Al-Nahrain University 2nd Year, 1st Semester

College of Information Engineering Engineering Mathematics I

Dept. of Info& Comms Engineering Lecture3: Solving ODE (2nd Order)

Lecture 3: Solving ODEs (2nd order)

3.1 Introduction
3.2 Solutions of 2nd Order ODE
3.2.1 Finding the Complementary Solution
3.2.2 Finding the Particular Solution
3.3 Series Solutions of ODE
3.3.1 Power Series Solution
3.3.2 Solutions for Bessel’s Equation
3.3.3 Extended Power Series

0
 Al-Nahrain University 2nd Year, 1st Semester
College of Information Engineering Engineering Mathematics I
Dept. of Info& Comms Engineering Lecture3: Solving ODE (2nd Order)

3.1 Introduction
The general form of a second-order linear ordinary differential equation (ODE) is
d2 𝑦 d𝑦
𝑝(𝑥) + 𝑞(𝑥) + 𝑟(𝑥)𝑦 = 𝑓(𝑥)
d𝑥 2 d𝑥
where 𝑝(𝑥), 𝑞(𝑥), 𝑟(𝑥) and 𝑓(𝑥) are functions of 𝑥 only.
An important relative of this equation is
d2 𝑦 d𝑦
𝑝(𝑥) + 𝑞(𝑥) + 𝑟(𝑥)𝑦 = 0
d𝑥 2 d𝑥
which is obtained from the first equation by ignoring the term which is independent of 𝑦. The second
equation is said to be a homogeneous equation - all its terms contain 𝑦 or its derivatives. The first
equation is said to be inhomogeneous.
For example,
2
d2 𝑦 d𝑦
𝑥 2
+ 𝑥 + (𝑥 2 − 1)𝑦 = e−𝑥
d𝑥 d𝑥
is an inhomogeneous 2nd order linear equation in which 𝑝(𝑥) = 𝑥 2 , 𝑞(𝑥) = 𝑥, 𝑟(𝑥) = 𝑥 2 − 1 and
𝑓(𝑥) = e−𝑥 . The associated homogeneous equation is
d2 𝑦 d𝑦
𝑥 + 𝑥 + (𝑥 2 − 1)𝑦 = 0
d𝑥 2 d𝑥
The following properties of linear equations are necessary for finding solutions of 2nd order equations.
Property 1
If 𝑦1 (𝑥) and 𝑦2 (𝑥) are any two linearly independent solutions of a second-order homogeneous
equation then the general solution, 𝑦H (𝑥), is
𝑦H (𝑥) = 𝐴𝑦1 (𝑥) + 𝐵𝑦2 (𝑥)
where 𝐴, 𝐵 are constants.
We see that the second-order linear ordinary differential equation has two arbitrary constants in its
general solution. The functions 𝑦1 (𝑥) and 𝑦2 (𝑥) are linearly independent if one is not simply a
multiple of the other.
Property 2
Let 𝑦P (𝑥) be any solution of an inhomogeneous equation. Let 𝑦H (𝑥) be the general solution of the
associated homogeneous equation. The general solution of the inhomogeneous equation is then
given by
𝑦(𝑥) = 𝑦H (𝑥) + 𝑦P (𝑥)

1
 Al-Nahrain University 2nd Year, 1st Semester
College of Information Engineering Engineering Mathematics I
Dept. of Info& Comms Engineering Lecture3: Solving ODE (2nd Order)

In other words, to find the general solution of an inhomogeneous equation, we must first find the
general solution of the corresponding homogeneous problem, and then add to it any solution of the
inhomogeneous equation.
The function 𝑦H (𝑥) is known as the complementary solution and 𝑦P (𝑥) is called the particular
solution. Clearly the complementary solution of a homogeneous problem is the same as its general
solution; we shall often write 𝑦(𝑥) for both. If conditions are given, they are applied to the general
solution of the inhomogeneous equation to determine any unknown constants. This yields the
particular solution satisfying the given conditions.
Example:
d2 𝑦
Verify that 𝑦1 (𝑥) = 𝑥 and 𝑦2 (𝑥) = 1 both satisfy = 0. Write down the general solution of this
d𝑥 2
equation and verify that this indeed satisfies the equation.
Solution:
d𝑦1 d2 𝑦1 d2 𝑦 d𝑦2
If 𝑦1 (𝑥) = 𝑥 then = 1 and = 0, so that 𝑦1 satisfies = 0. If 𝑦2 (𝑥) = 1, then = 0 and
d𝑥 d𝑥 2 d𝑥 2 d𝑥
d2 𝑦2 d2 𝑦 d2 𝑦
= 0, so that 𝑦2 satisfies = 0. From Property 1, the general solution of = 0 is
d𝑥 2 d𝑥 2 d𝑥 2
𝑦H (𝑥) = 𝐴𝑥 + 𝐵(1) = 𝐴𝑥 + 𝐵
To verify that this satisfies the equation proceed as follows:
d𝑦H d2 𝑦H
=𝐴 ; =0
d𝑥 d𝑥 2
d2 𝑦
and so 𝑦H (𝑥) satisfies 2 = 0.
d𝑥

Example: Given
d2 𝑦
+𝑦 =𝑥
d𝑥 2
(a) Show that 𝑦H = 𝐴cos 𝑥 + 𝐵sin 𝑥 is a solution of the corresponding homogeneous equation.
(b) Verify that 𝑦P = 𝑥 is a particular solution.
(c) Verify that 𝑦H + 𝑦P does indeed satisfy the inhomogeneous equation.
Solution:
(a) If 𝑦H = 𝐴cos 𝑥 + 𝐵sin 𝑥, then
𝑦H′ = −𝐴sin 𝑥 + 𝐵cos 𝑥 𝑦H′′ = −𝐴cos 𝑥 − 𝐵sin 𝑥
We see immediately that 𝑦H + 𝑦H′′ = 0 so that 𝑦H is a solution of the homogeneous equation.
(b) If 𝑦P = 𝑥 then 𝑦P′ = 1 and 𝑦P′′ = 0. Substitution into the inhomogeneous equation shows that 𝑦P
satisfies the equation given in the question, that is 𝑦P = 𝑥 is a particular solution.
(c) Writing

2
 Al-Nahrain University 2nd Year, 1st Semester
College of Information Engineering Engineering Mathematics I
Dept. of Info& Comms Engineering Lecture3: Solving ODE (2nd Order)

𝑦 = 𝐴cos 𝑥 + 𝐵sin 𝑥 + 𝑥
we have
𝑦 ′ = −𝐴 sin 𝑥 + 𝐵 cos 𝑥 + 1; 𝑦 ′′ = −𝐴cos 𝑥 − 𝐵sin 𝑥
Substitution into the l.h.s. of the equation given in the question gives
(−𝐴cos 𝑥 − 𝐵sin 𝑥) + (𝐴cos 𝑥 + 𝐵sin 𝑥 + 𝑥)
which equals 𝑥, and so the complementary solution plus the particular solution is indeed a solution of
the inhomogeneous equation, as required by Property 2.

3.2 Solutions of 2nd Order ODE

We now proceed to study in detail those second-order linear equations which have constant
coefficients. The general form of such an equation is
d2 𝑦 d𝑦
𝑎 + 𝑏 + 𝑐𝑦 = 𝑓(𝑥)
d𝑥 2 d𝑥
where 𝑎, 𝑏, 𝑐 are constants. The homogeneous form of the above equation is
d2 𝑦 d𝑦
𝑎 2
+𝑏 + 𝑐𝑦 = 0
d𝑥 d𝑥
Equations of this form arise in the analysis of circuits. Consider the following example.
Example:
Write down the differential equation governing the current flowing in
the series 𝐿𝐶𝑅 circuit shown in the figure.
Solution:
Using Kirchhoff's voltage law and the individual laws for each
component we find
d𝑖 1
𝐿 + 𝑖𝑅 + ∫ 𝑖 d𝑡 = 𝑣(𝑡)
d𝑡 𝐶
If this equation is now differentiated with respect to 𝑡 we find
d2 𝑖 d𝑖 1 d𝑣(𝑡)
𝐿 + 𝑅 + 𝑖 =
d𝑡 2 d𝑡 𝐶 d𝑡
This is an inhomogeneous second-order differential equation, with the inhomogeneity arising from
d𝑣
the term . When the circuit components 𝐿, 𝑅 and 𝐶 are constants.
d𝑡

Alternatively, we can write the ODE governing the charge, 𝑞, at a cross section of a conductor. Given
that the voltage across the capacitor is 𝑞/𝐶. Also, since 𝑖 = 𝑑𝑞/𝑑𝑡, the above equation becomes:
d2 𝑞 d𝑞 1
𝐿 2+𝑅 + 𝑞 = 𝑣(𝑡)
d𝑡 d𝑡 𝐶

3
 Al-Nahrain University 2nd Year, 1st Semester
College of Information Engineering Engineering Mathematics I
Dept. of Info& Comms Engineering Lecture3: Solving ODE (2nd Order)

