaced in (34) by either the positive values or the negative
̅̅̅̅̅
(2) ̅̅̅̅̅
(2) ̅̅̅̅̅
(2) (2) (3) (2)
ones. For this reason, we can say that the real solutions 𝑦1 , 𝑦2 , 𝑦3 of { 𝑦1 , 𝑦1 } , { 𝑦2 ,
(3) (2) (3)
𝑦2 } , { 𝑦3 , 𝑦3 } are, respectively,

̅̅̅̅̅
(2) ̅̅̅̅̅
(2) ̅̅̅̅̅
(2)
𝑦1 = 2 sin 4𝑡 , 𝑦2 = cos 4𝑡, 𝑦3 = 2 sin 4𝑡 .

̅̅̅̅̅
(3) ̅̅̅̅̅
(3) ̅̅̅̅̅
(3) (2) (3) (2) (3) (2) (3)
The imaginary solutions 𝑦1 , 𝑦2 , 𝑦3 of { 𝑦1 , 𝑦1 } , { 𝑦2 , 𝑦2 } , { 𝑦3 , 𝑦3 } are given by

̅̅̅̅̅
(3) ̅̅̅̅̅
(3) ̅̅̅̅̅
(3)
𝑦1 = 2 cos 4𝑡 , 𝑦2 = − sin 4𝑡 , 𝑦3 = 2 cos 4𝑡 .

Therefore, the general solution of the given system is

(1) ̅̅̅̅̅
(2) ̅̅̅̅̅̅
(3)
𝑦1 = 𝐶1 𝑦1 + 𝐶2 𝑦1 + 𝐶3 𝑦1 = −4𝐶1 𝑒 −2𝑡 + 2𝐶2 sin 4𝑡 + 2𝐶3 cos 4𝑡
(1) ̅̅̅̅̅
(2) ̅̅̅̅̅
(3)
𝑦2 = 𝐶1 𝑦2 + 𝐶2 𝑦2 + 𝐶3 𝑦2 = 𝐶1 𝑒 −2𝑡 + 𝐶2 cos 4𝑡 − 𝐶3 sin 4𝑡
(1) ̅̅̅̅̅
(2) ̅̅̅̅̅
(3)
𝑦3 = 𝐶1 𝑦3 + 𝐶2 𝑦3 + 𝐶3 𝑦3 = 𝐶1 𝑒 −2𝑡 + 2𝐶2 sin 4𝑡 + 2𝐶3 cos 4𝑡 }

Method for Integration of a System of Three Linear Nonhomogeneous DE with Constant

Coefficients (Method of Variation of Parameters)

Assume that for the given system

𝑑𝑦1
+ 𝑎1 𝑦1 + 𝑏1 𝑦2 + 𝑒1 𝑦3 = 𝑓1 (𝑡) (38𝐴)
𝑑𝑡
𝑑𝑦2
+ 𝑎2 𝑦1 + 𝑏2 𝑦2 + 𝑒2 𝑦3 = 𝑓2 (𝑡) (38𝐵) (38)
𝑑𝑡
𝑑𝑦3
+ 𝑎3 𝑦1 + 𝑏3 𝑦2 + 𝑒3 𝑦3 = 𝑓3 (𝑡) (38𝐶) }
𝑑𝑡
a general solution of the corresponding homogeneous system has been found.

23
By using the method of variation of parameters, let us assume that C1, C2, C3 are functions of t.
(1) (2) (3)
𝑦1 = 𝐶1 (𝑡)𝑦1 + 𝐶2 (𝑡)𝑦1 + 𝐶3 (𝑡)𝑦1
(1) (2) (3)
𝑦2 = 𝐶1 (𝑡)𝑦2 + 𝐶2 (𝑡)𝑦2 + 𝐶3 (𝑡)𝑦2 } (39)
(1) (2) (3)
𝑦3 = 𝐶1 (𝑡)𝑦3 + 𝐶2 (𝑡)𝑦3 + 𝐶3 (𝑡)𝑦3

In order to compute the values of C1(t), C2(t), C3(t), find the derivative of y1 with respect to t

(1) ′ (1) (2) ′ (2) (3) ′ (3)

