Differential Equation Lecture Notes Section 1 Onwards

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 46

Lecture Notes on Differential Equations (MAT-CC-513/MATF-CC-513) Dr Dilip Kumar

Lecture Notes on Differential Equations: Module 1


(MAT-CC-513/MATF-CC-513)
Dr Dilip Kumar
Assistant Professor of Mathematics,
Department of Mathematics, University of Kerala
Kariavattom Campus, Thiruvananthapuram ,Kerala- 695 581, India
Email: dilipkumar.cms@gmail.com, Mobile: 9446195433

1 Introduction
Definition 1. A general linear differential equation of order n with constant coefficients
is of the form
ao y (n) + a1 y (n−1) + a2 y (n−2) + · · · + an y = b(x) (1)
where a0 , a1 , a2 , · · · , an , are complex constants and b(x) is a certain continuous function
defined on some interval I.
Assume a0 = 1, then equation (1) takes the form

y (n) + a1 y (n−1) + a2 y (n−2) + · · · + an y = b(x). (2)

If L.H.S. of equation (2) is L(y) then (2) can be written as L(y) = b(x). Thus we have

L(y) = y (n) + a1 y (n−1) + a2 y (n−2) + · · · + an y.

Where L is the called a differential operator.


Note: If the function b(x) = 0, for all x in I, the corresponding equation L(y) = 0
is called a homogeneous equation, whereas if b(x) 6= 0 for some x in I, L(y) = b(x) is
called a nonhomogeneous equation.

2 The Second Order Homogeneous Differential Equations


Consider the solution of the first order differential equation y 0 + ay = 0 where a is
constant. It has a solution of the form e−ax . Also note that the constant −a is a
solution of the equation r + a = 0. Now consider a second order linear differential
equation of the form
L(y) = y 00 + a1 y 0 + a2 y = 0 (3)
where a1 and a2 are constants. Now since the derivative of erx is constant times erx , we
try y = erx as a solution of (3). Then y 0 = rerx ,y 00 = r2 erx . So

L(erx ) = r2 erx + a1 rerx + a2 erx


= (r2 + a1 r + a2 )erx

We see that erx is a solution of L(y) = 0, if r satisfies r2 + a1 r + a2 = 0. Here


p(r) = r2 + a1 r + a2 is called characteristic polynomial.

1
Lecture Notes on Differential Equations (MAT-CC-513/MATF-CC-513) Dr Dilip Kumar

Theorem 1. Let a1 and a2 be constants and consider the differential equation


L(y) = y 00 + a1 y 0 + a2 y = 0. (3)
If r1 and r2 are two distinct roots of characteristic polynomial p, where p(r) = r2 +
a1 r + a2 , then the functions, φ1 , φ2 defined by φ1 (x) = er1 x and φ2 (x) = er2 x forms the
solutions of L(y) = 0. If r1 is a repeated root of the characteristic polynomial p, then
the functions φ1 (x) = er1 x and φ2 (x) = xer1 x form the solutions to L(y) = 0.
Proof. The characteristic polynomial corresponding to the differential equation (3) is
given by p(r) = r2 +a1 r+a2 . Suppose r1 and r2 be two distinct roots of the characteristic
polynomial p, then consider
L(er1 x ) = r12 er1 x + a1 r1 er1 x + a2 er1 x
= (r12 + a1 r1 + a2 )er1 x = p(r1 )er1 x = 0, since r1 is a root, p(r1 ) = 0
Similarly
L(er2 x ) = r22 er2 x + a1 r2 er2 x + a2 er2 x
= (r22 + a1 r2 + a2 )er2 x = p(r2 )er2 x = 0, since r2 is a root, p(r2 ) = 0
Thus we see that both er1 x and er2 x satisfies the equation (3). Hence φ1 (x) = er1 x and
φ2 (x) = er2 x form solutions to the equation (3).

Now suppose r1 is a repeated root of the characteristic polynomial p, then cer-


tainly φ1 (x) = er1 x forms one of the solution to L(y) = 0. To find the other solution,
differentiate L(erx ) = r2 erx + a1 rerx + a2 erx with respect to r, then, we obtain
∂ ∂
L(erx ) = (p(r)erx )
∂r ∂r
= p(r)xerx + p0 (r)erx = (xp(r) + p0 (r))erx . (a)
Also if y = xerx , y 0 = xrerx + erx and y 00 = xr2 erx + 2rerx , then
L(erx x) = xr2 erx + 2rerx + a1 (xrerx + erx ) + a2 xerx
= x(r2 + a1 r + a2 )erx + (2r + a1 )erx
L(erx x) = p(r)xerx + p0 (r)erx = (xp(r) + p0 (r))erx . (b)
Therefore, from (a) and (b), we have

L(erx ) = (xp(r) + p0 (r))erx .
L(erx x) =
∂r
Now replace r by r1 , then, one can get
L(er1 x x) = (xp(r1 ) + p0 (r1 ))er1 x = 0,
since r1 is a repeated root of the polynomial p, we know that p0 (r1 ) = 0.
Therefore, we have
L(er1 x x) = 0.
Hence, φ2 (x) = xer1 x forms another solution to L(y) = 0.

2
Lecture Notes on Differential Equations (MAT-CC-513/MATF-CC-513) Dr Dilip Kumar

Example 1. Find the solution of y 00 + y 0 − 2y = 0.


Proof. The characteristic polynomial is given by

r2 + r − 2 = 0

Hence r1 = 1, r2 = −2 are the roots. Therefore, the solution is given by φ(x) =


c1 ex + c2 e−2x .
Example 2. Find the solution of y 00 − 4y = 0.
Proof. The characteristic polynomial is given by

r2 − 4 = 0

Hence r1 = 2, r2 = −2 are the roots. Therefore, the solution is given by φ(x) =


c1 e2x + c2 e−2x .
Example 3. Find the solution of y 00 + y 0 − 6y = 0 satisfying φ(0) = 1, φ0 (0) = 0.
Proof. The characteristic polynomial is given by

r2 + r − 6 = 0

Hence r1 = −3, r2 = 2 are the roots. Therefore, the solution is given by φ(x) =
c1 e−3x + c2 e2x .
From initial conditions, we have

φ(0) = c1 + c2 = 1 (4)
φ0 (0) = −3c1 + 2c2 = 0 (5)

By solving (4) and (5), we get


2 3
c1 = , c 2 =
5 5
1 −3x
Therefore, the solution is given by φ(x) = 5 (2e + 3e2x ).
Example 4. Consider the differential equation L(y) = y 00 +a1 y 0 +a2 y = 0, where a1 , a2
are real. Let α + iβ be the characteristic root of the polynomial p(r) then
a) Show that α − iβ is also a root.
b) Show that the solution is in the form φ(x) = eαx (d1 cos βx + d2 sin βx).
a2
c) Show that α = −a2 1 , β 2 = a2 − 41 .
Proof. Consider the equation L(y) = y 00 +a1 y 0 +a2 y = 0. The characteristic polynomial
corresponding to this differential equation is given by r2 + a1 r + a2 = 0.
Given z1 = α + iβ be a characteristic root of the polynomial. Hence,

z12 + a1 z1 + a2 = 0

3
Lecture Notes on Differential Equations (MAT-CC-513/MATF-CC-513) Dr Dilip Kumar

By taking bar on both sides we have

z12 + a1 z1 + a2 = 0
z12 + a1 z1 + a2 = 0
z1 2 + a1 z1 + a2 = 0 (since a1 and a2 are real).

Hence, z1 is also a root.


i.e, z1 = α − iβ is also the root of the polynomial p(r).
b) We know that α + iβ and α − iβ are two distinct roots of the characteristic
polynomial. Therefore, the solution is given by

φ(x) = c1 e(α+iβ)x + c2 e(α−iβ)x


= eαx (c1 eiβx + c2 e−iβx )
= eαx (c1 (cos βx + i sin βx) + c2 (cos βx − i sin βx))
= eαx ((c1 + c2 ) cos βx + i(c1 − c2 ) sin βx)

Put d1 = c1 + c2 , d2 = i(c1 − c2 ), hence,

φ(x) = eαx (d1 cos βx + d2 sin βx)

c) Given α + iβ and α − iβ are two distinct roots of quadratic polynomial P (r) =


r2 + a1 r + a2 . We know that the sum of the roots is given by

α + iβ + α − iβ = −a1
−a1
That is α = 2
. The product of two roots is given by

(α + iβ)(α − iβ) = a2
α 2 + β 2 = a2
 2
2 −a1
β = a2 −
2
a21
so that β 2 = a2 − 4
.

3 Initial Value Problem


Definition 2. An initial value problem is a problem of finding the solution φ to

L(y) = y 00 + a1 y 0 + a2 y = 0 (3)
satisfying φ(x0 ) = α, φ0 (x0 ) = β where α and β be any two given constants.
Theorem 2 (Existence Theorem). For any real x0 , and constants α, β, there exists a
solution φ of the initial value problem L(y) = y 00 +a1 y 0 +a2 y = 0, φ(x0 ) = α, φ0 (x0 ) = β,
on −∞ < x < ∞.