Example:
A spring has its upper end fastened to a rigid support, as shown in the Figure.
The propulsion force (weight) 𝐹P pulls the mass downward, but the spring
restoring force 𝐹s and frictional force 𝐹r pull the mass upward. The motion
starts at 𝑦 = 𝑦0 with the mass vibrating up and down. The 2nd order
inhomogeneous differential equation that describes the motion (vibrating) of
an object in a spring system is given as
𝑚𝑦 ′′ + 𝛿𝑦 ′ + 𝑘𝑦 = 𝐹(𝑡)
where y: displacement ,𝑦 ′ : velocity ,𝑦 ′′ : acceleration , 𝑚: mass
𝛿: damping constant , 𝑘: spring constant , 𝐹(𝑡): forcing function

3.2.1 Finding the complementary solution

As stated in Property 2, finding the general solution of 𝑎𝑦 ′′ + 𝑏𝑦 ′ + 𝑐𝑦 = 𝑓 is a two-stage process.
The first task is to determine the complementary solution. This is the general solution of the
corresponding homogeneous equation, that is 𝑎𝑦 ′′ + 𝑏𝑦 ′ + 𝑐𝑦 = 0. We now focus attention on the
solution of such equations.
Example:
Verify that 𝑦1 = e4𝑥 and 𝑦2 = e2𝑥 both satisfy the constant coefficient homogeneous equation
𝑑2 𝑦 𝑑𝑦
− 6 + 8𝑦 = 0 (1)
𝑑𝑥 2 𝑑𝑥
Write down the general solution of this equation.
Solution:
If 𝑦1 = e4𝑥 , differentiation yields
d𝑦1 d2 𝑦1
= 4e4𝑥 and 2
= 16e4𝑥
d𝑥 d𝑥
Substitution into Equation (1) gives
16e4𝑥 − 6(4e4𝑥 ) + 8e4𝑥 = 0
d𝑦2 d2 𝑦2
so that 𝑦1 = e4𝑥 is indeed a solution. Similarly, if 𝑦2 = e2𝑥 , then = 2e2𝑥 and = 4e2𝑥 .
d𝑥 d𝑥 2
Substitution into Equation (1) gives
4e2𝑥 − 6(2e2𝑥 ) + 8e2𝑥 = 0
so that 𝑦2 = e2𝑥 is also a solution of Equation (1). Now e2𝑥 and e4𝑥 are linearly independent
functions. So, from Property 1 we have the general solution as
𝑦H (𝑥) = 𝐴e4𝑥 + 𝐵e2𝑥

4
 Al-Nahrain University 2nd Year, 1st Semester
College of Information Engineering Engineering Mathematics I
Dept. of Info& Comms Engineering Lecture3: Solving ODE (2nd Order)

Example:
Find values of 𝑘 so that 𝑦 = e𝑘𝑥 is a solution of
d2 𝑦 d𝑦
− − 6𝑦 = 0
d𝑥 2 d𝑥
Hence state the general solution.
Solution:
As suggested, we try a solution of the form 𝑦 = e𝑘𝑥 . Differentiating and substituting into the given
equation yield
𝑘 2 e𝑘𝑥 − 𝑘e𝑘𝑥 − 6e𝑘𝑥 = 0
(𝑘 2 − 𝑘 − 6)e𝑘𝑥 = 0
The only way this equation can be satisfied for all values of 𝑥 is if
𝑘2 − 𝑘 − 6 = 0
The above equation for determining 𝑘 is called the auxiliary equation. That is,
(𝑘 − 3)(𝑘 + 2) = 0
so that 𝑘 = 3 or 𝑘 = −2. That is to say, if 𝑦 = e𝑘𝑥 is to be a solution of the differential equation 𝑘
must be either 3 or -2. We therefore have found two solutions
𝑦1 (𝑥) = e3𝑥 and 𝑦2 (𝑥) = e−2𝑥
These two functions are linearly independent and we can therefore apply Property 1 to give the
general solution:
𝑦H (𝑥) = 𝐴e3𝑥 + 𝐵e−2𝑥
Example:
Find the auxiliary equation of the differential equation
d2 𝑦 d𝑦
𝑎 2
+𝑏 + 𝑐𝑦 = 0
d𝑥 d𝑥
Solution:
We try a solution of the form 𝑦 = e𝑘𝑥 so that differentiating and substitution into the given
differential equation yield
𝑎𝑘 2 e𝑘𝑥 + 𝑏𝑘e𝑘𝑥 + 𝑐e𝑘𝑥 = 0
(𝑎𝑘 2 + 𝑏𝑘 + 𝑐 )e𝑘𝑥 = 0
Since this equation is to be satisfied for all values of 𝑥, then the required auxiliary equation is
𝑎𝑘 2 + 𝑏𝑘 + 𝑐 = 0
d2 𝑦 d𝑦
Hence, in general, the auxiliary equation of 𝑎 +𝑏 + 𝑐𝑦 = 0 is 𝑎𝑘 2 + 𝑏𝑘 + 𝑐 = 0
d𝑥 2 d𝑥

5
 Al-Nahrain University 2nd Year, 1st Semester
College of Information Engineering Engineering Mathematics I
Dept. of Info& Comms Engineering Lecture3: Solving ODE (2nd Order)

Solution of this quadratic equation gives the values of 𝑘 which we seek. Clearly the nature of the roots
will depend upon the values of 𝑎, 𝑏 and 𝑐. If 𝑏2 > 4𝑎𝑐 the roots will be real and distinct. The two
values of 𝑘 thus obtained, 𝑘1 and 𝑘2 , will allow us to write down two independent solutions:
𝑦1 (𝑥) = e𝑘1 𝑥 and 𝑦2 (𝑥) = e𝑘2 𝑥

Hence, if the auxiliary equation has real, distinct roots 𝑘1 and 𝑘2 , the complementary solution will be
𝒚(𝒙) = 𝑨𝐞𝒌𝟏 𝒙 + 𝑩𝐞𝒌𝟐 𝒙
On the other hand, if 𝑏2 = 4𝑎𝑐 the two roots of the auxiliary equation will be equal and this method
will therefore only yield one independent solution. In this case, special treatment is required. If 𝑏2 <
4𝑎𝑐 the two roots of the auxiliary equation will be complex, that is 𝑘1 and 𝑘2 will be complex
numbers. The procedure for dealing with such cases will become apparent in the following examples.
Example:
Find the general solution of
d2 𝑦 d𝑦
+ 3 − 10𝑦 = 0
d𝑥 2 d𝑥
Find the particular solution which satisfies the conditions 𝑦(0) = 1 and 𝑦 ′ (0) = 1.
Solution:
By letting 𝑦 = e𝑘𝑥 , so that, the auxiliary equation is found to be
𝑘 2 + 3𝑘 − 10 = 0
(𝑘 − 2)(𝑘 + 5) = 0
so that 𝑘 = 2 and 𝑘 = −5. Thus, there exist two solutions, 𝑦1 = e2𝑥 and 𝑦2 = e−5𝑥 . From Property 1
we can write the general solution as
𝑦 = 𝐴e2𝑥 + 𝐵e−5𝑥
To find the particular solution we must now impose the given conditions:
𝑦(0) = 1 gives 1=𝐴+𝐵
𝑦 ′ (0) = 1 gives 1 = 2𝐴 − 5𝐵
6 1 6 1
from which 𝐴 = and 𝐵 = . Finally, the required particular solution is 𝑦 = e2𝑥 + e−5𝑥 .
7 7 7 7

Example: Find the general solution of

d2 𝑦
+ 4𝑦 = 0
d𝑥 2
Solution:
As before, let 𝑦 = e𝑘𝑥 so that the auxiliary equation is easily found to be
𝑘 2 + 4 = 0 → 𝑘 = ±2j
that is, we have complex roots. The two independent solutions of the equation are thus

6
 Al-Nahrain University 2nd Year, 1st Semester
College of Information Engineering Engineering Mathematics I
Dept. of Info& Comms Engineering Lecture3: Solving ODE (2nd Order)

𝑦1 (𝑥) = e2j𝑥 and 𝑦2 (𝑥) = e−2j𝑥

so that the general solution can be written in the form
𝑦(𝑥) = 𝐴e2j𝑥 + 𝐵e−2j𝑥
However, in cases such as this, it is usual to rewrite the solution in the following way. The Euler's
relations give
e2j𝑥 = cos 2𝑥 + jsin 2𝑥 and e−2j𝑥 = cos 2𝑥 − jsin 2𝑥
so that
𝑦(𝑥) = 𝐴(cos 2𝑥 + jsin 2𝑥) + 𝐵(cos 2𝑥 − jsin 2𝑥)
If we now relabel the constants such that
𝐴 + 𝐵 = 𝐶 and 𝐴j − 𝐵j = 𝐷
we can write the general solution in the form
𝑦(𝑥) = 𝐶cos 2𝑥 + 𝐷sin 2𝑥

If the auxiliary equation has complex roots 𝛼 + 𝛽𝑗 and 𝛼 − 𝛽𝑗, the complementary solution will be
𝒚(𝒙) = 𝐞𝜶𝒙 (𝑨𝐜𝐨𝐬 𝜷𝒙 + 𝐁𝐬𝐢𝐧 𝜷𝒙)

Engineering Application
Oscillating mass-spring system
Consider the simple mechanical problem of a mass resting on a smooth frictionless table. A spring is
attached to the mass and an adjacent anchor point as shown in the Figure below. The spring is
capable of being compressed as well as stretched. When the spring is neither compressed nor
stretched the mass is located at 𝑥 = 0.