𝑦1′ = 𝐶1 𝑦1 + 𝐶1′ 𝑦1 + 𝐶2 𝑦1 + 𝐶2′ 𝑦1 + 𝐶3 𝑦1 + 𝐶3′ 𝑦1 (40)

By substituting (39) and (40) into (38A), we have

(1) (2) (3) (1) ′ (1) (1) (1)

𝐶1′ 𝑦1 + 𝐶2′ 𝑦1 + 𝐶3′ 𝑦1 + 𝐶1 [𝑦1 + 𝑎1 𝑦1 + 𝑏1 𝑦2 + 𝑒1 𝑦3 ]
(2) ′ (2) (2) (2) (3) ′ (3) (3) (3)
+ 𝐶2 [𝑦1 + 𝑎1 𝑦1 + 𝑏1 𝑦2 + 𝑒1 𝑦3 ] + 𝐶3 [𝑦1 + 𝑎1 𝑦1 + 𝑏1 𝑦2 + 𝑒1 𝑦3 ]
= 𝑓1 (𝑡).

All sums in brackets are zero (they represent the expressions of the corresponding homogeneous system).
Thus, we obtain
(1) (2) (3)
𝐶1′ 𝑦1 + 𝐶2′ 𝑦1 + 𝐶3′ 𝑦1 = 𝑓1 (𝑡)

In a similar way, from (38B) and (38C), we get the remaining two equations with which we can then form
the system
(1) (2) (3)
𝐶1′ 𝑦1 + 𝐶2′ 𝑦1 + 𝐶3′ 𝑦1 = 𝑓1 (𝑡)
(1) (2) (3)
𝐶1′ 𝑦2 + 𝐶2′ 𝑦2 + 𝐶3′ 𝑦2 = 𝑓2 (𝑡) } (41)
(1) (2) (3)
𝐶1′ 𝑦3 + 𝐶2′ 𝑦3 + 𝐶3′ 𝑦3 = 𝑓3 (𝑡)

System (41) with unknowns C1, C2, C3 has a solution since its determinant
(1) (2) (3)
𝑦1 𝑦1 𝑦1
Det(A) = |𝑦2(1) (2)
𝑦2
(3)
𝑦2 | ≠ 0.
(1) (2) (3)
𝑦3 𝑦3 𝑦3

This is due to the fact that the particular solutions of the corresponding homogeneous system are linearly
independent (see [1].)

Once C1, C2, C3 are found, by using integration, we can determine C1, C2, C3 which permits us to write
the general solution (39) of (38).

Example 5. Integrate the system

24
 𝑑𝑦1
= 2𝑦1 + 𝑦2 − 2𝑦3 + (2 − 𝑡)
𝑑𝑡
𝑑𝑦2
= −𝑦1 +1 (42)
𝑑𝑡
𝑑𝑦3
= 𝑦1 + 𝑦2 − 𝑦3 + (1 − 𝑡)
𝑑𝑡

Firstly, let us solve the homogeneous system

𝑑𝑦1
= 2𝑦1 + 𝑦2 − 2𝑦3
𝑑𝑡
𝑑𝑦2
= −𝑦1
𝑑𝑡
𝑑𝑦3
= 𝑦1 + 𝑦2 − 𝑦3
𝑑𝑡

Let us form the system of type (20)

(2 − 𝑟)𝜆 + 𝜇 − 2𝜈 = 0
− 𝜆 − 𝑟𝜇 =0} (43)
𝜆 + 𝜇 + (−1 − 𝑟)𝜈 = 0

The characteristic equation is

2−𝑟 1 −2
1 −2 2−𝑟 −2
∆ = | −1 −𝑟 0 |=| |−𝑟| |
1 −1 − 𝑟 1 −1 − 𝑟
1 1 −1 − 𝑟
= (−1 − 𝑟 + 2) − 𝑟[(2 − 𝑟)(−1 − 𝑟) + 2] = (1 − 𝑟) − 𝑟(−𝑟 + 𝑟 2 )
= −𝑟 3 + 𝑟 2 − 𝑟 + 1 = 0 ⇒ 𝑟 3 − 𝑟 2 + 𝑟 − 1 = 0.