4
Lecture Notes on Differential Equations (MAT-CC-513/MATF-CC-513) Dr Dilip Kumar

Proof. To prove this theorem, we have to show that there exists unique constants c1
and c2 such that the function φ = c1 φ1 + c2 φ2 satisfies (3), where φ1 and φ2 are the
solutions. Since φ satisfies the initial conditions, we must have

c1 φ1 (x0 ) + c2 φ2 (x0 ) = α
c1 φ01 (x0 ) + c2 φ02 (x0 ) = β

To show c1 and c2 are unique, we have to prove

φ1 (x0 ) φ2 (x0 )
∆=
φ01 (x0 ) φ02 (x0 )
= (φ1 φ02 − φ01 φ2 )(x0 ) 6= 0

If r1 and r2 are two distinct roots of the characteristic polynomial, then from the
previous section, we know that φ1 (x) = er1 x and φ2 (x) = er2 x .
Hence,

∆ = er1 x0 r2 er2 x0 − r1 er1 x0 er2 x0


= (r2 − r1 )e(r1 +r2 )x0 6= 0

Now, if r1 is a repeated root of the characteristic polynomial then again from the
previous section, we know that
φ1 (x) = er1 x and φ2 (x) = xer1 x .

Therefore, the determinant is

∆ = er1 x0 (x0 r1 er1 x0 + er1 x0 ) − r1 er1 x0 x0 er1 x0


= e2r1 x0 6= 0

so that the constants c1 and c2 uniquely exists and hence, the function φ(x) = c1 φ1 (x)+
c2 φ2 (x) forms the solution for the given differential equation.

Theorem 3. Let φ be any solution of (3) on an interval I containing a point x0 . Then,


for all x in I, we have

e−k|x−x0 | kφ(x0 )k ≤ kφ(x)k ≤ kφ(x0 )kek|x−x0 |


1
where kφ(x)k = [|φ(x)|2 + |φ0 (x)|2 ] 2 and k = 1 + |a1 | + |a2 |.

Proof. Let u(x) = kφ(x)k2 . Then

kφk2 = |φ|2 + |φ0 |2


= φφ + φ0 φ0 .

5
Lecture Notes on Differential Equations (MAT-CC-513/MATF-CC-513) Dr Dilip Kumar

The derivative becomes


u0 = φφ0 + φ0 φ + φ0 φ00 + φ00 φ0
Taking modulus on both sides, we have

|u0 | ≤ |φ||φ0 | + |φ0 ||φ| + |φ0 ||φ00 | + |φ00 ||φ0 |.

That is
|u0 | ≤ 2|φ||φ0 | + 2|φ0 ||φ00 |. (6)
Since φ is a solution of L(y) = 0,we have

φ00 + a1 φ0 + a2 φ = 0.

Hence, φ00 = −a1 φ0 − a2 φ and

|φ00 | ≤ |a1 ||φ0 | + |a2 ||φ|.

By substituting the expression for |φ00 | in (6), we get

|u0 | ≤ 2|φ||φ0 | + 2|φ0 |(|a1 ||φ0 | + |a2 ||φ|)


= 2|φ||φ0 | + 2|a1 ||φ0 |2 + 2|a2 ||φ||φ0 |.

Therefore
|u0 | ≤ 2(1 + |a2 |)|φ||φ0 | + 2|a1 ||φ0 |2 . (7)
But we know that
0 ≤ (|a| − |b|)2 = |a|2 + |b|2 − 2|a||b|
Hence,
2|a||b| ≤ |a|2 + |b|2 .
By using this inequality in (7), we get

|u0 | ≤ (1 + |a2 |)(|φ|2 + |φ0 |2 ) + 2|a1 ||φ0 |2


= (1 + 2|a1 | + |a2 |)|φ0 |2 + (1 + |a2 |)|φ|2
|u0 | ≤ 2(1 + |a1 | + |a2 |)(|φ|2 + |φ0 |2 ).

Hence
|u0 | ≤ 2ku where k = 1 + |a1 | + |a2 |
so that
−2ku ≤ u0 ≤ 2ku. (8)
By considering the right inequality in (8), we get

u0 − 2ku ≤ 0.

6
Lecture Notes on Differential Equations (MAT-CC-513/MATF-CC-513) Dr Dilip Kumar

Multiplying throughout the above equation by e−2kx , we get

e−2kx (u0 − 2ku) ≤ 0.

It can be written as
[ue−2kx ]0 ≤ 0
Integrate the above term from x0 to x, where x0 < x, to get

u(x)e−2kx − u(x0 )e−2kx0 ≤ 0

or
u(x)e−2kx ≤ u(x0 )e−2kx0 .
That is
u(x) ≤ u(x0 )e2k(x−x0 ) ∀x > x0 .
Using the definition of norm, we get

kφ(x)k2 ≤ kφ(x0 )k2 e2k|x−x0 | , ∀x > x0 .

kφ(x)k ≤ kφ(x0 )kek|x−x0 | , ∀x > x0 . (9)


Again by considering the left inequality of equation (8), we get

kφ(x0 )ke−k|x−x0 | ≤ kφ(x)k, ∀x > x0 . (10)

By clubbing (9) and (10), we get the required inequality as

kφ(x0 )ke−k|x−x0 | ≤ kφ(x)k ≤ kφ(x0 )kek|x−x0 | , ∀x > x0

If x < x0 , then the inequality becomes

kφ(x0 )kek|x−x0 | ≤ kφ(x)k ≤ kφ(x0 )ke−k|x−x0 | , ∀x < x0

Theorem 4 (Uniqueness Theorem). Let α and β be any constants and let x0 be any
real number. On any interval I containing a point x0 , there exist at most one solution
φ to an initial value problem L(y) = y 00 + a1 y 0 + a2 y = 0, φ(x0 ) = α, φ0 (x0 ) = β.
Proof. Suppose φ and ϕ were two solutions. Let χ = φ − ϕ, then

L(χ) = L(φ − ϕ) = (φ − ϕ)00 + a1 (φ − ϕ)0 + a2 (φ − ϕ) = L(φ) − L(ϕ) = 0


χ(x0 ) = φ(x0 ) − ϕ(x0 ) = α − α = 0
χ0 (x0 ) = φ0 (x0 ) − ϕ0 (x0 ) = β − β = 0.

From theorem 3, we know athat

e−k|x−x0 | kχ(x0 )k ≤ kχ(x)k ≤ kχ(x0 )kek|x−x0 |

7
Lecture Notes on Differential Equations (MAT-CC-513/MATF-CC-513) Dr Dilip Kumar

1
where kχ(x)k = [|χ(x)|2 + |χ0 (x)|2 ] 2 and k = 1 + |a1 | + |a2 |.
By definition of norm, we have kχ(x0 )k = 0. Therefore, kχ(x)k = 0
That is χ(x) = 0
or φ(x) − ϕ(x) = 0 which implies φ(x) = ϕ(x). That is φ is the unique solution of the
given initial value problem.
Example 5. Solve an initial value problem y 00 − 2y 0 − 3y = 0 satisfying φ(0) = 0,
φ0 (0) = 1.
Proof. The characteristic polynomial is given by

r2 − 2r − 3 = 0

Hence r1 = 3, r2 = −1 are the roots of the characteristic polynomial. Therefore, the


solution is given by φ(x) = c1 e3x + c2 e−x .
φ0 (x) = 3c1 e3x − c2 e−x .
Given that φ(0) = 0 and φ0 (0) = 1 From initial conditions, we have

φ(0) = c1 + c2 = 0 (11)
φ0 (0) = 3c1 − c2 = 1 (12)

By solving (11) and (12), we get


1 1
c1 = , c 2 = −
4 4
Therefore, the solution is given by φ(x) = 41 (e3x − e−x ).
Example 6. Find a function φ which has a continuous derivative on 0 ≤ x ≤ 2 and
satisfies φ(0) = 0, φ0 (0) = 1 and y 00 − y = 0, 0 ≤ x ≤ 1
and y 00 − 9y = 0, 1 ≤ x ≤ 2.
Proof. Consider the equation y 00 − y = 0, 0 ≤ x ≤ 1.
The characteristic polynomial is given by

r2 − 1 = 0.

Hence r1 = 1, r2 = −1 are the roots of the characteristic polynomial. Therefore the


solution is given by

φ(x) = c1 ex + c2 e−x
φ0 (x) = c1 ex − c2 e−x

Given that φ(0) = 0 and φ0 (0) = 1 From initial conditions, we have

c1 + c2 = 0 (13)
c1 − c2 = 1 (14)

8
Lecture Notes on Differential Equations (MAT-CC-513/MATF-CC-513) Dr Dilip Kumar

By solving the above equations, we get


1 1
c1 = , c 2 = −
2 2
Therefore, the solution is given by φ(x) = 21 (ex − e−x ) in the interval 0 ≤ x ≤ 1.
Now consider y 00 − 9y = 0, 1 ≤ x ≤ 2.
The characteristic polynomial is given by

r2 − 9 = 0.