If the mass is pulled in the 𝑥 direction and let go, it will oscillate about the 𝑥 = 0 position. We wish
to find the position of the mass, 𝑥, as a function of time, 𝑡. Differential equations are needed to
describe this problem fully.
Newton's second law states that if a force 𝐹 is applied to a body of mass 𝑚 then the motion of the
body is governed by 𝐹 = 𝑚𝑎, where 𝑎 is the acceleration. Applying Newton's second law, and
d2 𝑥
noting that 𝑎 = , we obtain
d𝑡 2
d2 𝑥
𝐹=𝑚 2
d𝑡
7
 Al-Nahrain University 2nd Year, 1st Semester
College of Information Engineering Engineering Mathematics I
Dept. of Info& Comms Engineering Lecture3: Solving ODE (2nd Order)

The force, 𝐹, is provided by the spring. The force exerted by a spring is given by Hooke's law, which
states that the force is proportional to the extension or compression of the spring,
𝐹 = −𝑘𝑥
where 𝑘 is the spring constant for the spring in use. The minus sign is required so that when the
spring is stretched (𝑥 > 0) the force is in the negative 𝑥 direction. When the spring is compressed
(𝑥 < 0) the force is in the positive 𝑥 direction.
d2 𝑥
If the table is sufficiently smooth and if there are no other external forces acting, then 𝑚 2 =
d𝑡
−𝑘𝑥.
Therefore, the differential equation that governs motion in the system is
d2 𝑥
𝑚 2 + 𝑘𝑥 = 0
d𝑡
We write this as
d2 𝑥 𝑘
+ 𝑥=0
d𝑡 2 𝑚
𝑘
Let = 𝜔2 , giving
𝑚
d2 𝑥
2
+ 𝜔2 𝑥 = 0
d𝑡
Using the technique illustrated in the previous Example, we obtain
𝑥(𝑡) = 𝐴cos 𝜔𝑡 + 𝐵sin 𝜔𝑡
where 𝐴, 𝐵 are constants. Note this solution may also be expressed as a single wave
𝑥 = 𝐶cos (𝜔𝑡 + 𝜙)
where 𝐶 and 𝜙 are constants.
We note that this solution is sinusoidal and oscillates with time. It gives the position of the mass at
a given point in time. The constants 𝐶 and 𝜙 depend on the position and velocity of the mass when
it is released. These are known as the initial conditions of the differential equation. It is intuitive
that an equation that describes the position of the mass at a given time must take these into
account.
Mathematically describing or modelling systems like this one using differential equations is an
extremely important discipline. Mathematical models of mechanical, electrical and other
subsystems can be linked together and as a result whole systems can be accurately characterized.
Example:
Find the general solution of 𝑦 ′′ + 2𝑦 ′ + 4𝑦 = 0.
Solution:
The auxiliary equation is 𝑘 2 + 2𝑘 + 4 = 0. This equation has complex roots given by

−2 ± √4 − 16 −2 ± √12j
𝑘 = =
2 2
= −1 ± √3j
Try to find the general solution to be:
𝑦 = e−𝑥 (𝐴cos √3𝑥 + 𝐵sin √3𝑥)

8
 Al-Nahrain University 2nd Year, 1st Semester
College of Information Engineering Engineering Mathematics I
Dept. of Info& Comms Engineering Lecture3: Solving ODE (2nd Order)

Important Note:
If the auxiliary equation has two equal roots, 𝑘1 , the complementary solution is
𝑦 = 𝐴e𝑘1𝑥 + 𝐵𝑥e𝑘1𝑥
For example: 𝑦 ′′ + 8𝑦 ′ + 16𝑦 = 0
Exercises:
Obtain the general solutions, that is the d2 𝑦 d𝑦
(g) −2 +𝑦 =0
d𝑥 2 d𝑥
complementary solutions, of the following d2 𝑦 d𝑦
(h) + + 5𝑦 = 0
homogeneous equations: d𝑡 2 d𝑡
d2 𝑦 d𝑦
d2 𝑦 d𝑦 (i) + − 2𝑦 = 0
(a) −3 + 2𝑦 = 0 d𝑥 2 d𝑥
d𝑥 2 d𝑥 d2 𝑦
d2 𝑦 d𝑦 (j) + 9𝑦 = 0
(b) +7 + 6𝑦 = 0 d𝑥 2
d𝑥 2 d𝑥 d2 𝑦 d𝑦
d2 𝑥 d𝑥 (k) −2 =0
d𝑥 2 d𝑥
(c) 2
+5 + 6𝑥 = 0 d2 𝑥
d𝑡 d𝑡
d2 𝑦
(l) − 16𝑥 = 0
d𝑦 d𝑡 2
(d) 2
+2 +𝑦=0
d𝑡 d𝑡
d2 𝑦 d𝑦
(e) −4 + 4𝑦 = 0
d𝑥 2 d𝑥
d2 𝑦 d𝑦
(f) + + 8𝑦 = 0
d𝑡 2 d𝑡
Solutions:
(a) 𝑦 = 𝐴e𝑥 + 𝐵e2𝑥 (g) 𝑦 = 𝐴e𝑥 + 𝐵𝑥e𝑥
(b) 𝑦 = 𝐴e−𝑥 + 𝐵e−6𝑥 (h) 𝑦 = e−0.5𝑡 (𝐴cos 2.18𝑡 + 𝐵sin 2.18𝑡)
(c) 𝑥 = 𝐴e−2𝑡 + 𝐵e−3𝑡 (i) 𝑦 = 𝐴e−2𝑥 + 𝐵e𝑥
(d) 𝑦 = 𝐴e−𝑡 + 𝐵𝑡e−𝑡 (j) 𝑦 = 𝐴cos 3𝑥 + 𝐵sin 3𝑥 (k) 𝑦 = 𝐴 + 𝐵e2𝑥
(e) 𝑦 = 𝐴e2𝑥 + 𝐵𝑥e2𝑥 (k) 𝑥 = 𝐴 + 𝐵e2x
(f) 𝑦 = e−0.5𝑡 (𝐴cos 2.78𝑡 + 𝐵sin 2.78𝑡) (l) 𝑥 = 𝐴e4𝑡 + 𝐵e−4𝑡
3.2.2 Finding a particular solution
We stated in Property 2 (Page no.1) that the general solution of an inhomogeneous equation is the
sum of the complementary solution and a particular solution. We have seen how to find the
complementary solution in the case of a constant coefficient equation. We shall now deal with the
problem of finding a particular solution. Recall that the particular solution is any solution of the
inhomogeneous equation. There are a number of advanced techniques available for finding such
solutions we shall adopt a simpler strategy. Since any solution will do, we shall try to find such a
solution by a combination of educated guesswork and trial and error.
Example:
Find the general solution of the equation
d2 𝑦 d𝑦
2
− − 6𝑦 = e2𝑥
d𝑥 d𝑥

9
 Al-Nahrain University 2nd Year, 1st Semester
College of Information Engineering Engineering Mathematics I
Dept. of Info& Comms Engineering Lecture3: Solving ODE (2nd Order)

Solution: The complementary solution for this equation has already been shown in the Example (in
page no.4) to be
𝑦H = 𝐴e3𝑥 + 𝐵e−2𝑥
We shall attempt to find a solution of the inhomogeneous problem by trying a function of the same
form as that on the r.h.s. In particular, let us try 𝑦P (𝑥) = 𝛼e2𝑥 , where 𝛼 is a constant that we shall
now determine. If 𝑦P (𝑥) = 𝛼e2𝑥 then so that differentiating and substitution into the given
differential equation yield
4𝛼e2𝑥 − 2𝛼e2𝑥 − 6𝛼e2𝑥 = e2𝑥
−4𝛼e2𝑥 = e2𝑥
so that 𝑦P will be a solution if 𝛼 is chosen so that −4𝛼 = 1, that is 𝛼 = −1/4. Therefore, the
particular solution is 𝑦P (𝑥) = −e2𝑥 /4. From Property 2 the general solution of the inhomogeneous
equation is found by summing this particular solution and the complementary solution
1
𝑦(𝑥) = 𝐴e3𝑥 + 𝐵e−2𝑥 − e2𝑥
4
Example:
Obtain a particular solution of the equation
𝑑2 𝑦 𝑑𝑦
2
−6 + 8𝑦 = 𝑥
𝑑𝑥 𝑑𝑥
Solution: In the last example, we found that a fruitful approach was to assume a solution in the same
form as that on the r.h.s. Suppose we assume a solution 𝑦P (𝑥) = 𝛼𝑥 and proceed to determine 𝛼.
d𝑦P d2 𝑦P
This approach will actually fail, but let us see why. If 𝑦P (𝑥) = 𝛼𝑥 then = 𝛼 and = 0.
d𝑥 d𝑥 2
Substitution into the differential equation yields
0 − 6𝛼 + 8𝛼𝑥 = 𝑥
and 𝛼 ought now to be chosen so that this expression is true for all 𝑥. If we equate the coefficients of
1
𝑥 we find 8𝛼 = 1 so that 𝛼 = , but with this value of 𝛼 the constant terms are inconsistent. Clearly a
8
particular solution of the form 𝛼𝑥 is not possible. The problem arises because differentiation of the
term 𝛼𝑥 produces constant terms which are unbalanced on the r.h.s. So, we try a solution of the form
𝑦P (𝑥) = 𝛼𝑥 + 𝛽
d𝑦P d2 𝑦P
with 𝛼, 𝛽 constants. Proceeding as before, = 𝛼, = 0. Substitution in the differential equation
d𝑥 d𝑥 2
now gives
0 − 6𝛼 + 8(𝛼𝑥 + 𝛽) = 𝑥
Equating coefficients of 𝑥 we find
8𝛼 = 1
and equating constant terms we find
−6𝛼 + 8𝛽 = 0

10
 Al-Nahrain University 2nd Year, 1st Semester
College of Information Engineering Engineering Mathematics I
Dept. of Info& Comms Engineering Lecture3: Solving ODE (2nd Order)

1
−6 ( ) + 8𝛽 = 0 → 𝛽 = 3/32
8
𝑥 3
The required particular solution is 𝑦P (𝑥) = + .
8 32

Experience leads to the trial solutions (to find particular solution) suggested in the Table below.