So, we obtain

(𝑟 − 1)(𝑟 2 + 1) = 0

𝑟1 = 1, 𝑟2 = 𝑖, 𝑟3 = −𝑖

If r = r1 = 1, then from (43), we get the system

𝜆1 + 𝜇1 − 2𝜈1 = 0
−𝜆1 − 𝜇1 = 0}
𝜆1 + 𝜇1 − 2𝜈1 = 0

whose solution is

𝜆1 −1
(𝜇1 ) = 𝐴 ( 1 ) , 𝐴 = 𝜇1 , 𝐴 ∈ ℝ.
𝜈1 0

If A = 1, then

𝜆1 = −1, 𝜇1 = 1, 𝜈1 = 0

25
 (1) (1) (1)
𝑦1 = −𝑒 𝑡 , 𝑦2 = 𝑒𝑡, 𝑦3 =0

If r = r2 = i, then from (43), we get the system

(2 − 𝑖)𝜆2 + 𝜇2 − 2𝜈2 = 0
− 𝜆2 − 𝑖𝜇2 = 0}
𝜆2 + 𝜇2 + (−1 − 𝑖) 𝜈2 = 0

whose solution is

𝜆2 1
(𝜇2 ) = 𝐴1 (−1⁄𝑖 ) , 𝐴1 = 𝜈2 , 𝐴1 ∈ ℂ.
𝜈2 1

If A1 = i, then

𝜆2 = 𝑖, 𝜇2 = −1, 𝜈2 = 𝑖
(2) (2) (2)
𝑦1 = 𝑖𝑒 𝑖𝑡 , 𝑦2 = −𝑒 𝑖𝑡 , 𝑦3 = 𝑖𝑒 𝑖𝑡

(2) (2) (2)

𝑦1 = 𝑖(cos 𝑡 + 𝑖 sin 𝑡), 𝑦2 = − cos 𝑡 − 𝑖 sin 𝑡 , 𝑦3 = 𝑖(cos 𝑡 + 𝑖 sin 𝑡)

(2) (2) (2)

𝑦1 = − sin 𝑡 + 𝑖 cos 𝑡 , 𝑦2 = −cos 𝑡 − 𝑖 sin 𝑡 , 𝑦3 = − sin 𝑡 + 𝑖 cos 𝑡.

Lastly, If r = r3 = -i, then from (43), we get the system

(2 + 𝑖)𝜆3 + 𝜇3 − 2𝜈3 = 0
−𝜆3 + 𝑖𝜇3 =0}
𝜆3 + 𝜇3 + (−1 + 𝑖) 𝜈3 = 0

whose solution is

𝜆3 1
(𝜇3 ) = 𝐴2 (1⁄𝑖 ) , 𝐴2 = 𝜈3 , 𝐴2 ∈ ℂ.
𝜈3 1

If A2 = i, then

𝜆3 = 𝑖, 𝜇3 = 1, 𝜈3 = 𝑖
(3) (3) (3)
𝑦1 = 𝑖𝑒 −𝑖𝑡 , 𝑦2 = 𝑒 −𝑖𝑡 , 𝑦3 = 𝑖𝑒 −𝑖𝑡

(3) (3) (3)

𝑦1 = 𝑖(cos 𝑡 − 𝑖 sin 𝑡), 𝑦2 = cos 𝑡 − 𝑖 sin 𝑡 , 𝑦3 = 𝑖(cos 𝑡 − 𝑖 sin 𝑡)

(3) (3) (3)

𝑦1 = sin 𝑡 + 𝑖 cos 𝑡 , 𝑦2 = cos 𝑡 − 𝑖 sin 𝑡 , 𝑦3 = sin 𝑡 + 𝑖 cos 𝑡.

Therefore,

̅̅̅̅̅
(2) ̅̅̅̅̅
(2) ̅̅̅̅̅
(2)
𝑦1 = sin 𝑡 , 𝑦2 = cos 𝑡, 𝑦3 = sin 𝑡 ,

26
 ̅̅̅̅̅
(3) ̅̅̅̅̅
(3) ̅̅̅̅̅
(3)
𝑦1 = cos 𝑡 , 𝑦2 = − sin 𝑡 , 𝑦3 = cos 𝑡 .