Hence r1 = 3, r2 = −3 are the roots of the characteristic polynomial. Therefore the


solution is given by

ϕ(x) = d1 e3x + d2 e−3x


ϕ0 (x) = 3d1 e3x − 3d2 e−3x .

Since ‘10 is the common point in the intervals 0 ≤ x ≤ 2 and 1 ≤ x ≤ 2, we equate φ(1)
and ϕ(1); φ0 (1), ϕ0 (1) respectively, and get
1
(e − e−1 ) = d1 e3 + d2 e−3
2
1
(e + e−1 ) = 3d1 e3 − 3d2 e−3 .
2
Now by solving the above two equations, we get
2e − e−1 e − 2e−1
d1 = , d 2 =
6e3 6e−3
Therefore, the solution ϕ becomes
2e − e−1 3x e − 2e−1 −3x
ϕ(x) = e + e .
6e3 6e−3

in the interval 1 ≤ x ≤ 2.

4 Linear Dependence and Independence


Linear Dependence

Two functions φ1 and φ2 defined on an interval I are said to be linearly dependent if


there exist two constants c1 and c2 not both zero such that c1 φ1 +c2 φ2 = 0 for all x in I.

Linear Independence

Two functions φ1 and φ2 defined on an interval I are said to be linearly independent


if there exist two constants c1 and c2 both zero such that c1 φ1 + c2 φ2 = 0 for all x in I.

9
Lecture Notes on Differential Equations (MAT-CC-513/MATF-CC-513) Dr Dilip Kumar

φ1 φ2
Definition 3. The = φ1 φ02 − φ01 φ2 is called Wronskian of φ1 and φ2 and it is
φ01 φ02
denoted as W (φ1 , φ2 ). It is a function and its value at x is denoted by W (φ1 , φ2 )(x).
Theorem 5. Two solutions φ1 and φ2 of L(y) = 0 are linearly independent on an
interval I, if and only if W (φ1 , φ2 )(x) 6= 0, ∀ x ∈ I.
Proof. Suppose W (φ1 , φ2 )(x) 6= 0, ∀ x ∈ I. Now, we have to prove that the solutions
φ1 and φ2 are linearly independent. Let c1 and c2 be two constants such that

c1 φ1 + c2 φ2 = 0, ∀ x ∈ I. (15)

Then, also
c1 φ01 (x) + c2 φ02 (x) = 0, ∀ x ∈ I. (16)
For a fixed x, equations (15) and (16) are linear homogeneous equations satisfied by
c1 and c2 . The determinant of the coefficients is just W (φ1 , φ2 )(x) which is not zero.
Hence, by Cramer’s rule, we see that c1 = 0, c2 = 0. So we conclude that solutions φ1
and φ2 are linearly independent.

Conversely, suppose that φ1 and φ2 are linearly independent solutions of L(y) = 0.


Now it is required to show that W (φ1 , φ2 )(x) 6= 0. Suppose that there is an x0 in I
such that W (φ1 , φ2 )(x0 ) = 0. This gives system of two equations

c1 φ1 (x0 ) + c2 φ2 (x0 ) = 0
. (17)
c1 φ01 (x0 ) + c2 φ02 (x0 ) = 0

These equations have solution c1 and c2 , where atleast one of these numbers is not zero.
Let c1 and c2 be such solution and consider the function

ϕ = c1 φ1 + c2 φ2 . (18)

Now, L(ϕ) = 0 and from equation (17), we see that ϕ(x0 ) = 0, ϕ0 (x0 ) = 0. Hence,
kϕ(x0 )k = 0 We know the inequality

e−k|x−x0 | kϕ(x0 )k ≤ kϕ(x)k ≤ kϕ(x0 )kek|x−x0 |


1
where kϕ(x)k = [|ϕ(x)|2 + |ϕ0 (x)|2 ] 2 and k = 1 + |a1 | + |a2 |.
From this, we get ϕ(x) = 0, ∀x ∈ I and thus, from equation (18), one can get

c1 φ1 (x) + c2 φ2 (x) = 0, ∀x ∈ I.

But this is a contradiction to the fact that φ1 and φ2 are linearly independent on I.
Thus, the supposition that there was a point x0 in I such that W (φ1 , φ2 )(x0 ) = 0 is
wrong Hence, W (φ1 , φ2 )(x) 6= 0.

Therefore, we have consequently proved that W (φ1 , φ2 )(x) 6= 0, ∀x ∈ I.

10
Lecture Notes on Differential Equations (MAT-CC-513/MATF-CC-513) Dr Dilip Kumar

Theorem 6. Let φ1 and φ2 be two solutions of L(y) = 0 defined on an initial I,


containing a point x0 .Then φ1 and φ2 are linearly independent on I if and only if
W (φ1 , φ2 )(x0 ) 6= 0.
Proof. Suppose the solutions φ1 and φ2 are linearly independent on I then, by theorem
5, we get W (φ1 , φ2 )(x) 6= 0, ∀x ∈ I.
Hence, in particular, W (φ1 , φ2 )(x0 ) 6= 0.
Conversely, suppose W (φ1 , φ2 )(x0 ) 6= 0, and c1 and c2 are constants such that

c1 φ1 (x) + c2 φ2 (x) = 0, ∀x ∈ I.

then, we see that

c1 φ1 (x0 ) + c2 φ2 (x0 ) = 0
c1 φ01 (x0 ) + c2 φ02 (x0 ) = 0

and the determinant of coefficients is W (φ1 , φ2 )(x0 ) 6= 0. from Cramer’s rule, we obtain
c1 = 0 and c2 = 0. Hence, the solutions φ1 and φ2 are linearly independent on I.
Example 7. Show that the following functions are linearly independent.

φ1 = x, φ2 = |x|

Proof. Choose any two constants c1 and c2 such that c1 x + c2 x = 0 for all x in I. By
definition of |x|, we have

c1 x + c2 x = 0 if x > 0
and c1 x − c2 x = 0 if x < 0

Solving these two equations, we get c1 = 0 and c2 = 0. Therefore, φ1 and φ2 are linearly
independent.
4.1 A Formula for Wronskian
Theorem 7. If φ1 and φ2 are any two solutions of L(y) = 0 on an interval I containing
a point x0 , then
W (φ1 , φ2 )(x) = e−a1 (x−x0 ) W (φ1 , φ2 )(x0 ).
Proof. Given φ1 and φ2 are solutions of L(y) = 0, therefore,

φ001 + a1 φ01 + a2 φ1 = 0 (a)


φ002 + a1 φ02 + a2 φ2 = 0 (b)

Solving these the above two equations, we have

(a) × φ2 ⇒ φ2 φ001 + a1 φ2 φ01 + a2 φ2 φ1 = 0 (c)


(b) × φ1 ⇒ φ1 φ002 + a1 φ1 φ02 + a2 φ1 φ2 = 0 (d)

11
Lecture Notes on Differential Equations (MAT-CC-513/MATF-CC-513) Dr Dilip Kumar

(d) − (c) ⇒ (φ1 φ002 − φ2 φ001 ) + a1 (φ1 φ02 − φ2 φ01 )


But we have W (φ1 , φ2 )(x) = (φ1 φ02 − φ2 φ01 )(x) and W 0 (φ1 , φ2 )(x) = (φ1 φ002 − φ2 φ001 )(x).
Thus we get
W 0 (φ1 , φ2 )(x) + a1 W (φ1 , φ2 )(x) = 0.
This is in the form of first order linear homogeneous differential equation and hence,
the solution of this equation is given by

W (φ1 , φ2 )(x) = ce−a1 x .

Putting x = x0 in the above, we get

W (φ1 , φ2 )(x0 ) = ce−a1 x0 .

Hence
c = ea1 x0 W (φ1 , φ2 )(x0 ).
By using this expression of c, we have

W (φ1 , φ2 )(x) = ea1 x0 W (φ1 , φ2 )(x0 )e−a1 x .

or
W (φ1 , φ2 )(x) = e−a1 (x−x0 ) W (φ1 , φ2 )(x0 )
which is the required solution.

Problem 1. Determine whether the following are linearly dependent or independent?

1. φ1 = x and φ2 = erx where r is a complex constant.

2. φ1 = cos x and φ2 = sin x

1. Proof. φ1 = x and φ2 = erx then φ01 = 1 and φ02 = rerx


W (φ1 , φ2 ) = xrerx − erx = (xr − 1)erx 6= 0.
Therefore φ1 = x and φ2 = erx are linearly independent by Theorem.

2. Proof. φ1 = cos x and φ2 = sin x φ01 = − sin x and φ02 = cos x


W (φ1 , φ2 ) = cos2 x + sin2 x = 1 6= 0.
Therefore φ1 = cos x and φ2 = sin x are linearly independent by Theorem.