𝒇(𝒙) Trial solution 𝒚𝐏 (𝒙)

𝒃𝐞 𝒌𝒙
𝛼e𝑘𝑥
Constant 𝒃 Constant 𝛼
th
n degree polynomial in 𝒙 𝛼𝑛 𝑥 + 𝛼𝑛−1 𝑥 𝑛−1 + ⋯ + 𝛼1 𝑥 + 𝛼0
𝑛

𝒃𝐜𝐨𝐬 𝒌𝒙 𝛼cos 𝑘𝑥 + 𝛽sin 𝑘𝑥

𝒃𝐬𝐢𝐧 𝒌𝒙 𝑎cos 𝑘𝑥 + 𝑏sin 𝑘𝑥
Example: Find a particular solution for the equation
d2 𝑦 d𝑦
− 6 + 8𝑦 = 3cos 𝑥
d𝑥 2 d𝑥
Solution: We shall try a solution of the form
𝑦P (𝑥) = 𝛼cos 𝑥 + 𝛽sin 𝑥
Differentiating 𝑦P (𝑥) and substitution into the differential equation give
(−𝛼cos 𝑥 − 𝛽sin 𝑥) − 6(−𝛼sin 𝑥 + 𝛽cos 𝑥) + 8(𝛼cos 𝑥 + 𝛽sin 𝑥) = 3cos 𝑥
Equating coefficients of cos 𝑥 we find
−𝛼 − 6𝛽 + 8𝛼 = 3
while those of sin 𝑥 give
−𝛽 + 6𝛼 + 8𝛽 = 0
21 18
Solving Equations simultaneously gives 𝛼 = and 𝛽 = − , so that the particular solution is
85 85

21 18
𝑦P (𝑥) = cos 𝑥 − sin 𝑥
85 85
Engineering Application
An LC circuit with sinusoidal input
Write down the differential equation governing the flow of
current 𝑖 in a series 𝐿𝐶 circuit when subject to an applied voltage
𝑣(𝑡) = 𝑉0 sin 𝜔𝑡. Then, obtain its general solution.

Solution
Kirchhoff's voltage law and the component laws give
d𝑖 1
𝐿 + ∫ 𝑖 d𝑡 = 𝑉0 sin 𝜔𝑡
d𝑡 𝐶

11
 Al-Nahrain University 2nd Year, 1st Semester
College of Information Engineering Engineering Mathematics I
Dept. of Info& Comms Engineering Lecture3: Solving ODE (2nd Order)

To avoid processes of differentiation and integration in the same equation let us differentiate this
equation with respect to 𝑡. This yields
d2 𝑖 1
𝐿 2 + 𝑖 = 𝜔𝑉0 cos 𝜔𝑡
d𝑡 𝐶
d2 𝑖 𝑖
The homogeneous equation is 𝐿 2 + = 0. Letting 𝑖 = e𝑘𝑡 we find the auxiliary equation is 𝐿𝑘 2 +
d𝑡 𝐶
1 j 1
= 0 so that 𝑘 = ± . Therefore, using the result of Example 19.22, with 𝛼 = 0 and 𝛽 = , the
𝐶 √𝐿𝐶 √𝐿𝐶
complementary solution is
𝑡 𝑡
𝑖 = 𝐴cos + 𝐵sin
√𝐿𝐶 √𝐿𝐶
To find a particular solution, try 𝑖 = 𝐸cos 𝜔𝑡 + 𝐹sin 𝜔𝑡, where 𝐸 and 𝐹 are constants. We find
d𝑖
= −𝜔𝐸sin 𝜔𝑡 + 𝜔𝐹cos 𝜔𝑡
d𝑡
d2 𝑖
= −𝜔2 𝐸cos 𝜔𝑡 − 𝜔2 𝐹sin 𝜔𝑡
d𝑡 2
Substitution into the inhomogeneous equation yields
1
𝐿(−𝜔2 𝐸cos 𝜔𝑡 − 𝜔2 𝐹sin 𝜔𝑡 ) + (𝐸cos 𝜔𝑡 + 𝐹sin 𝜔𝑡) = 𝑉0 𝜔cos 𝜔𝑡
𝐶
Equating coefficients of sin 𝜔𝑡 gives
𝐹
−𝜔2 𝐿𝐹 + = 0
𝐶
Equating coefficients of cos 𝜔𝑡 gives
𝐸
−𝜔2 𝐿𝐸 + = 𝑉0 𝜔
𝐶
0 𝐶𝑉 𝜔 0 𝐶𝑉 𝜔
so that 𝐹 = 0 and 𝐸 = . It follows that the particular solution is 𝑖 = cos 𝜔𝑡. Finally,
1−𝜔2 𝐿𝐶 1−𝜔2 𝐿𝐶
the general solution is
𝑡 𝑡 𝐶𝑉0 𝜔cos 𝜔𝑡
𝑖 = 𝐴cos + 𝐵sin +
√𝐿𝐶 √𝐿𝐶 1 − 𝜔 2 𝐿𝐶
The term ‘’the zero-input response’’ is the solution when 𝑣(𝑡) = 0. Also, the term ‘’the zero-state
response’’ represents the solution of the inhomogeneous equation subject to certain initial
d𝑖
conditions (e.g., 𝑖 = 0 and = 0 at 𝑡 = 0). It corresponds to there being no initial energy in the
d𝑡
circuit.

Engineering Application
Parallel RLC circuit
Write down the second-order differential equation governing
the flow of inductor current 𝑖 in a parallel 𝑅𝐿𝐶 circuit when
subject to an applied a current source 𝑖S (𝑡).

12
 Al-Nahrain University 2nd Year, 1st Semester
College of Information Engineering Engineering Mathematics I
Dept. of Info& Comms Engineering Lecture3: Solving ODE (2nd Order)

Suppose 𝐿 = 10H, 𝑅 = 10Ω, 𝐶 = 0.1 F and 𝑖S (𝑡) = e−2𝑡 and that the initial conditions are 𝑖 = 1
d𝑖
and = 2 when 𝑡 = 0.
d𝑡
(a) Obtain the solution of this equation subject to the given initial conditions and hence state the
inductor current 𝑖.
(b) Obtain the zero-input response. This is the solution when 𝑖S (𝑡) = 0.
(c) Obtain the zero-state response. This is the solution of the inhomogeneous equation subject to
d𝑖
the conditions 𝑖 = 0 and = 0 at 𝑡 = 0.
d𝑡
(d) Show that the solution in (a) is the sum of the zero-input response and the zero-state response.
Solution: the 2nd order inhomogeneous ODE is obtained as:
d2 𝑖 𝐿 d𝑖
𝐿𝐶 2 + + 𝑖 = 𝑖S (𝑡) for 𝑡 ⩾ 0
d𝑡 𝑅 d𝑡
(a) With the given parameter values the differential equation becomes
d2 𝑖 d𝑖
2
+ + 𝑖 = e−2𝑡
d𝑡 d𝑡
It is first necessary to find the complementary solution. Letting 𝑖 = e𝑘𝑡 , the auxiliary equation is
𝑘 2 + 𝑘 + 1 = 0 which has complex solutions
1 √3
𝑘=− ± j
2 2
The complementary solution is therefore
√3 √3
𝑖 = e−𝑡/2 (𝐴cos 𝑡 + 𝐵sin 𝑡)
2 2
For a particular solution we try a solution of the form 𝑖 = 𝛼e−2𝑡 . Substitution into the differential
equation gives
4𝛼e−2𝑡 − 2𝛼e−2𝑡 + 𝛼e−2𝑡 = e−2𝑡
so that
1
3𝛼 = 1, that is 𝛼 =
3
1 −2𝑡
Hence a particular solution is 𝑖 = e . The general solution is the sum of the complementary
3
solution and the particular solution:
√3 √3 1
𝑖 = e−𝑡/2 (𝐴cos 𝑡 + 𝐵sin 𝑡) + e−2𝑡
2 2 3
We now apply the initial conditions to find the constants 𝐴 and 𝐵. Given 𝑖 = 1 when 𝑡 = 0 means
1 2
1 = 𝐴 + so that 𝐴 =
3 3
d𝑖
To apply the second condition, we need to find :
d𝑡