Therefore, from the general solution of the corresponding homogeneous system, we get

(1) ̅̅̅̅̅
(2) ̅̅̅̅̅̅
(3)
𝑦1 = 𝐶1 𝑦1 + 𝐶2 𝑦1 + 𝐶3 𝑦1 = −𝐶1 (𝑡)𝑒 𝑡 + 𝐶2 (𝑡)sin 𝑡 + 𝐶3 (𝑡) cos 𝑡
(1) ̅̅̅̅̅
(2) ̅̅̅̅̅
(3)
(44)
𝑦2 = 𝐶1 𝑦2 + 𝐶2 𝑦2 + 𝐶3 𝑦2 = 𝐶1 (𝑡)𝑒 𝑡 + 𝐶2 (𝑡)cos 𝑡 − 𝐶3 (𝑡) sin 𝑡
(1) ̅̅̅̅̅
(2) ̅̅̅̅̅
(3)
𝑦3 = 𝐶1 𝑦3 + 𝐶2 𝑦3 + 𝐶3 𝑦3 = 𝐶2 (𝑡)sin 𝑡 + 𝐶3 (𝑡) cos 𝑡}

Now, let us form the system

−𝑒 𝑡 𝐶1′ + sin 𝑡 𝐶2′ + cos 𝑡 𝐶3′ = 2 − 𝑡 (45𝐴)

𝑒 𝑡 𝐶1′ + cos 𝑡 𝐶2′ − sin 𝑡 𝐶3′ = 1 (45𝐵)} (45)
sin 𝑡 𝐶2′ + cos 𝑡 𝐶3′ = 1 − 𝑡 (45𝐶)

By subtracting (45C) from (45A), we have

𝐶1′ = −𝑒 −𝑡 (46)

𝐶1 (𝑡) = 𝑒 −𝑡 + 𝐶1 . (46𝐴)

According to (45C)

1 − 𝑡 − sin 𝑡 𝐶2′
𝐶3′ = (45𝐷)
cos 𝑡
Substituting (46) and (45D) into (45B), we get

sin 𝑡 (1 − 𝑡 − sin 𝑡 𝐶2′ )

−1 + cos 𝑡 𝐶2′ − =1
cos 𝑡

𝐶2′ = 2 cos 𝑡 + sin 𝑡 − 𝑡 sin 𝑡 (47)

𝐶2 (𝑡) = 2 sin 𝑡 − cos 𝑡 − ∫ 𝑡 sin 𝑡 𝑑𝑡

𝐶2 (𝑡) = 2 sin 𝑡 − cos 𝑡 + 𝑡 cos 𝑡 − sin 𝑡 + 𝐶2

𝐶2 (𝑡) = sin 𝑡 − cos 𝑡 + 𝑡 cos 𝑡 + 𝐶2 (47𝐴)

By substituting (47) into (45D), we have

1 − 𝑡 − sin 𝑡 (2 cos 𝑡 + sin 𝑡 − 𝑡 sin 𝑡)

𝐶3′ =
cos 𝑡

𝐶3′ = cos 𝑡 − 2 sin 𝑡 − 𝑡 cos 𝑡

𝐶3 (𝑡) = sin 𝑡 + 2 cos 𝑡 − ∫ 𝑡 cos 𝑡 𝑑𝑡

27
 𝐶3 (𝑡) = sin 𝑡 + 2 cos 𝑡 − 𝑡 sin 𝑡 − cos 𝑡 + 𝐶3

𝐶3 (𝑡) = sin 𝑡 + cos 𝑡 − 𝑡 sin 𝑡 + 𝐶3 (48)

Substituting (46A), (47A), (48) into (44), we obtain

𝑦1 = (−𝑒 −𝑡 − 𝐶1 )𝑒 𝑡 + (sin 𝑡 − cos 𝑡 + 𝑡 cos 𝑡 + 𝐶2 ) sin 𝑡 + (sin 𝑡 + cos 𝑡 − 𝑡 sin 𝑡 + 𝐶3 ) cos 𝑡