5 Second Order Nonhomogeneous Differential Equation


Next we discuss how to find all the solutions of the differential equation of the form

L(y) = y 00 + a1 y 0 + a2 y = b(x)

where a1 and a2 are constants and b(x) is a continuous function defined on an interval I.

12
Lecture Notes on Differential Equations (MAT-CC-513/MATF-CC-513) Dr Dilip Kumar

Let ϕp be the particular solution of L(y) = b(x) and φ be any other solution of
L(y) = b(x). Then, L(φ) = b(x) and L(ϕp ) = b(x).
Therefore
L(φ − ϕp ) = L(φ) − L(ϕp ) = b(x) − b(x) = 0
on I. This shows that (φ − ϕp ) is a solution of homogeneous differential equation
L(y) = 0. So the solution (φ − ϕp ) must be in the form c1 φ1 + c2 φ2 = 0, where φ1 and
φ2 are linearly independent solutions of L(y) = 0. In other words, every solution φ of
L(y) = b(x) can be written in form
φ = ϕp + c1 φ1 + c2 φ2 .
Thus to find all the solution of L(y) = b(x) it is enough to find a particular solution ϕp
and two linearly independent solutions φ1 and φ2 .
5.1 Method of Undetermined Coefficients
Method of Undetermined Coefficients is mainly used to find the solution of the equation
of the form
L(y) = y 00 + a1 y 0 + a2 y = b(x)
where b(x) contains a polynomial or trigonometric functions such as sin px, cos px,
tan px or exponential function epx . Here p is a constant or combinations of sums and
products of these.
Example 8. Solve the nonhomogeneous equation
4y 00 − y = ex (i)
Proof. Here the corresponding homogenous equation is 4y 00 − y = 0. The characteristic
polynomial is given by
4r2 − 1 = 0.
Hence r1 = − 21 and r2 = 1
2
are the solutions. Therefore, the solution of the homogeneous
equation is given by
1 1
ψ(x) = c1 e− 2 x + c2 e 2 x .
To find the particular solution, consider a function y = aex as a solution, where a is an
undetermined coefficient.
Substituting this equation in (i) we get,
4aex − aex = ex
3aex = ex
1
This implies a =
3
Thus the particular solution is ψp = 13 ex . Hence the general solution is given by
1 1 1
φ(x) = c1 e− 2 x + c2 e 2 x + ex .
3

13
Lecture Notes on Differential Equations (MAT-CC-513/MATF-CC-513) Dr Dilip Kumar

Example 9. Find the solution of a nonhomogeneous equation

y 00 + 4y 0 + 4y = 6 sin 3x. (ii)

Proof. The corresponding homogeneous equation is given by y 00 + 4y 0 + 4y = 0.


The characteristic polynomial is given by r2 + 4r + 4 = 0.
Hence r1 = −2 and r2 = −2 are the roots.
Therefore, the solution of the homogeneous equation is given by

ψ(x) = (c1 + c2 x)e−2x

which is the complementary solution.


To find the particular solution, consider the function y = a sin 3x+b cos 3x as a solution
where a and b are the undetermined coefficients.

y = a sin 3x + b cos 3x
y 0 = 3a cos 3x − 3b sin 3x
y 00 = −9a sin 3x − 9b cos 3x.
Substituting these in equation (ii) we get,

(−9a sin 3x − 9b cos 3x) + 4(3a cos 3x − 3b sin 3x) + 4(a sin 3x + b cos 3x) = 6 sin 3x
(−5a − 12b) sin 3x + (12a − 5b) cos 3x = 6 sin 3x

Equating the coefficients of sin 3x and cos 3x on both sides, we get −5a − 12b = 6 and
5
12a − 5b = 0. Solving these two equations we have a = 12 b.
−25b − 144b
=6
12
−25b − 144b = 72.
72 30
Hence b = − 169 and a = − 169 .
Therefore the particular solution is
30 72
ψp = − sin 3x − cos 3x.
169 169
The general solution is given by
30 72
φ = (c1 + c2 x)e−2x − sin 3x − cos 3x.
169 169

Example 10. Solve the nonhomogeneous equation

y 00 + 4y = cos x. (iii)

14
Lecture Notes on Differential Equations (MAT-CC-513/MATF-CC-513) Dr Dilip Kumar

Proof. The corresponding homogeneous equation is given by y 00 + 4y = 0.


The characteristic polynomial is given by r2 + 4 = 0.
Hence r1 = 2i and r2 = −2i are the roots.
Therefore, the solution of the homogeneous equation is given by

ψ(x) = c1 cos 2x + c2 sin 2x

which is the complementary solution.


To find the particular solution, consider the function y = a sin x + b cos x as a solution
where a and b are the undetermined coefficients.

y = a sin x + b cos x
y 0 = a cos x − b sin x
y 00 = −a sin x − b cos x.
Substituting these in equation (iii) we get,

(−a sin x − b cos x) + 4(a sin x + b cos x) = cos x


(3a) sin x + 3b cos x = cos x

Equating the coefficients of sin x and cos x on both sides, we get 3a = 0 and 3b = 1.
Solving these two equations we have a = 0 and b = 31
Therefore the particular solution is
1
ψp = cos x.
3
The general solution is given by
1
φ = c1 cos 2x + c2 sin 2x + cos x.
3

Example 11. Find the solution of a nonhomogeneous equation

y 00 + 9y = sin 3x. (iv)

Proof. The corresponding homogeneous equation is given by y 00 + 9y = 0.


The characteristic polynomial is given by r2 + 9 = 0.
Hence r1 = 3i and r2 = −3i are the roots.
Therefore, the solution of the homogeneous equation is given by

ψ(x) = c1 cos 3x + c2 sin 3x

which is the complementary solution.


To find the particular solution, consider the function y = a sin 3x+b cos 3x as a solution

15
Lecture Notes on Differential Equations (MAT-CC-513/MATF-CC-513) Dr Dilip Kumar

where a and b are the undetermined coefficients.

y = a sin 3x + b cos 3x
y 0 = 3a cos 3x − 3b sin 3x
y 00 = −9a sin 3x − 9b cos 3x.
Substituting these in equation (iv) we get,

(−9a sin 3x − 9b cos 3x) + 9(a sin 3x + b cos 3x) = sin 3x

which is impossible. Hence consider another function y = x(a sin 3x + b cos 3x) as a
solution where a and b are the undetermined coefficients.

y = x(a sin 3x + b cos 3x)

y 0 = x(3a cos 3x − 3b sin 3x) + (a sin 3x + b cos 3x)


y 00 = x(−9a sin 3x − 9b cos 3x) + (3a cos 3x − 3b sin 3x) + (3a cos 3x − 3b sin 3x).
Substituting these in equation (iv) we get,

x(−9a sin 3x − 9b cos 3x) + 2(3a cos 3x − 3b sin 3x) + 9x(a sin 3x + b cos 3x) = sin 3x

6a cos 3x − 6b sin 3x = sin 3x


Equating the coefficients of sin 3x, cos 3x, on both sides, we get b = − 61 and a = 0.
Therefore the particular solution is
1
ψp = − x cos 3x
6
The general solution is given by
1
φ = c1 cos 3x + c2 sin 3x − x cos 3x
6

Example 12. Find the solution of a nonhomogeneous equation

y 00 + 4y 0 + 9y = x2 + 3x. (v)

Proof. The corresponding homogeneous equation is given by y 00 + 4y 0 + 9y = 0.


The characteristic√polynomial is given√by r2 + 4r + 9 = 0.
Hence r1 = −2 + 5i and r2 = −2 − 5i are the roots.
Therefore, the solution of the homogeneous equation is given by
√ √
ψ(x) = e−2x (c1 cos 5x + c2 sin 5x)

16
Lecture Notes on Differential Equations (MAT-CC-513/MATF-CC-513) Dr Dilip Kumar

which is the complementary solution.


To find the particular solution, consider the function y = ax2 + bx + c as a solution
where a, b and c are the undetermined coefficients.

y = ax2 + bx + c
y 0 = 2ax + b
y 00 = 2a
Substituting these in equation (v) we get,

(2a) + 4(2ax + b) + 9(ax2 + bx + c) = x2 + 3x

Equating the coefficients of x2 , x and constant terms, on both sides, we get

9a = 1, 8a + 9b = 3 and 2a + 4b + 9c = 0.

On solving we get a = 91 , b = 19
81
94
and c = − 729 .
Therefore the particular solution is
1 19 94
ψp = x2 + x −
9 81 729
The general solution is given by
√ √ 1 19 94
φ = e−2x (c1 cos 5x + c2 sin 5x) + x2 + x −
9 81 729

Example 13. Find the solution of a nonhomogeneous equation

y 00 + 2y 0 + y = 2 cos 2x − 3x + 2 + 3ex . (vi)

Proof. The corresponding homogeneous equation is given by y 00 + 2y 0 + y = 0


The characteristic polynomial is given by r2 + 2r + 1 = 0.
Hence r1 = −1 and r2 = −1 are the roots.
Therefore, the solution of the homogeneous equation is given by

ψ(x) = (c1 + c2 x)e−x

which is the complementary solution.