13
 Al-Nahrain University 2nd Year, 1st Semester
College of Information Engineering Engineering Mathematics I
Dept. of Info& Comms Engineering Lecture3: Solving ODE (2nd Order)

d𝑖 √3 √3 √3 √3
= e−𝑡/2 (− 𝐴sin 𝑡+ 𝐵cos 𝑡)
d𝑡 2 2 2 2
1 √3 √3 2
− e−𝑡/2 (𝐴cos 𝑡 + 𝐵sin 𝑡) − e−2𝑡
2 2 2 3
d𝑖
Given = 2 when 𝑡 = 0 means
d𝑡
√3 1 2
2= 𝐵 − (𝐴) −
2 2 3
6
𝐵 = = 2√3
√3
Finally, the required particular solution which gives the current through the inductor at any time is
2 √3 √3 1
𝑖 = e−𝑡/2 ( cos 𝑡 + 2√3sin 𝑡) + e−2𝑡
3 2 2 3
(b) The zero-input response is obtained by ignoring the source term 𝑖S (𝑡). From (a) we see that this
is just the complementary solution:
√3 √3
𝑖 = e−𝑡/2 (𝐴cos 𝑡 + 𝐵sin 𝑡)
2 2
5
Applying the given initial conditions to this solution gives 𝐴 = 1 and 𝐵 = and so the zero-input
√3
response is
√3 5 √3
𝑖 = e−𝑡/2 (cos
𝑡+ sin 𝑡)
2 √3 2
(c) The zero-state response of the inhomogeneous equation is found by applying the conditions 𝑖 =
d𝑖
0 and = 0 at 𝑡 = 0 to the general solution already obtained in (a). It is straightforward to show
d𝑡
1 1
that 𝐴 = − and 𝐵 = . So, the zero-state response is
3 √3
1 √3 1 √3 1
𝑖 = e−𝑡/2 (− cos 𝑡+ sin 𝑡) + e−2𝑡
3 2 √3 2 3
(d) Inspection of the previous working shows that the particular solution obtained in (a) is the sum
of the zero-state response and zero-input response.

In some examples, terms which form part of the complementary solution also appear in the
inhomogeneous term. This gives rise to an additional complication. Consider the following Example.
Example
Consider the equation 𝑦 ′′ − 𝑦 ′ − 6𝑦 = e3𝑥 . It is straightforward to show that the complementary
solution is
𝑦 = 𝐴e3𝑥 + 𝐵e−2𝑥
Find a particular solution and deduce the general solution.

14
 Al-Nahrain University 2nd Year, 1st Semester
College of Information Engineering Engineering Mathematics I
Dept. of Info& Comms Engineering Lecture3: Solving ODE (2nd Order)

Solution:
Suppose we try to find a particular solution by using a trial solution of the form 𝑦P = 𝛼e3𝑥 .
Substitution into the l.h.s. of the inhomogeneous equation yields
9𝛼e3𝑥 − 3𝛼e3𝑥 − 6𝛼e3𝑥 which simplifies to 0
so that 𝛼e3𝑥 is clearly not a solution of the inhomogeneous equation. The reason is that e3𝑥 is part of
the complementary solution and so causes the l.h.s. to vanish. To obtain a particular solution in such a
case, we carry out the procedure required when the auxiliary equation has equal roots. That is, we try
𝑦P = 𝛼𝑥e3𝑥 . We find
𝑦 ′ = 𝛼e3𝑥 (3𝑥 + 1) and 𝑦 ′′ = 𝛼e3𝑥 (9𝑥 + 6)
Substitution into the inhomogeneous equation yields
𝛼e3𝑥 (9𝑥 + 6) − 𝛼e3𝑥 (3𝑥 + 1) − 6𝛼𝑥e3𝑥 = e3𝑥
5𝛼e3𝑥 = e3𝑥
𝑥e3𝑥
So that 𝛼 = 1/5. Finally, the required particular solution is 𝑦P = . The general solution is then
5

3𝑥 −2𝑥
𝑥e3𝑥
𝑦 = 𝐴e + 𝐵e +
5
Exercises:
1- Find the general solution of the 2- Find a particular solution for the equations:
d2 𝑥 d𝑥
following equations: (a) −3 + 2𝑥 = 5e3𝑡
d𝑡 2 d𝑡
d2 𝑥 d𝑥 d2 𝑥
(a) −2 − 3𝑥 = 6 (b) − 𝑥 = 4e−2𝑡
d𝑡 2 d𝑡 d𝑡 2
d2 𝑦 d𝑦 d2 𝑦 d𝑦
(b) +5 + 4𝑦 = 8 (c) 2 + + 𝑦 = 1 + 𝑥
d𝑥 2 d𝑥 d𝑥 d𝑥
d2 𝑦 d𝑦
(c) +5 + 6𝑦 = 2𝑡 3- Obtain the general solution of the equation.
d𝑡 2 d𝑡
d2 𝑥 d𝑥 d2 𝑦 d𝑦
(d) + 11 + 30𝑥 = 8𝑡 2
+3 + 2𝑦 = 10 cos 2𝑥
d𝑡 2 d𝑡 d𝑥 d𝑥
d2 𝑦 d𝑦
(e) +2 + 3𝑦 = 2sin 2𝑥 Then, find the particular solution satisfying 𝑦(0) = 1,
d𝑥 2 d𝑥 d𝑦
d2 𝑦 d𝑦 (0) = 0
(f) + + 𝑦 = 4cos 3𝑡 d𝑥
d𝑡 2 d𝑡
d2 𝑦
4- For the circuit in the figure
(g) + 9𝑦 = 4e8𝑥 show that
d𝑥 2
d2 𝑥 d2 𝑖 d𝑖
(h) 2
− 16𝑥 = 9e6𝑡 𝑅𝐶𝐿 22 + 𝐿 2 + 𝑅𝑖2 = 𝐸(𝑡)
d𝑡 d𝑡 d𝑡
(i) 𝑦 ′′ − 𝑦 ′ − 2𝑦 = 6 If 𝐿 = 1mH, 𝑅 = 10Ω, 𝐶 = 1𝜇F
and 𝐸(𝑡) = 2sin 100𝜋𝑡, find the complementary solution.

Solutions:
1- (a) 𝑥 = 𝐴e−𝑡 + 𝐵e3𝑡 − 2 2-(a) 𝑥 = 2.5e3𝑡
4
(b) 𝑦 = 𝐴e−𝑥 + 𝐵e−4𝑥 + 2 (b) 𝑥 = e−2𝑡
3

15
 Al-Nahrain University 2nd Year, 1st Semester
College of Information Engineering Engineering Mathematics I
Dept. of Info& Comms Engineering Lecture3: Solving ODE (2nd Order)

𝑡 5
(c) 𝑦 = 𝐴e−2𝑡 + 𝐵e−3𝑡 + − (c) 𝑥
3 18
(d) 𝑥 = 𝐴e−6𝑡 + 𝐵e−5𝑡 + 0.267𝑡 − 3 1 3 3
0.0978 3-𝐴e−2𝑥 + 𝐵e−𝑥 + sin 2𝑥 − cos 2𝑥, e−2𝑥 + sin 2𝑥 −
2 2 2 2
1
(e) 𝑦 = e−𝑥 [𝐴sin √2𝑥 + cos 2𝑥
8 2 2
𝐵cos √2𝑥] − cos 2𝑥 − sin 2𝑥
17 17
(f) 𝑦 = e−0.5𝑡 (𝐴cos 0.866𝑡 + 4- 𝐴e−11270𝑡 + 𝐵e−88730𝑡
𝐵sin 0.866𝑡) − 0.438cos 3𝑡 +
0.164sin 3𝑡
(g) 𝑦 = 𝐴cos 3𝑥 + 𝐵sin 3𝑥 +
8𝑥
0.0548e
9
(h) 𝑥 = 𝐴e4𝑡 + 𝐵e−4𝑡 + e6𝑡
20
(i) 𝐴e2𝑥 + 𝐵e−𝑥 − 3

3.3 Series Solutions of ODE

3.3.1 Power Series Solution
We first need to review the three important Maclaurin series (A special commonly used case of a
Taylor series) as stated below:

We now introduce a technique for finding solutions of differential equations that can be expressed in
the form of a power series
∞

𝑦 = ∑ 𝑎𝑚 𝑥 𝑚 = 𝑎0 + 𝑎1 𝑥 + 𝑎2 𝑥 2 + ⋯
𝑚=0
where, 𝑎𝑚 , are constants whose values need to be found.
With the given power series form of 𝑦, we can differentiate term by term to obtain power series for
d𝑦 d2 𝑦
and . By substituting these series into certain classes of differential equation we can determine
d𝑥 d𝑥 2
the constants 𝑎𝑚 , for 𝑚 = 0,1,2,3 … and thus obtain the power series solution for 𝑦. This technique
of representing a solution in the form of a power series paves the way for the introduction of an
important family of differential equations, known as Bessel's equations, which can be solved by this

16
 Al-Nahrain University 2nd Year, 1st Semester
College of Information Engineering Engineering Mathematics I
Dept. of Info& Comms Engineering Lecture3: Solving ODE (2nd Order)

method. The power series solutions so formed are known as Bessel functions which will be introduced
later.
Example:
The general solution of the differential equation 𝑦 ′′ + 𝑦 = 0 was shown in Example (page no.2) to be
𝑦 = 𝐴cos 𝑥 + 𝐵sin 𝑥
where 𝐴 and 𝐵 are arbitrary constants. Obtain this result by looking for a solution of the equation in
the form of a power series 𝑦 = ∑∞ 𝑚
𝑚=0 𝑎𝑚 𝑥 .