𝑦2 = (𝑒 −𝑡 + 𝐶1 )𝑒 𝑡 + (sin 𝑡 − cos 𝑡 + 𝑡 cos 𝑡 + 𝐶2 ) cos 𝑡 − (sin 𝑡 + cos 𝑡 − 𝑡 sin 𝑡 + 𝐶3 ) sin 𝑡 }
𝑦3 = (sin 𝑡 − cos 𝑡 + 𝑡 cos 𝑡 + 𝐶2 ) sin 𝑡 + (sin 𝑡 + cos 𝑡 − 𝑡 sin 𝑡 + 𝐶3 ) cos 𝑡

Thus, from the last system, we get the general solution

𝑦1 = −𝐶1 𝑒 𝑡 + 𝐶2 sin 𝑡 + 𝐶3 cos 𝑡

𝑦2 = 𝐶1 𝑒 𝑡 + 𝐶2 cos 𝑡 − 𝐶3 sin 𝑡 + 𝑡 }
𝑦3 = 𝐶2 sin 𝑡 + 𝐶3 cos 𝑡 + 1

Solving a System of Two Linear Nonhomogeneous DE with Given Initial Conditions by Using the
Laplace Transform

Assume we need to find the solution of the system

𝑑𝑦1
= 𝑎1 𝑦1 + 𝑏1 𝑦2 + 𝑓1 (𝑡)
𝑑𝑡 } (49)
𝑑𝑦2
= 𝑎2 𝑦1 + 𝑏2 𝑦2 + 𝑓2 (𝑡)
𝑑𝑡
with initial conditions
(0) (0)
𝑦1 (0) = 𝑦1 , 𝑦2 (0) = 𝑦2 . (50)

Assume now that 𝑌1 (𝑠), 𝑌2 (𝑠) are the corresponding ℒ-transforms of 𝑦1 (𝑡), 𝑦2 (𝑡). Under these
conditions, we have

ℒ{𝑦1 (𝑡)} = 𝑌1 (𝑠), ℒ{𝑦2 (𝑡)} = 𝑌2 (𝑠)

ℒ{𝑦1′ (𝑡)} = 𝑠ℒ{𝑦1 (𝑡)} − 𝑦1 (0), ℒ{𝑦2′ (𝑡)} = 𝑠ℒ{𝑦2 (𝑡)} − 𝑦2 (0)

ℒ{𝑓1 (𝑡)} = 𝐹1 (𝑠), ℒ{𝑓2 (𝑡)} = 𝐹2 (𝑠)

Thus, by applying the Laplace transform to both sides of each equation of (49), we have the operational
system

ℒ{𝑦1′ (𝑡)} = 𝑎1 ℒ{𝑦1 (𝑡)} + 𝑏1 ℒ{𝑦2 (𝑡)} + ℒ{𝑓1 (𝑡)}

}
ℒ{𝑦2′ (𝑡)} = 𝑎2 ℒ{𝑦1 (𝑡)} + 𝑏2 ℒ{𝑦2 (𝑡)} + ℒ{𝑓2 (𝑡)}

which is equivalent to

𝑠𝑌1 (𝑠) = 𝑎1 𝑌1 (𝑠) + 𝑏1 𝑌2 (𝑠) + 𝐹1 (𝑠) + 𝑦1 (0)

}. (51)
𝑠𝑌2 (𝑠) = 𝑎2 𝑌1 (𝑠) + 𝑏2 𝑌2 (𝑠) + 𝐹2 (𝑠) + 𝑦2 (0)

28
(51) is an algebraic system of two linear equations with unknowns 𝑌1 (𝑠), 𝑌2 (𝑠). Once (51) is solved, the
inverse Laplace transform ℒ −1 is applied to the values of 𝑌1 (𝑠), 𝑌2 (𝑠) in order to obtain the solution
𝑦1 (𝑡), 𝑦2 (𝑡) of system (49) with initial conditions (50).