To find the particular solution, consider the function y = a sin 2x+b cos 2x+cx+d+f ex
as a solution where a, b, c, d and f are the undetermined coefficients.

y = a sin 2x + b cos 2x + cx + d + f ex
y 0 = 2a cos 2x − 2b sin 2x + c + f ex

17
Lecture Notes on Differential Equations (MAT-CC-513/MATF-CC-513) Dr Dilip Kumar

y 00 = −4a sin 2x − 4b cos 2x + f ex


Substituting these in equation (vi) we get,
−4a sin 2x − 4b cos 2x + f ex + 2(2a cos 2x − 2b sin 2x + c + f ex ) + a sin 2x + b cos 2x
+ cx + d + f ex = 2 cos 2x − 3x + 2 + 3ex .
Equating the coefficients of ex , sin 2x, cos 2x, x and constant terms, on both sides, we
get
−3a − 4b = 0, 4a − 3b = 2, c = −3, d + 2c = 2, and 4f = 3.
8 6
On solving we get a = 25 , b = − 25 , c = −3, d = 8 and f = 43 .
Therefore the particular solution is
8 6 3
ψp =sin 2x − cos 2x − 3x + 8 + ex
25 25 4
The general solution is given by
8 6 3
φ = (c1 + c2 x)e−x + sin 2x − cos 2x − 3x + 8 + ex
25 25 4

5.2 Exceptions in Method of Undetermined Coefficients


Example 14. Solve the nonhomogeneous equation
y 00 + 3y 0 + 2y = 4e−2x (vii)
Proof. Here the corresponding homogenous equation is y 00 + 3y 0 + 2y = 0. The charac-
teristic polynomial is given by
r2 + 3r + 2 = 0.
Hence r1 = −2 and r2 = −1 are the roots. Therefore, the solution of the homogeneous
equation is given by
ψ(x) = c1 e−2x + c2 e−x
which is the complementary solution. To find the particular solution, consider a function
y = ae−2x as a solution, where a is an undetermined coefficient.
Substituting this equation in (vii) we get,
4ae−2x − 6ae−2x + 2ae−2x = 4e−2x
0 = 4e−2x
which is impossible. Hence consider another function y = axe−2x as the solution, where
a is an undetermined coefficient.
Substituting this equation in (vii) we get,
4axe−2x − 4ae−2x − 6axe−2x + 3ae−2x + 2axe−2x = 4e−2x
ae−2x = 4e−2x
a = −4

18
Lecture Notes on Differential Equations (MAT-CC-513/MATF-CC-513) Dr Dilip Kumar

Thus the particular solution is ψp = −4xe−2x . Hence the general solution is given by

φ(x) = c1 e−2x + c2 e−x − 4xe−2x .

Example 15. Solve the equation

y 00 + 4y = 6 sin 2x + 3x2 . (viii)

Proof. The corresponding homogeneous equation is given by y 00 + 4y = 0


The characteristic polynomial is given by r2 + 4 = 0.
Hence r1 = −2i and r2 = 2i are the roots.
Therefore, the solution of the homogeneous equation is given by

ψ(x) = (c1 cos 2x + c2 sin 2x)

which is the complementary solution.


To find the particular solution, consider the function y = a sin 2x+b cos 2x+cx2 +dx+f
as a solution where a, b, c, d and f are the undetermined coefficients.

y = a sin 2x + b cos 2x + cx2 + dx + f


y 0 = 2a cos 2x − 2b sin 2x + 2cx + d
y 00 = −4a sin 2x − 4b cos 2x + 2c
Substituting these in equation (viii) we see that this is an impossible solution. Hence
assuming the function y = x(a sin 2x + b cos 2x) + cx2 + dx + f as a solution where a,
b, c, d and f are the undetermined coefficients.

y = x(a sin 2x + b cos 2x) + cx2 + dx + f


y 0 = x(2a cos 2x − 2b sin 2x) + (a sin 2x + b cos 2x) + 2cx + d
y 00 = x(−4a sin 2x − 4b cos 2x) + 2(2a cos 2x − 2b sin 2x) + 2c
Substituting these in equation (viii) we

x(−4a sin 2x − 4b cos 2x) + 2(2a cos 2x − 2b sin 2x) + 2c + 4(x(a sin 2x + b cos 2x)
+ cx2 + dx + f ) = 6 sin 2x + 3x2
(4a cos 2x − 4b sin 2x) + 4cx2 + 4dx + 4f + 2c = 6 sin 2x + 3x2

Equating the coefficients of sin 2x, cos 2x, x2 , x and constant terms, on both sides, we
get
4a = 0, −4b = 6, 4c = 3, 4d = 0, and 4f + 2c = 0.

19
Lecture Notes on Differential Equations (MAT-CC-513/MATF-CC-513) Dr Dilip Kumar

On solving we get a = 0, b = − 32 , c = 34 , d = 0 and f = − 38 .


Therefore the particular solution is
3 3 3
ψp = − x cos 2x + x2 −
2 4 8
The general solution is given by
3 3 3
φ = (c1 cos 2x + c2 sin 2x) − x cos 2x + x2 −
2 4 8

Example 16. Solve the equation


y 000 − 3y 00 + 3y 0 − y = 2ex . (ix)
Proof. Here the corresponding homogenous equation is y 000 − 3y 00 + 3y 0 − y = 0. The
characteristic polynomial is given by
r3 − 3r2 + 3r − 1 = 0.
Hence r1 = 1, r2 = 1 and r3 = 1 are the roots. Therefore, the solution of the homoge-
neous equation is given by
ψ(x) = (c1 + c2 x + c3 x2 )ex
which is the complementary solution. To find the particular solution, considering a
function y = aex as a solution, where a is an undetermined coefficient, we get an im-
possible situation. Considering y = axex and y = ax2 ex also gives impossible situations.
Hence we consider y = ax3 ex as a solution, where a is an undetermined coefficient.

y = ax3 ex
y0 = ax3 ex + 3ax2 ex
y 00 = ax3 ex + 6ax2 ex + 6axex
y 000 = ax3 ex + 9ax2 ex + 18axex + 6aex
Substituting these in equation (ix) we get,
ax3 ex + 9ax2 ex + 18axex + 6aex − 3(ax3 ex + 6ax2 ex + 6axex )
+ 3(ax3 ex + 3ax2 ex ) − ax3 ex = 2ex
6aex = 2ex
1
a=
3
Thus the particular solution is ψp = 31 x3 ex . Hence the general solution is given by
1
φ(x) = (c1 + c2 x + c3 x2 )ex + x3 ex .
3

20
Lecture Notes on Differential Equations (MAT-CC-513/MATF-CC-513) Dr Dilip Kumar

General Rule
The general rule to solve a linear equation with constant coefficients are
1. Write the complementary solution ψ.

2. Assume the particular solution corresponding to the terms on the right hand side
of the equation.

(a) For a polynomial of degree n, assume a polynomial of degree n.


(b) For terms sin px, cos px or sum and differences of such terms, assume a sin px+
b cos px
(c) For terms like epx assume aepx .

3. If any of the assumed terms in 2(a), (b) or (c) occur in the complementary solution,
we must multiply these assumed terms by a power of x which is sufficiently high
so that none of these assumed terms will occur in the complementary solution.

4. Write the assumed form for the particular solution and evaluate the coefficients,
thus, obtaining, ψp .

5. Adding ψ and ψp , one get the required general solution.


5.3 Method of Undetermined Coefficients for More Complicated Cases
Example 17. Solve the equation

y 00 + y 0 + y = x3 ex . (x)

Proof. The corresponding homogeneous equation is given by y 00 + y 0 + y = 0.


The characteristic√polynomial is given√by r2 + r + 1 = 0.
Hence r1 = − 21 + 23i and r2 = − 12 − 23i are the roots.
Therefore, the solution of the homogeneous equation is given by
√ √
− x2 3 3
ψ(x) = e (c1 cos x + c2 sin x)
2 2
which is the complementary solution.
To find the particular solution, consider the function y = ax3 ex as a solution where a
is the undetermined coefficient.
On substituting this in equation (x), we cannot find a, and our assumption is wrong.
Hence we assume another function y = ax3 ex + bx2 ex + cxex + dex as the solution where
a, b, c, and d are undetermined coefficients.

y = ax3 ex + bx2 ex + cxex + dex

y 0 = ax3 ex + (3a + b)x2 ex + (2b + c)xex + (c + d)ex


y 00 = ax3 ex + (6a + b)x2 ex + (6a + 4b + c)xex + (2b + 2c + d)ex

21
Lecture Notes on Differential Equations (MAT-CC-513/MATF-CC-513) Dr Dilip Kumar

Substituting these in equation (x) we get,

ax3 ex + (6a + b)x2 ex + (6a + 4b + c)xex + (2b + 2c + d)ex + ax3 ex + (3a + b)x2 ex + (2b + c)xex
+ (c + d)ex + ax3 ex + bx2 ex + cxex + dex = x3 ex
3ax3 ex + (9a + 3b)x2 ex + (6a + 6b + 3c)xex + (2b + 3c + 3d)ex = x3 ex

Equating the coefficients of x3 ex , x2 ex , xex and ex , respectively on both sides, we get

3a = 1, 9a + 3b = 0, 6a + 6b + 3c = 0 and 2b + 3c + 3d = 0.