Solution: We let
∞

𝑦 = ∑ 𝑎𝑚 𝑥 𝑚 = 𝑎0 + 𝑎1 𝑥 + 𝑎2 𝑥 2 + 𝑎3 𝑥 3 + 𝑎4 𝑥 4 + 𝑎5 𝑥 5 + ⋯
𝑚=0

from which, by differentiating with respect to 𝑥,

∞
′
d𝑦
𝑦 = = ∑ 𝑚𝑎𝑚 𝑥 𝑚−1 = 𝑎1 + 2𝑎2 𝑥 + 3𝑎3 𝑥 2 + 4𝑎4 𝑥 3 + 5𝑎5 𝑥 4 + ⋯
d𝑥
𝑚=0
and
∞
′′
d2 𝑦
𝑦 = = ∑ (𝑚 − 1)𝑚𝑎𝑚 𝑥 𝑚−2 = 2𝑎2 + (2)(3)𝑎3 𝑥 + (3)(4)𝑎4 𝑥 2 + (4)(5)𝑎5 𝑥 3 + ⋯
d𝑥 2
𝑚=0

These expressions for 𝑦 and 𝑦 ′′ are substituted into the given differential equation 𝑦 ′′ + 𝑦 = 0. Thus
(2𝑎2 + (2)(3)𝑎3 𝑥 + (3)(4)𝑎4 𝑥 2 + (4)(5)𝑎5 𝑥 3 + ⋯ ) +
(𝑎0 + 𝑎1 𝑥 + 𝑎2 𝑥 2 + 𝑎3 𝑥 3 + 𝑎4 𝑥 4 + 𝑎5 𝑥 5 + ⋯ ) = 0.
We now use the technique of equating coefficients. Equating the constant terms:
1
2𝑎2 + 𝑎0 = 0 so that 𝑎2 = − 𝑎0 .
2
Equating coefficients of 𝑥 :
1
(2) (3) 𝑎3 + 𝑎1 = 0 so that 𝑎3 = − 𝑎 .
(2)(3) 1
Equating coefficients of 𝑥 2 :
1 1
(3) (4) 𝑎4 + 𝑎2 = 0 so that 𝑎4 = − 𝑎2 = 𝑎 .
(3)(4) (2)(3)(4) 0
Equating coefficients of 𝑥 3 :
1 1
(4) (5) 𝑎5 + 𝑎3 = 0 so that 𝑎5 = − 𝑎3 = 𝑎 .
(4)(5) (2)(3)(4)(5) 1
Note that by using factorial notation we can express these coefficients concisely as:
1 1 1 1
𝑎2 = − 𝑎0 , 𝑎3 = − 𝑎1 , 𝑎4 = 𝑎0 , 𝑎5 = 𝑎1 .
2! 3! 4! 5!

17
 Al-Nahrain University 2nd Year, 1st Semester
College of Information Engineering Engineering Mathematics I
Dept. of Info& Comms Engineering Lecture3: Solving ODE (2nd Order)

By equating higher order terms in the same way further coefficients 𝑎𝑚 can be determined and the
power series 𝑦 = ∑∞ 𝑚
𝑚=0 𝑎𝑚 𝑥 can then be written out explicitly. Observe that all coefficients 𝑎𝑚 are
multiples of 𝑎0 when 𝑚 is even, and multiples of 𝑎1 when 𝑚 is odd. We can therefore collect terms
together as follows:
∞

𝑦 = ∑ 𝑎𝑚 𝑥 𝑚
𝑚=0
= 𝑎0 + 𝑎1 𝑥 + 𝑎2 𝑥 2 + 𝑎3 𝑥 3 + 𝑎4 𝑥 4 + 𝑎5 𝑥 5 + ⋯
1 1 1 1
= 𝑎0 + 𝑎1 𝑥 − 𝑎0 𝑥 2 − 𝑎1 𝑥 3 + 𝑎0 𝑥 4 + 𝑎1 𝑥 5 + ⋯
2! 3! 4! 5!
1 1 1 1
= 𝑎0 (1 − 𝑥 2 + 𝑥 4 − ⋯ ) + 𝑎1 (𝑥 − 𝑥 3 + 𝑥 5 − ⋯ )
2! 4! 3! 5!
Observe that the series in the brackets above are the power series expansions of cos 𝑥 and sin 𝑥 as
stated earlier. Hence, we can write the solution as
𝑦 = 𝑎0 cos 𝑥 + 𝑎1 sin 𝑥.
We see that this is simply the general solution of the differential equation as given in the question but
where the coefficients 𝑎0 and 𝑎1 are the arbitrary constants.
Example:
Find the general solution of the differential equation 𝑦 ′′ − 2𝑥𝑦 ′ + y = 0 by looking for a solution of
the equation in the form of a power series 𝑦 = ∑∞ 𝑚=0 𝑎𝑚 𝑥 .
𝑚

Solution:
We first find the expressions for 𝑦 and 𝑦 ′′ (as in the previous example). Then, we substitute 𝑦 and 𝑦 ′′
into the given differential equation. Thus
(2𝑎2 + (2)(3)𝑎3 𝑥 + (3)(4)𝑎4 𝑥 2 + (4)(5)𝑎5 𝑥 3 + ⋯ ) −
2𝑥 (𝑎1 + 2𝑎2 𝑥 + 3𝑎3 𝑥 2 + 4𝑎4 𝑥 3 + 5𝑎5 𝑥 4 + ⋯ )
+(𝑎0 + 𝑎1 𝑥 + 𝑎2 𝑥 2 + 𝑎3 𝑥 3 + 𝑎4 𝑥 4 + 𝑎5 𝑥 5 + ⋯ ) = 0.
Equating the constant terms: (to be completed in class)
?
Equating coefficients of 𝑥 :
?
Equating coefficients of 𝑥 2 :
?
Equating coefficients of 𝑥 3 :
?

18
 Al-Nahrain University 2nd Year, 1st Semester
College of Information Engineering Engineering Mathematics I
Dept. of Info& Comms Engineering Lecture3: Solving ODE (2nd Order)

We can therefore collect terms together as follows: (to be completed in class)

∞

𝑦 = ∑ 𝑎𝑚 𝑥 𝑚
𝑚=0
= 𝑎0 + 𝑎1 𝑥 + 𝑎2 𝑥 2 + 𝑎3 𝑥 3 + 𝑎4 𝑥 4 + 𝑎5 𝑥 5 + ⋯

=
=
Exercise:
Use the power series form 𝑦 = ∑∞ 𝑚
𝑚=0 𝑎𝑚 𝑥 to find the general solution of the differential equations:

(a) 𝑦 ′′ + 𝑥𝑦 ′ + y = 0 (b) (1 − 𝑥 2 )𝑦 ′′ − 6𝑥𝑦 ′ − 4y = 0, −1 < 𝑥 < 1

3.3.2 Solutions for Bessel’s Equation

A linear differential equation that has important applications in the theory of wave propagation,
particularly when using cylindrical or spherical polar coordinates, is known as Bessel’s equation. In
fact, the term Bessel’s equation represents a whole family of equations having the form
𝑥 2 𝑦 ′′ + 𝑥 𝑦 ′ + (𝑥 2 − 𝑣 2 )y = 0
Here 𝑥 and 𝑦 are the independent variable and dependent variable respectively. The third quantity 𝑣,
termed a parameter, is known as the order of the equation.1 As we vary the parameter 𝑣 we obtain
the family of differential equations. Note that 𝑣 can be any real number. In Sections 19.8.1 and 19.8.2
we will solve Bessel’s equation when 𝑣 = 0 and 𝑣 = 1. In the engineering applications which follow
later in the chapter you will see that 𝑣 may be other integers including those that are negative.
Observe that the equation is linear and homogeneous, but it does not have constant coefcients: the
coefcient of 𝑦 ′′ is 𝑥 2 . In what follows we show that solutions of Bessel’s equation can be found in the
form of infinite power series which we call Bessel functions.
Bessel functions have many applications in physics and engineering. In electrical engineering they are
commonly used in the field of communications. They are important in finding the bandwidth and
frequency content of some radio signals. They are also essential in the analysis of propagating modes
in cylindrical waveguides and cylindrical cavity resonators, both of which are important devices used
in high frequency microwave engineering. Also, Bessel functions are utilized in the design of filters.
Example: (Bessel’s equation of order zero)
By seeking a power series solution in the form ∑∞ 𝑚
𝑚=0 𝑎𝑚 𝑥 , obtain the first four non-zero terms in a
power series solution of Bessel’s equation in the special case when 𝑣 = 0
Solution: when 𝑣 = 0 the Bessel’s equation is:
𝑥 2 𝑦 ′′ + 𝑥 𝑦 ′ + 𝑥 2 𝑦 = 0
which for non-zero 𝑥 the equation becomes