Example 6. Solve the system of DE with the given initial conditions

𝑑𝑦1
+ 𝑦2 = 0
𝑑𝑡 }, 𝑦1 (0) = 2, 𝑦2 (0) = 0.
𝑑𝑦2
+ 𝑦1 = 0
𝑑𝑡
Firstly, let us write the system in the form

𝑑𝑦1
= −𝑦2
𝑑𝑡 }
𝑑𝑦2
= −𝑦1
𝑑𝑡
It follows then that

ℒ{𝑦1′ (𝑡)} = −ℒ{𝑦2 (𝑡)}

′ (𝑡)} }
ℒ{𝑦2 = −ℒ{𝑦1 (𝑡)}

which can be written in the form

𝑠𝑌1 (𝑠) − 2 = −𝑌2 (𝑠)

}
𝑠𝑌2 (𝑠) − 0 = −𝑌1 (𝑠)

or simply

𝑠𝑌1 (𝑠) + 𝑌2 (𝑠) = 2

}.
𝑌1 (𝑠) + 𝑠𝑌2 (𝑠) = 0

Applying Cramer’s rule, this system has a solution given by

2𝑠 2
𝑌1 (𝑠) = , 𝑌2 (𝑠) = − .
𝑠2 −1 𝑠2 −1
Therefore, according to Table 1,

2𝑠 𝑠
𝑦1 (𝑡) = ℒ −1 {ℒ{𝑦1 (𝑡)}} = ℒ −1 {𝑌1 (𝑠)} = ℒ −1 { } = 2ℒ −1 { 2 } = 2 cosh 𝑡,
𝑠2−1 𝑠 −1

2 1
𝑦2 (𝑡) = ℒ −1 {ℒ{𝑦2 (𝑡)}} = ℒ −1 {𝑌2 (𝑠)} = ℒ −1 {− } = −2ℒ −1 { 2 } = −2 sinh 𝑡
𝑠2 −1 𝑠 −1

Example 7. Solve the system of DE with the given initial conditions

29
 𝑑𝑦1
+ 2𝑦2 = 3𝑡
𝑑𝑡 }, 𝑦1 (0) = 2, 𝑦2 (0) = 3.
𝑑𝑦2
− 2𝑦1 = 4
𝑑𝑡
Let us write the system in the form

𝑑𝑦1
= −2𝑦2 + 3𝑡
𝑑𝑡 }
𝑑𝑦2
= 2𝑦1 +4
𝑑𝑡
It follows that

ℒ{𝑦1′ (𝑡)} = −2ℒ{𝑦2 (𝑡)} + 3ℒ{𝑡}

′ (𝑡)} }
ℒ{𝑦2 = 2ℒ{𝑦1 (𝑡)} + 4ℒ{1}

which is equivalent to

3
𝑠𝑌1 (𝑠) − 2 = −2𝑌2 (𝑠) +
𝑠2}
4
𝑠𝑌2 (𝑠) − 3 = 2𝑌1 (𝑠) +
𝑠
or simply

2𝑠 2 + 3
𝑠𝑌1 (𝑠) + 2𝑌2 (𝑠) =
𝑠2 }
3𝑠 + 4
−2𝑌1 (𝑠) + 𝑠𝑌2 (𝑠) =
𝑠
Applying Cramer’s rule, this system has a solution given by

2𝑠 2 − 6𝑠 − 5 3𝑠 3 + 8𝑠 2 + 6
𝑌1 (𝑠) = , 𝑌2 (𝑠) = . (52)
𝑠(𝑠 2 + 4) 𝑠 2 (𝑠 2 + 4)

Now, we need to transform the last two solutions as follows:

2𝑠 2 − 6𝑠 − 5 2𝑠 6 5
2
= 2 − 2 − 2
.
𝑠(𝑠 + 4) 𝑠 + 4 𝑠 + 4 𝑠(𝑠 + 4)

We need to use the partial fraction expansion of a rational expression

5 𝐴 𝐵𝑠 + 𝐶
= + 2
𝑠(𝑠 2 + 4) 𝑠 (𝑠 + 4)

whose values for A, B, C are

5 5
𝐴= , 𝐵=− , 𝐶=0
4 4

30
which gives

5 5 5𝑠
= − .
𝑠(𝑠 2 + 4) 2
4𝑠 4(𝑠 + 4)

On the other hand,

3𝑠 3 + 8𝑠 2 + 6 3𝑠 8 6
2 2
= 2 + 2 + 2 2
𝑠 (𝑠 + 4) 𝑠 + 4 𝑠 + 4 𝑠 (𝑠 + 4)