On solving we get a = 31 , b = −1, c = 43 and d = − 23 .


Therefore the particular solution is
1 4 2
ψp = x3 ex − x2 ex + xex − ex
3 3 3
The general solution is given by
√ √
− x2 3 3 1 4 2
φ = e (c1 cos x + c2 sin x) + x3 ex − x2 ex + xex − ex .
2 2 3 3 3

Example 18. Solve the equation

y 00 + y = x2 cos 5x (xi)

Proof. The corresponding homogeneous equation is given by y 00 + y = 0.


The characteristic polynomial is given by r2 + 1 = 0.
Hence r1 = i and r2 = −i are the roots.
Therefore, the solution of the homogeneous equation is given by

ψ(x) = (c1 cos x + c2 sin x)

which is the complementary solution.


To find the particular solution, consider the function y = (ax2 + bx + c) sin 5x + (dx2 +
f x + g) cos 5x as the solution where a, b and c are the undetermined coefficients.

y = (ax2 + bx + c) sin 5x + (dx2 + f x + g) cos 5x

y 0 = 5(ax2 + bx + c) cos 5x + (2ax + b) sin 5x − 5(dx2 + f x + g) sin 5x + (2dx + f ) cos 5x

y 00 = −25(ax2 + bx + c) sin 5x + 5(2ax + b) cos 5x + 5(2ax + b) cos 5x + 2a sin 5x


−25(dx2 + f x + g) cos 5x − 5(2dx + f ) sin 5x − 5(2dx + f ) sin 5x + 2d cos 5x

22
Lecture Notes on Differential Equations (MAT-CC-513/MATF-CC-513) Dr Dilip Kumar

Substituting these in equation (xi) we get,


−25(ax2 + bx + c) sin 5x + 5(2ax + b) cos 5x + 5(2ax + b) cos 5x + 2a sin 5x
−25(dx2 + f x + g) cos 5x − 5(2dx + f ) sin 5x − 5(2dx + f ) sin 5x + 2d cos 5x
+(ax2 + bx + c) sin 5x + (dx2 + f x + g) cos 5x = x2 cos 5x
Equating the coefficients of x2 sin 5x, x2 cos 5x, x sin 5x, x cos 5x, sin 5x and cos 5x re-
spectively on both sides, we get
−24a = 0, −25d+d = 1, −24b−20d = 0, −24f +20a = 0, −24c+2a−10f = 0 and −24g+2d+10b = 0.
5 1 19
On solving we get a = 0, b = 144 , c = 0,d = − 24 , f = 0 and g = 1728
.
Therefore the particular solution is
5 1 19
ψp =x sin 5x + (− x2 + ) cos 5x
144 24 1728
The general solution is given by
5 1 19
φ = (c1 cos x + c2 sin x) + x sin 5x + (− x2 + ) cos 5x.
144 24 1728

5.4 Method of Variation of Parameters


Consider a nonhomogeneous differential equation
L(y) = y 00 + a1 y 0 + a2 y = b(x). (19)
Our aim is to find the particular solution ϕp . Let
ϕp = u1 φ1 + u2 φ2
where u1 and u2 are certain functions defined on I. Now it is required to find u1 and
u2 . Suppose ϕp is a solution of L(y) = b(x),
ϕp = u1 φ1 + u2 φ2
ϕ0p = u1 φ01 + u01 φ1 + u2 φ02 + u02 φ2
ϕ00p = u1 φ001 + u01 φ01 + u01 φ01 + u001 φ1 + u2 φ002 + u02 φ02 + u02 φ02 + u002 φ2
then we get

L(ϕp ) = (u1 φ1 + u2 φ2 )00 + a1 (u1 φ1 + u2 φ2 )0 + a2 (u1 φ1 + u2 φ2 ) = b


= (u1 φ001 + u01 φ01 + u01 φ01 + u001 φ1 + u2 φ002 + u02 φ02 + u02 φ02 + u002 φ2 )
+ a1 (u1 φ01 + u01 φ1 + u2 φ02 + u02 φ2 ) + a2 (u1 φ1 + u2 φ2 ) = b
= u1 φ001 + a1 (u1 φ01 ) + a2 (u1 φ1 ) + u2 φ002 + a1 (u2 φ02 ) + a2 (u2 φ2 )
+ (u001 φ1 + u002 φ2 ) + 2(u01 φ01 + u02 φ02 ) + a1 (u01 φ1 + u02 φ2 ) = b
= u1 L(φ1 ) + u2 L(φ2 ) + (u001 φ1 + u002 φ2 ) + 2(u01 φ01 + u02 φ02 ) + a1 (u01 φ1 + u02 φ2 ) = b

23
Lecture Notes on Differential Equations (MAT-CC-513/MATF-CC-513) Dr Dilip Kumar

Since L(φ1 ) = 0 and L(φ2 ) = 0 we have

L(ϕp ) = (u001 φ1 + u002 φ2 ) + 2(u01 φ01 + u02 φ02 ) + a1 (u01 φ1 + u02 φ2 ) = b. (20)

Now suppose
(u01 φ1 + u02 φ2 ) = 0 (21)
Differentiating (21) with respect to x implies

0 = (u01 φ1 + u02 φ2 )0 = (u001 φ1 + u01 φ01 + u002 φ2 + u02 φ02 )

Substituting this in equation (20), we get

u01 φ01 + u02 φ02 = b (22)

Solving the equation (21) and (22), we get

(21) × φ01 ⇒ u01 φ1 φ01 + u02 φ2 φ01 = 0 (x)


(22) × φ1 ⇒ u01 φ1 φ01 + u02 φ1 φ02 = φ1 b (y)

(y) − (x) ⇒ u02 (φ1 φ02 − φ2 φ01 ) = φ1 b


φ1 b
u02 =
W (φ1 , φ2 )(x)
Substituting in (21) we get
−φ2 b
u01 =
W (φ1 , φ2 )(x)
Rx φ2 b
For u1 and u2 , integrate above equations from x0 to x, then one can get u1 = − x0 W (φ1 ,φ2 )(t)
dt
Rx
and u2 = x0 W (φφ1 ,φ
1b
2 )(t)
dt.
Hence, the particular solution ϕp becomes,
Z x
[φ2 (x)φ1 (t) − φ2 (t)φ1 (x)]b(t)
ϕp = dt
x0 W (φ1 , φ2 )(t)

Example 19. Find the solution of a non-homogeneous equation

y 00 − y 0 − 2y = e−x .

Proof. The corresponding homogeneous equation is given by y 00 − y 0 − 2y = 0.


The characteristic polynomial is given by r2 − r − 2 = 0.
Hence r1 = 2 and r2 = −1 are the roots.
Therefore, the solution of the homogeneous equation is given by

ψ(x) = (c1 e2x + c2 e−x )

24
Lecture Notes on Differential Equations (MAT-CC-513/MATF-CC-513) Dr Dilip Kumar

which is the complementary solution.


Here φ1 (x) = e2x and φ2 (x) = e−x . Then φ01 (x) = 2e2x and φ02 (x) = −e−x
W (φ1 , φ2 )(x) = −e2x e−x − 2e2x e−x = −ex − 2ex = −3ex
The particular solution is
ϕp = u1 φ1 + u2 φ2 .
Then
−e−2x
Z Z Z
φ2 b 1 1
u1 = − dx = x
= e−3x dx = − e−3x
W (φ1 , φ2 )(x) −3e 3 9
and
ex
Z Z Z
φ1 b 1 1
u2 = dx = dx = − dx = − x.
W (φ1 , φ2 )(x) −3ex 3 3
Therefore
1 1
ϕp = u1 φ1 + u2 φ2 = − e2x e−3x − xe−x .
9 3
Hence the complete solution is
1 1 1 1
φ = ψ + ϕp = (c1 e2x + c2 e−x ) − e2x e−3x − xe−x = (c1 e2x + c2 e−x ) − e−x − xe−x
9 3 9 3

Example 20. Find the solution of a non-homogeneous equation


π π
y 00 + y = tan x, − < x < .
2 2
Proof. The corresponding homogeneous equation is given by y 00 + y = 0.

The characteristic polynomial is given by r2 + 1 = 0.