19
 Al-Nahrain University 2nd Year, 1st Semester
College of Information Engineering Engineering Mathematics I
Dept. of Info& Comms Engineering Lecture3: Solving ODE (2nd Order)

𝑥𝑦 ′′ + 𝑦 ′ + 𝑥𝑦 = 0
As was done in the previous section, we seek a power series solution in the form
∞

𝑦 = ∑ 𝑎𝑚 𝑥 𝑚 = 𝑎0 + 𝑎1 𝑥 + 𝑎2 𝑥 2 + 𝑎3 𝑥 3 + 𝑎4 𝑥 4 + 𝑎5 𝑥 5 + ⋯
𝑚=0
The strategy is to substitute 𝑦, and its first and second derivatives into the Bessel’s equation and
choose values 𝑎𝑚 so that the equation is satisfied. 𝑦 and its first two derivatives can be written:
𝑦 = 𝑎0 + 𝑎1 𝑥 + 𝑎2 𝑥 2 + 𝑎3 𝑥 3 + 𝑎4 𝑥 4 + 𝑎5 𝑥 5 + ⋯
𝑦′ = 𝑎1 + 2𝑎2 𝑥 + 3𝑎3 𝑥 2 + 4𝑎4 𝑥 3 + 5𝑎5 𝑥 4 + ⋯
𝑦 ′′ = 2𝑎2 + 6𝑎3 𝑥 + 12𝑎4 𝑥 2 + 20𝑎5 𝑥 3 + ⋯
Substitution into 𝑥𝑦 ′′ + 𝑦 ′ + 𝑥𝑦 = 0 gives
(2𝑎2 𝑥 + 6𝑎3 𝑥 2 + 12𝑎4 𝑥 3 + 20𝑎5 𝑥 4 … ) +(𝑎1 + 2𝑎2 𝑥 + 3𝑎3 𝑥 2 + 4𝑎4 𝑥 3 … )
+(𝑎0 𝑥 + 𝑎1 𝑥 2 + 𝑎2 𝑥 3 + 𝑎3 𝑥 4 + 𝑎4 𝑥 5 … ) = 0
We now equate coefficients on both sides of this equation starting with the constant term:
𝑎1 = 0.
Equating coefficients of 𝑥 :
1
4𝑎2 + 𝑎0 = 0 so that 𝑎2 = − 𝑎0 .
4
2
Equating coefficients of 𝑥 :
1
9𝑎3 + 𝑎1 = 0 so that 𝑎3 = − 𝑎1 = 0 since 𝑎1 = 0.
9
3
Equating coefficients of 𝑥 :
1 1 1 1
16𝑎4 + 𝑎2 = 0 so that 𝑎4 = − 𝑎2 = (− ) (− ) 𝑎0 = 𝑎 .
16 16 4 64 0
It is easy to verify that all subsequent terms 𝑎𝑚 , when 𝑚 is odd, are zero. You should verify, in a
1
similar way, that 𝑎6 = − 𝑎0 . When 𝑚 is even, all 𝑎𝑚 are multiples of 𝑎0 . We can now write down
2304
a power series that satisfies Bessel's differential Equation 19.25:
1 1 1
𝑦 = 𝑎0 (1 − 𝑥 2 + 𝑥 4 − 𝑥6 + ⋯ )
4 64 2304
We have succeeded in obtaining a power series solution of Bessel's Equation where 𝑎0 is an arbitrary
constant. As we have already noted, 𝑎0 is an arbitrary constant. However, historically its value is
chosen to be 1 in the last equation.
The resulting power series is used to define a function called a Bessel function of the first kind of
order zero that is conventionally denoted 𝐽0 (𝑥). It is possible to show that it can be expressed
concisely using sigma notation as
∞
(−1)𝑚 𝑥 2𝑚
𝐽0 (𝑥) = ∑ .
22𝑚 (𝑚!)2
𝑚=0

20
 Al-Nahrain University 2nd Year, 1st Semester
College of Information Engineering Engineering Mathematics I
Dept. of Info& Comms Engineering Lecture3: Solving ODE (2nd Order)

Note that this method of solving the differential equation

produces a power series solution and is used to introduce a new
function, 𝐽0 (𝑥). It is not practical to try to draw a graph of 𝐽0 (𝑥)
by hand but this can be obtained readily using a technical
computing language such as MATLAB®. The figure on the right-
hand shows a graph obtained in this way.
Recall that the general solution of a second order linear equation
requires two independent solutions. The function 𝐽0 (𝑥) is one
such solution. A second independent solution is called a Bessel
function of the second kind of order zero and is denoted by
𝑌0 (𝑥). Derivation of this function is beyond the scope of this course. The general solution of Bessel's
equation of order zero can then be written
𝑦 = 𝐴𝐽0 (𝑥) + 𝐵𝑌0 (𝑥)
where 𝐴 and 𝐵 are arbitrary constants.
Example: (Bessel’s equation of order one)
By seeking a power series solution in the form ∑∞ 𝑚=0 𝑎𝑚 𝑥
𝑚
obtain the first three non-zero terms in
the solution of Bessel's equation in the case when 𝑣 = 1 :
Solution: when 𝑣 = 1 the Bessel’s equation becomes:
𝑥 2 𝑦 ′′ + 𝑥 𝑦 ′ + (𝑥 2 − 1)𝑦 = 0
As before we seek a power series solution in the form
∞

𝑦 = ∑ 𝑎𝑚 𝑥 𝑚 = 𝑎0 + 𝑎1 𝑥 + 𝑎2 𝑥 2 + 𝑎3 𝑥 3 + 𝑎4 𝑥 4 + 𝑎5 𝑥 5 …
𝑚=0

The strategy is to substitute 𝑦, and its first and second derivatives into the Bessel’s equation and
choose values 𝑎𝑚 so that the equation is satisfied. 𝑦 and its first two derivatives can be written:
𝑦 = 𝑎0 + 𝑎1 𝑥 + 𝑎2 𝑥 2 + 𝑎3 𝑥 3 + 𝑎4 𝑥 4 + 𝑎5 𝑥 5 …
𝑦′ = 𝑎1 + 2𝑎2 𝑥 + 3𝑎3 𝑥 2 + 4𝑎4 𝑥 3 + 5𝑎5 𝑥 4 …
𝑦 ′′ = 2𝑎2 + 6𝑎3 𝑥 + 12𝑎4 𝑥 2 + 20𝑎5 𝑥 3 …
Substitution into 𝑥 2 𝑦 ′′ + 𝑥 𝑦 ′ + (𝑥 2 − 1)𝑦 = 0 gives
(2𝑎2 𝑥 2 + 6𝑎3 𝑥 3 + 12𝑎4 𝑥 4 + 20𝑎5 𝑥 5 … ) + (𝑎1 𝑥 + 2𝑎2 𝑥 2 + 3𝑎3 𝑥 3 + 4𝑎4 𝑥 4 + 5𝑎5 𝑥 5 … ) +
(𝑎0 𝑥 2 + 𝑎1 𝑥 3 + 𝑎2 𝑥 4 + 𝑎3 𝑥 5 + 𝑎4 𝑥 6 + 𝑎5 𝑥 7 … ) − (𝑎0 + 𝑎1 𝑥 + 𝑎2 𝑥 2 + 𝑎3 𝑥 3 + 𝑎4 𝑥 4 + 𝑎5 𝑥 5 … ) = 0
We now equate coefficients on both sides of this equation starting with the constant term:
−𝑎0 = 0 so that 𝑎0 = 0.
Equating coefficients of 𝑥 :

21
 Al-Nahrain University 2nd Year, 1st Semester
College of Information Engineering Engineering Mathematics I
Dept. of Info& Comms Engineering Lecture3: Solving ODE (2nd Order)

𝑎1 − 𝑎1 = 0 so that 𝑎1 is arbitrary.
2
Equating coefficients of 𝑥 :
2𝑎2 + 2𝑎2 + 𝑎0 − 𝑎2 = 0 so that 3𝑎2 = −𝑎0 , from which 𝑎2 = 0.
Equating coefficients of 𝑥 3 :
1
6𝑎3 + 3𝑎3 + 𝑎1 − 𝑎3 = 0 so that 8𝑎3 = −𝑎1 , from which 𝑎3 = − 𝑎1 .
8
It is easy to verify that all subsequent terms 𝑎𝑚 when 𝑚 is even, are zero.
Equating coefficients of 𝑥 5 :
20𝑎5 + 5𝑎5 + 𝑎3 − 𝑎5 = 0 so that 24𝑎5 = −𝑎3
from which
1 1
𝑎5 = − 𝑎3 = 𝑎
24 192 1
Note that all non-zero coefficients are multiples of 𝑎1 . We can now write down the first three non-
zero terms in the power series solution:
1 1 5
𝑦 = 𝑎1 (𝑥 − 𝑥 3 + 𝑥 − ⋯)
8 192
Here 𝑎1 is an arbitrary constant. We have succeeded in obtaining
a power series solution of Bessel's Equation. Although 𝑎1 is an
1
arbitrary constant, historically its value is chosen to equal in
2
the last equation and the resulting power series is used to define
a function called a Bessel function of the first kind of order one
that is conventionally denoted 𝐽1 (𝑥). It is possible to show that it
can be expressed concisely using sigma notation as
∞
(−1)𝑚 𝑥 2𝑚+1
𝐽1 (𝑥) = ∑ 2𝑚+1 .
2 𝑚! (𝑚 + 1)!
𝑚=0