We use again the partial fraction expansion of a rational expression to obtain

6 3 1 1
= ( 2− 2 )
𝑠 2 (𝑠 2 + 4) 2 𝑠 𝑠 +4

Now, we can rewrite (52) as follows:

2𝑠 6 5 5𝑠 13𝑠 6 5
𝑌1 (𝑠) = − 2 − + = − 2 − ,
𝑠2 2 2
+ 4 𝑠 + 4 4𝑠 4(𝑠 + 4) 4(𝑠 + 4) 𝑠 + 4 4𝑠

3𝑠 8 3 3 3𝑠 13 3
𝑌2 (𝑠) = + + − = + + .
𝑠 2 + 4 𝑠 2 + 4 2𝑠 2 2(𝑠 2 + 4) 𝑠 2 + 4 2(𝑠 2 + 4) 2𝑠 2

Therefore, according to table 1,

13 −1 𝑠 2 5 −1 1
𝑦1 (𝑡) = ℒ −1 {ℒ{𝑦1 (𝑡)}} = ℒ −1 {𝑌1 (𝑠)} = ℒ { 2 } − 3ℒ −1
{ } − ℒ { }
4 𝑠 + 22 𝑠 2 + 22 4 𝑠
13 5
= cos 2𝑡 − 3 sin 2𝑡 − ,
4 4
𝑠 13 2 3 1
𝑦2 (𝑡) = ℒ −1 {ℒ{𝑦2 (𝑡)}} = ℒ −1 {𝑌2 (𝑠)} = 3ℒ −1 { 2 2 } + ℒ −1 { 2 2 } + ℒ −1 { 2 }
𝑠 +2 4 𝑠 +2 2 𝑠
13 3
= 3 cos 2𝑡 + sin 2𝑡 + 𝑡 .
4 2

Exercises.

1. Solve the initial value problems

a)

𝑦 ′′ − 𝑦 ′ = 2 sin 𝑡, 𝑦(0) = 2, 𝑦 ′ (0) = 0

Solution.

𝑦 = cos 𝑡 − sin 𝑡 + 𝑒 𝑡

b)
𝑦 ′′′ + 𝑦 ′′ − 2𝑦 = 5𝑒 𝑡 , 𝑦(0) = 0, 𝑦 ′ (0) = 1, 𝑦 ′′ (0) = 2

Solution.

31
 𝑦 = 𝑡𝑒 𝑡

2. Integrate the following system by reducing it to a second order DE:

𝑑𝑦1
= 3𝑦1 + 2𝑦2 + 𝑒 𝑡
𝑑𝑡
𝑑𝑦2
= 𝑦1 + 2𝑦2 + 𝑒 4𝑡
𝑑𝑡

Solution.

2 1
𝑦1 = 𝐶1 𝑒 𝑡 + 𝐶2 𝑒 4𝑡 + 𝑡𝑒 4𝑡 + 𝑡𝑒 𝑡
3 3
1 1 1 1 1
𝑦2 = −𝐶1 𝑒 𝑡 + 𝐶2 𝑒 4𝑡 + 𝑡𝑒 4𝑡 + 𝑒 4𝑡 − 𝑡𝑒 𝑡 − 𝑒 𝑡
2 3 3 3 3
3. Integrate the following homogeneous systems:

a)

𝑑𝑦1
= 𝑦2 + 𝑦3
𝑑𝑡
𝑑𝑦2
= 3𝑦1 + 𝑦3
𝑑𝑡
𝑑𝑦3
= 3𝑦1 + 𝑦2
𝑑𝑡

Solution.

2
𝑦1 = −𝐶2 𝑒 −2𝑡 + 𝐶3 𝑒 3𝑡
3
𝑦2 = −𝐶1 𝑒 −𝑡 + 𝐶2 𝑒 −2𝑡 + 𝐶3 𝑒 3𝑡
𝑦3 = 𝐶1 𝑒 −𝑡 + 𝐶2 𝑒 −2𝑡 + 𝐶3 𝑒 3𝑡 }

b)

𝑑𝑦1
= 𝑦2 + 𝑦3
𝑑𝑡
𝑑𝑦2
= 𝑦1 + 𝑦3
𝑑𝑡
𝑑𝑦3
= 𝑦1 + 𝑦2
𝑑𝑡

Solution.