Hence r1 = i and r2 = −i are the roots.
Therefore, the solution of the homogeneous equation is given by
ψ(x) = (c1 cos x + c2 sin x)
which is the complementary solution.
Here φ1 (x) = cos x and φ2 (x) = sin x. Then φ01 (x) = − sin x and φ02 (x) = cos x
W (φ1 , φ2 )(x) = cos2 x + sin2 x = 1
The particular solution is
ϕp = u1 φ1 + u2 φ2 .
Then
sin2 x
Z Z Z
φ2 b sin x tan x
u1 = − dx = − dx = − dx
W (φ1 , φ2 )(x) 1 cos x
1 − cos2 x
Z Z
=− dx = − (sec x − cos x)dx = sin x − ln(sec x + tan x)
cos x

25
Lecture Notes on Differential Equations (MAT-CC-513/MATF-CC-513) Dr Dilip Kumar

and Z Z Z
φ1 b cos x tan x
u2 = dt = dx = sin xdx = − cos x.
W (φ1 , φ2 )(x) 1
Therefore

ϕp = u1 φ1 +u2 φ2 = (sin x−ln(sec x+tan x)) cos x−cos x sin x = − ln(sec x+tan x) cos x.

Hence the complete solution is

φ = ψ + ϕp = (c1 cos x + c2 sin x) − ln(sec x + tan x) cos x.

5.5 Operational Methods


Consider a non-homogeneous differential equation of the form

L(D)y = b(x) (23)


d
where L(D)y is called a linear differential operator and in which D = dx , so L(D) =
2
D + a1 D + a2 and here a1 and a2 are constants. To find the solution of equation (23),
generally, we have
1
y= b(x).
L(D)
1
Here, L(D) meant by an operation to be performed on b(x).
” Consider a simple function as y 0 = x. Symbolically we write y = 1
D
x. By this we
mean
x2
Z
y = xdx = +c
2
So, Z
1
x = xdx.
D
In a similar way
1
D2
= twice integration.
1
D3
= trice integration and so on.
Now consider
(D − p)y = b(x) (24)
where p is a constant. For this equation we have
1
y= b(x) (25)
D−p
The solution of (24), which is in the form of a linear differential equation can be obtained
as Z
y=e px
e−px b(x)dx

26
Lecture Notes on Differential Equations (MAT-CC-513/MATF-CC-513) Dr Dilip Kumar

R
dy P (x)dx
[Linear differential
R equation is dx + P (x)y = Q(x). IF = e and Solution is
IF × y = IF × Q(x)dx.]
So by equation (25), we mean that
Z
1
b(x) = epx
e−px b(x)dx.
D−p
If the operator is of the form

(D − p)(D − q)y = b(x)

Then Z
1 1
y= b(x) = (eqx e−qx b(x)dx).
(D − p)(D − q) D−p
Therefore, Z Z
−px
y=e px
e qx
(e e−qx b(x)dx)dx
or
A1 A2
y= b(x) + b(x)(by partial fraction)
(D − p) (D − q)
That is, Z Z
−px
y = A1 (e px
e b(x)dx) + A2 (e qx
e−qx b(x)dx)

Example 21. Solve the equation

y 00 − y = e−x .

Proof. The corresponding homogeneous equation is given by y 00 − y = 0.


The characteristic polynomial is given by r2 − 1 = 0.
Hence r1 = 1 and r2 = −1 are the roots.
Therefore, the solution of the homogeneous equation is given by

ψ(x) = (c1 ex + c2 e−x )

which is the complementary solution.


Now the given equation can be written in the operator form as

(D2 − 1)y = e−x

(D − 1)(D + 1)y = e−x


1
y= e−x .
(D − 1)(D + 1)
By the method of partial fraction we have
1 A B
= +
(D − 1)(D + 1) (D − 1) (D + 1)

27
Lecture Notes on Differential Equations (MAT-CC-513/MATF-CC-513) Dr Dilip Kumar

Thus we get
1 = A(D + 1) + B(D − 1)
1
Putting D = 1 and D = −1 we respectively get A = 2
and B = − 12 . Thus we have

1 1 1 1 1
y= e−x = e−x − e−x .
(D − 1)(D + 1) 2 (D − 1) 2 (D + 1)
Hence Z Z
1 x 1 −x
y = (e ) e e dx − (e ) ex e−x dx
−x −x
2 2
1 1 1 1
y = − (ex e−2x ) − (e−x x) = − e−x − xe−x
4 2 4 2
Therefore, the complete solution is given by
1 1
φ = (c1 ex + c2 e−x ) − e−x − xe−x .
4 2

Example 22. Solve the equation

y 00 + 4y 0 + 4y = x3 e−2x .

Proof. The corresponding homogeneous equation is given by y 00 + 4y 0 + 4y = 0.


The characteristic polynomial is given by r2 + 4r + 4 = 0.
Hence r1 = −2 and r2 = −2 are the roots.
Therefore, the solution of the homogeneous equation is given by

ψ(x) = (c1 + c2 x)e−2x

which is the complementary solution.


Now the given equation can be written in the operator form as

(D2 + 4D + 4)y = x3 e−2x

(D + 2)(D + 2)y = x3 e−2x

1
y= x3 e−2x
(D + 2)(D + 2)
   Z 
1 1 3 −2x 1 −2x 2x 3 −2x
= xe = e e x e dx
(D + 2) (D + 2) (D + 2)
4
x4
  Z
1 −2x x −2x
= e =e e2x e−2x dx
(D + 2) 4 4
5
x
= e−2x
20

28
Lecture Notes on Differential Equations (MAT-CC-513/MATF-CC-513) Dr Dilip Kumar

Therefore, the complete solution is given by

x5 −2x
φ = (c1 + c2 x)e−2x + e .
20

Example 23. Solve the equation

y 00 − y 0 + y = x3 − 3x2 + 1.

Proof. The corresponding homogeneous equation is given by y 00 − y 0 + y = 0.


The characteristic
√ polynomial is given
√ by r2 − r + 1 = 0.
Hence r1 = 21 + 23i and r2 = 12 − 23i are the roots.
Therefore, the solution of the homogeneous equation is given by
√ √
x 3 3
ψ(x) = e 2 (c1 cos x + c2 sin x)
2 2
which is the complementary solution.
Now the given equation can be written in the operator form as

(D2 − D + 1)y = x3 − 3x2 + 1

This operator is not possible to write in simple factors and hence we write in the form
1
y= x3 − 3x2 + 1
(1 − D + D2 )
= (1 − (D − D2 ))−1 (x3 − 3x2 + 1)
= (1 + (D − D2 ) + (D − D2 )2 + (D − D2 )3 + (D − D2 )4 + · · · )(x3 − 3x2 + 1)
= (1 + (D − D2 ) + (D2 − 2D3 + D4 ) + (D3 − 3D4 + 3D5 − D6 )
+ (D4 − 4D5 + 6D6 − 4D7 + D8 ) + · · · )(x3 − 3x2 + 1)
= (1 + D − D3 − D4 + · · · )(x3 − 3x2 + 1)
= x3 − 3x2 + 1 + 3x2 − 6x − 6 = (x3 − 6x − 5)

[(1 − x)−1 = 1 + x + x2 + x3 + · · · , |x| < 1]


Therefore, the complete solution is given by
√ √
x 3 3
φ = e 2 (c1 cos x + c2 sin x) + x3 − 6x − 5.
2 2

29
Lecture Notes on Differential Equations (MAT-CC-513/MATF-CC-513) Dr Dilip Kumar

6 Linear Homogeneous Equations of Order n


The linear homogeneous differential equation of order n is of the form

L(y) = y (n) + a1 y (n−1) + · · · + an y = 0 (26)

where a1 , a2 , . . . , an are constants.


Now we try to find the solution of equation (27) using the exponential function erx as
in the previous section(Section 2) for second order equations.

L(erx ) = rn erx + a1 rn−1 erx + · · · + an erx


= erx (rn + a1 rn−1 + · · · + an )
= erx p(r)

Where p(r) = rn + a1 rn−1 + · · · + an is called as the characteristic polynomial.


Note:If r1 is a repeated root of p(r) then p(r1 ) = p0 (r1 ) = · · · = p(n−1) (r1 ) = 0.

Example 24. Find the solution of y 000 − 8y = 0.

Proof. The characteristic


√ polynomial
√ is given by r3 − 8 = 0 or r3 = 8. Then the roots
are r = 2, −1 + i 3, −1 − i 3.
Therefore, the solution is given by
√ √
φ(x) = c1 e2x + e−x (c2 cos 3x + c3 sin 3x).

Example 25. Find the solution of y (4) + λy = 0 in the following cases:


(a)λ < 0 (b)λ = 0 (c)λ > 0.

Proof. Consider the equation


y (4) + λy = 0.
Case a) Suppose λ is negative. ie, λ < 0 ⇒ λ = −k 4 , where k is positive, then

y (4) − k 4 y = 0.