The figure on the right-hand shows a graph of 𝐽1 (𝑥) obtained

using computer software.
A second independent solution is called a Bessel function of the second kind of order one and is
denoted by 𝑌1 (𝑥). Derivation of this function is beyond the scope of this course but, like 𝐽1 (𝑥), values
of 𝑌1 (𝑥) can be obtained from tables and by using computer software. The general solution of
Bessel's equation of order one can then be written
𝑦 = 𝐴𝐽1 (𝑥) + 𝐵𝑌1 (𝑥)
where 𝐴 and 𝐵 are arbitrary constants.
A solution of Bessel’s equation of positive integer order 𝑛 is a Bessel function of the first kind of order
𝑛, denoted 𝐽𝑛 (𝑥), and can be shown to be

22
 Al-Nahrain University 2nd Year, 1st Semester
College of Information Engineering Engineering Mathematics I
Dept. of Info& Comms Engineering Lecture3: Solving ODE (2nd Order)

∞
(−1)𝑚 𝑥 2𝑚+𝑛
𝐽𝑛 (𝑥) = ∑ 2𝑚+𝑛 . 𝑓𝑜𝑟 𝑛 = 0,1,2,3, …
2 𝑚! (𝑚 + 𝑛)!
𝑚=0

Whilst to this point we have only discussed cases where the order is non-negative, in fact Bessel’s
functions are defined for any order 𝑣, −∞ < 𝑣 < ∞. To cope with more general cases, including
those of a non-integer order, the usual practice is to seek solutions in the form ∑∞
𝑚=0 𝑎𝑚 𝑥
𝑚+𝑟
where
𝑟 is an unknown real number that is chosen during the course of the calculation in order to generate a
solution. The solution method that deals with such a case is discussed next.

3.3.3 Extended Power Series Solution

Several 2nd order ODEs of considerable practical importance have coefficients that are not constant,
but are “not too bad,” so that these ODEs can still be solved by series (power series times a fractional
power of x). Indeed, the following theorem permits an extension of the power series method. The
new method is called the Frobenius method. Both methods, that is, the power series method and the
Frobenius method, have gained in significance due to the use of software in actual calculations.
Let 𝑏(𝑥) and 𝑐(𝑥) be any functions that are analytic at 𝑥 = 0. Then, the ODE
𝑥 2 𝑦 ′′ + 𝑥 𝑏(𝑥)𝑦 ′ + 𝑐 (𝑥)y = 0 (Eq. 1)
has at least one solution that can be represented in the form
∞

𝑦 = ∑ 𝑎𝑚 𝑥 𝑚+𝑟 = 𝑥 𝑟 (𝑎0 + 𝑎1 𝑥 + 𝑎2 𝑥 2 + 𝑎3 𝑥 3 + 𝑎4 𝑥 4 + 𝑎5 𝑥 5 + ⋯ ) (Eq. 2)

𝑚=0

where the exponent 𝑟 may be any (real or complex) number and is chosen so that 𝑎0 ≠ 0

23
 Al-Nahrain University 2nd Year, 1st Semester
College of Information Engineering Engineering Mathematics I
Dept. of Info& Comms Engineering Lecture3: Solving ODE (2nd Order)

Example:
Use Frobenius method (or extended power series method) to obtain the first three non-zero terms in
the solution of Bessel’s equation of order 1.
Solution: when 𝑣 = 1 the Bessel’s equation becomes:
𝑥 2 𝑦 ′′ + 𝑥 𝑦 ′ + (𝑥 2 − 1)𝑦 = 0
According to Frobenius method, the solution in the form of Eq.2 above.
The strategy is to substitute 𝑦, and its first and second derivatives into the Bessel’s equation and
choose values 𝑎𝑚 so that the equation is satisfied. 𝑦 and its first two derivatives can be written:
∞

𝑦 = ∑ 𝑎𝑚 𝑥 𝑚+𝑟 = 𝑥 𝑟 (𝑎0 + 𝑎1 𝑥 + 𝑎2 𝑥 2 + 𝑎3 𝑥 3 + 𝑎4 𝑥 4 + 𝑎5 𝑥 5 … )
𝑚=0
∞

𝑦 ′ = ∑ (𝑚 + 𝑟) 𝑎𝑚 𝑥 𝑚+𝑟−1 = 𝑥 𝑟−1 (𝑟𝑎0 + (𝑟 + 1)𝑎1 𝑥 + (𝑟 + 2)𝑎2 𝑥 2 + (𝑟 + 3)𝑎3 𝑥 3 … )

𝑚=0
∞

𝑦 ′′ = ∑ (𝑚 + 𝑟)(𝑚 + 𝑟 − 1) 𝑎𝑚 𝑥 𝑚+𝑟−2
𝑚=0
= 𝑥 𝑟−2 (𝑟(𝑟 − 1)𝑎0 + 𝑟(𝑟 + 1)𝑎1 𝑥 + (𝑟 + 2)(𝑟 + 1)𝑎2 𝑥 2 + (𝑟 + 3)(𝑟 + 2)𝑎3 𝑥 3 … )
Substitution into 𝑥 2 𝑦 ′′ + 𝑥 𝑦 ′ + (𝑥 2 − 1)𝑦 = 0 gives
[𝑥 𝑟 (𝑟(𝑟 − 1)𝑎0 + 𝑟(𝑟 + 1)𝑎1 𝑥 + (𝑟 + 2)(𝑟 + 1)𝑎2 𝑥 2 + (𝑟 + 3)(𝑟 + 2)𝑎3 𝑥 3 … )]
+ [𝑥 𝑟 (𝑟𝑎0 + (𝑟 + 1)𝑎1 𝑥 + (𝑟 + 2)𝑎2 𝑥 2 + (𝑟 + 3)𝑎3 𝑥 3 … )]
+[𝑥 𝑟+2 (𝑎0 + 𝑎1 𝑥 + 𝑎2 𝑥 2 + 𝑎3 𝑥 3 + 𝑎4 𝑥 4 + 𝑎5 𝑥 5 … )] (Eq.3)
𝑟 (𝑎 2 3 4 5
−[𝑥 0 + 𝑎1 𝑥 + 𝑎2 𝑥 + 𝑎3 𝑥 + 𝑎4 𝑥 + 𝑎5 𝑥 … )]
Equating coefficients of 𝑥 𝑟 :
𝑟(𝑟 − 1)𝑎0 + 𝑟𝑎0 − 𝑎0 = 0
𝑟 2 𝑎0 − 𝑟𝑎0 + 𝑟 𝑎0 − 𝑎0 = 0
(𝑟 2 − 1)𝑎0 = 0
𝑟 = ±1
For 𝑟 = 1, (Eq.3) becomes
(2𝑎1 𝑥 2 + 6𝑎2 𝑥 3 + 12𝑎3 𝑥 4 … ) + (𝑎0 𝑥 + 2𝑎1 𝑥 2 + 3𝑎2 𝑥 3 + 4𝑎3 𝑥 4 … )
+(𝑎0 𝑥 3 + 𝑎1 𝑥 4 + 𝑎2 𝑥 5 + 𝑎3 𝑥 6 + 𝑎4 𝑥 7 + 𝑎5 𝑥 8 … ) (Eq.4)
−(𝑎0 𝑥 + 𝑎1 𝑥 2 + 𝑎2 𝑥 3 + 𝑎3 𝑥 4 + 𝑎4 𝑥 5 + 𝑎5 𝑥 6 … )
Equating coefficients of 𝑥 :
𝑎0 − 𝑎0 = 0 so that 𝑎0 is arbitrary.
2
Equating coefficients of 𝑥 :
2𝑎1 + 2𝑎1 − 𝑎1 = 0 so that 𝑎1 = 0.
Equating coefficients of 𝑥 3 :

24
 Al-Nahrain University 2nd Year, 1st Semester
College of Information Engineering Engineering Mathematics I
Dept. of Info& Comms Engineering Lecture3: Solving ODE (2nd Order)

1
6𝑎2 + 3𝑎2 + 𝑎0 − 𝑎2 = 0 so that 𝑎2 = − 𝑎0 .
8
It is easy to verify that all subsequent terms 𝑎𝑚 when 𝑚 is odd, are zero.
Equating coefficients of 𝑥 5 : (to be completed in class)
?= 0
from which
𝑎4 = ?
Note that all non-zero coefficients are multiples of 𝑎0 . We can now write down the first three non-
zero terms in the power series solution:
𝑦 = 𝑥 𝑟 (𝑎0 + 𝑎1 𝑥 + 𝑎2 𝑥 2 + 𝑎3 𝑥 3 + 𝑎4 𝑥 4 + 𝑎5 𝑥 5 + ⋯ )
𝑦 = 𝑎0 𝑥 + 𝑎2 𝑥 3 + 𝑎4 𝑥 5 …
𝑦 =?
is the solution of Bessel’s equation of order 1 using the Frobenius method equal to the solution using
the power series method (solved in page no.22) ?

25