𝑦1 = 𝐶1 𝑒 2𝑡 − (𝐶2 + 𝐶3 )𝑒 −𝑡
𝑦2 = 𝐶1 𝑒 2𝑡 + 𝐶2 𝑒 −𝑡 }
2𝑡 −𝑡
𝑦3 = 𝐶1 𝑒 + 𝐶3 𝑒

32
c)

𝑑𝑦1
= 2𝑦1 + 𝑦2
𝑑𝑡
𝑑𝑦2
= 𝑦1 + 3𝑦2 − 𝑦3
𝑑𝑡
𝑑𝑦3
= −𝑦1 + 2𝑦2 + 3𝑦3
𝑑𝑡

Solution.

𝑦1 = 𝐶1 𝑒 2𝑡 + 𝐶2 𝑒 3𝑡 cos 𝑡 + 𝐶3 𝑒 3𝑡 sin 𝑡
𝑦2 = 𝐶2 𝑒 3𝑡 (cos 𝑡 − sin 𝑡) + 𝐶3 𝑒 3𝑡 (sin 𝑡 + cos 𝑡) }
𝑦3 = 𝐶1 𝑒 2𝑡 + 𝐶2 𝑒 3𝑡 (2cos 𝑡 + sin 𝑡) + 𝐶3 𝑒 3𝑡 (2 sin 𝑡 − cos 𝑡)

4. Integrate the following nonhomogeneous system:

𝑑𝑦1
= −𝑦1 + 𝑦2 + 𝑦3 + 𝑒 𝑡
𝑑𝑡
𝑑𝑦2
= 𝑦1 − 𝑦2 + 𝑦3 + 𝑒 3𝑡
𝑑𝑡
𝑑𝑦3
= 𝑦1 + 𝑦2 + 𝑦3 + 4
𝑑𝑡

Solution.

1 1 1 1 3
𝑦1 = 𝐶1 𝑒 −𝑡 + 𝐶2 𝑒 2𝑡 + 𝐶3 𝑒 −2𝑡 + 𝑒 𝑡 + 𝑒 3𝑡 − 2
3 6 2 6 20
1 1 1 1 7
𝑦2 = 𝐶1 𝑒 −𝑡 + 𝐶2 𝑒 2𝑡 − 𝐶3 𝑒 −2𝑡 − 𝑒 𝑡 + 𝑒 3𝑡 − 2
3 6 2 6 20
1 1 1 1
𝑦3 = − 𝐶1 𝑒 −𝑡 + 𝐶2 𝑒 2𝑡 − 𝑒 𝑡 + 𝑒 3𝑡 }
3 3 2 4
5. Integrate the following systems with the given initial conditions:

a)
𝑑𝑦1
+ 𝑦1 − 2𝑦2 = 0
𝑑𝑡 }, 𝑦1 (0) = 𝑦2 (0) = 1.
𝑑𝑦2
+ 𝑦1 + 4𝑦2 = 0
𝑑𝑡

Solution.

𝑦1 (𝑡) = 4𝑒 −2𝑡 − 3𝑒 −3𝑡

}
𝑦2 (𝑡) = 3𝑒 −3𝑡 − 2𝑒 −2𝑡

b)

33
 𝑑𝑦1
+ 𝑦1 = 𝑦2 + 𝑒 𝑡
𝑑𝑡 }, 𝑦1 (0) = 𝑦2 (0) = 1.
𝑑𝑦2 𝑡
+ 𝑦2 = 𝑦1 + 𝑒
𝑑𝑡

Solution.

𝑦1 (𝑡) = 𝑒 𝑡
}
𝑦2 (𝑡) = 𝑒 𝑡

References

1. C.E. Frasser. The Method of Variation of Parameters and the Higher Order Linear Nonhomogeneous
Differential Equation with Constant Coefficients. https://www.researchgate.net/publication/329861238

2. N. Piskunov. Differential and Integral Calculus, Vol II, CBS Publishers & Distributors, New Delhi,
India, 1999.

34

View publication stats