The characteristic polynomial of this equation is r4 − k 4 = 0 which has the roots


r = ±k, ±ik. Hence the solution is

φ(x) = c1 ekx + c2 e−kx + c3 cos kx + c4 sin kx.

Case b) Suppose λ is positive. ie, λ > 0 ⇒ λ = k 4 , where k is positive, then

y (4) + k 4 y = 0.

30
Lecture Notes on Differential Equations (MAT-CC-513/MATF-CC-513) Dr Dilip Kumar

The characteristic polynomial of this equation is r4 + k 4 = 0 or r4 = −k 4 which has the


1
roots r = √k2 (1±i), √k2 (−1±i)[Hint: Take r = k(−1) 4 = k(cos (2n−1)π
4
+i sin (2n−1)π
4
), n =
0, 1, 2, 3]. Hence the solution is
k
√ x k k − √k x k k
φ(x) = e 2 (c1 cos √ x + c2 sin √ x) + e 2 (c3 cos √ x + c4 sin √ x).
2 2 2 2
Case c) Suppose λ is equal to zero. ie, λ = 0, then

y (4) = 0.

The roots of the equation are r = 0, 0, 0, 0. Hence the solution is

φ(x) = c1 + c2 x + c3 x2 + c4 x3 .

Example 26. Find the solution of y (4) + 16y = 0.


Proof. The characteristic polynomial is given by r4 + 16 = 0 or r4 = −16.
By case b) of the previous example, we have the roots as r = √22 (1 ± i), √22 (−1 ± i) =
√ √
2(1 ± i), 2(−1 ± i).
Therefore, the solution is given by
√ √ √ √ √ √
φ(x) = e 2x (c1 cos 2x + c2 sin 2x) + e− 2x (c3 cos 2x + c4 sin 2x).

7 Non-homogeneous Equations of Order n


Definition 4. The n functions φ1 , φ2 , . . . , φn having (n − 1) derivatives on an interval
I is defined as the determinant function

φ1 φ2 ... φn
φ01 φ02 ... φ0n
W (φ1 , φ2 , . . . , φn ) = .. .. ..
. . ··· .
(n−1) (n−1) (n−1)
φ1 φ2 . . . φn

its value at x in I being W (φ1 , φ2 , . . . , φn )(x).


7.1 Method of Variation of Parameters
The linear non-homogeneous differential equation of order n is of the form

L(y) = y (n) + a1 y (n−1) + · · · + an y = b(x) (27)

where a1 , a2 , . . . , an are arbitrary constants and b(x) is certain continuous function


defined on an interval I.

31
Lecture Notes on Differential Equations (MAT-CC-513/MATF-CC-513) Dr Dilip Kumar

Let ϕp be any particular solution of L(y) = b(x) and ϕ be any other solution of L(y) =
b(x), so we have,
L(ϕ − ϕp ) = L(ϕ) − L(ϕp ) = b − b = 0.
Hence (ϕ − ϕp ) forms a solution to the homogeneous equation L(y) = 0. So the solution
of L(y) = 0 is in the form

ϕ − ϕp = c1 φ1 + c2 φ2 + · · · + cn φn

or
ϕ = ϕp + c1 φ1 + c2 φ2 + · · · + cn φn
where ϕp is a particular solution and φ1 , φ2 , . . . , φn are assumed to be linearly indepen-
dent solutions of L(y) = 0.
Now our aim is to find a particular solution ϕp of L(y) = b(x). To find the particular
solution we use the method of variation of parameters. Let

ϕp = u1 φ1 + u2 φ2 + · · · + un φn (28)

Now it requires to find the functions u1 , u2 , . . . , un . Now differentiating the equation


(28), we get

ϕ0p = u1 φ01 + u01 φ1 + u2 φ02 + u02 φ2 + · · · + un φ0n + u0n φn


ϕ0p = (u1 φ01 + u2 φ02 + · · · + un φ0n ) + (u01 φ1 + u02 φ2 + · · · + u0n φn ).

Let
u01 φ1 + u02 φ2 + · · · + u0n φn = 0 (29)
Again differentiating ϕ0p , we get

ϕ00p = u1 φ001 + u01 φ01 + u2 φ002 + u02 φ02 + · · · + un φ00n + u0n φ0n
ϕ00p = (u1 φ001 + u2 φ002 + · · · + un φ00n ) + (u01 φ01 + u02 φ02 + · · · + u0n φ0n ).

Again set
u01 φ01 + u02 φ02 + · · · + u0n φ0n = 0 (30)
The third derivative of ϕp gives

ϕ000 000 0 00 000 0 00 000 0 00


p = u1 φ1 + u1 φ1 + u2 φ2 + u2 φ2 + · · · + un φn + un φn
ϕ000 000 000 000 0 00 0 00 0 00
p = (u1 φ1 + u2 φ2 + · · · + un φn ) + (u1 φ1 + u2 φ2 + · · · + un φn ).

Further, let
u01 φ001 + u02 φ002 + · · · + u0n φ00n = 0 (31)
In general,
(n) (n) 0 (n−1) (n−1)
ϕ(n) (n)
p = (u1 φ1 + u2 φ2 + · · · + un φn ) + (u1 φ1 + u02 φ2 + · · · + u0n φ(n−1)
n ).

32
Lecture Notes on Differential Equations (MAT-CC-513/MATF-CC-513) Dr Dilip Kumar

By adding all of the above equation, we get


(n−1) (n−1)
L(ϕp ) = (u1 L(φ1 ) + u2 L(φ2 ) + · · · un L(φn ) + (u01 φ1 + u02 φ2 + · · · + u0n φ(n−1)
n ).
Hence
(n−1) (n−1)
u01 φ1 + u02 φ2 + · · · + u0n φ(n−1)
n =b (32)
Now, by solving the n equations from (29) and (32), we get
b(x)Wk (x)
u0k =
W (x)
where W (x) is the Wronskian of φ1 , φ2 , . . . , φn and Wk (x) is the determinant in which
the k th column is replaced by (0, 0, . . . , 1)T . Since
W (φ1 , φ2 , . . . , φn ) 6= 0
u0k ’s are well defined. For a fixed x0 , we get
Z x
b(t)Wk (t)
uk = dt, for all k = 1, 2, 3, . . . , n.
x0 W (t)
So the particular solution is given by
n Z x
X b(t)Wk (t)
φk (x) dt.
k=1 x0 W (t)

Example 27. Find the solution of a non-homogeneous differential equation


y 000 + y 00 + y 0 + y = 1
satisfying φ(0) = 1, φ0 (0) = 1, φ00 (0) = 0
Proof. The corresponding homogeneous equation is
y 000 + y 00 + y 0 + y = 0.
The characteristic polynomial of the given equation is r3 + r2 + r + 1 = 0. By trial
method r = −1 is a root. Dividing r3 + r2 + r + 1 = 0 by r + 1 we get r2 + 1 which has
roots r = +i and r = −i. Hence the roots of r3 + r2 + r + 1 = 0 are r = −1, +i, −i.
Therefore the complementary solution is ψ = c1 e−x + c2 cos x + c3 sin x.
Hence φ1 = e−x , φ2 = cos x, φ3 = sin x. Let
ϕp = u1 φ1 + u2 φ2 + u3 φ3
be the particular solution. Then using the method of variation of parameters, we have
e−x cos x sin x
−x
W (φ1 , φ2 , φ3 )(x) = −e − sin x cos x
e−x − cos x − sin x
= e−x (sin2 x + cos2 x) − cos x(e−x sin x − e−x cos x) + sin x(e−x cos x + e−x sin x)
= 2e−x .

33
Lecture Notes on Differential Equations (MAT-CC-513/MATF-CC-513) Dr Dilip Kumar

0 cos x sin x
W1 (x) = 0 − sin x cos x
1 − cos x − sin x
= 1(cos2 x + sin2 x) = 1.

e−x 0 sin x
W2 (x) = −e−x 0 cos x
e−x 1 − sin x
= −1(e−x cos x + e−x sin x) = −e−x (cos x + sin x).

e−x cos x 0
−x
W3 (x) = −e − sin x 0
e−x − cos x 1
= 1(−e−x sin x + e−x cos x) = e−x (cos x − sin x).

Therefore,
Z Z
1 1 x 1 x
u1 = dx = e dx = e
2e−x 2 2
−e−x (cos x + sin x)
Z Z
1
u2 = dx = − (cos x + sin x)dx
2e−x 2
1
= (cos x − sin x)
Z2 −x
e (cos x − sin x)
Z
1
u3 = dx = (cos x − sin x)dx
2e−x 2
1
= (cos x + sin x)
2

Hence, the particular solution is given by

ϕp = u1 φ1 + u2 φ2 + u3 φ3

Therefore,
1 1 1
ϕp = ex e−x + (cos x − sin x) cos x + (cos x + sin x) sin x
2 2 2
1
= (1 + cos x − sin x cos x + cos x sin x + sin2 x) = 1
2
2
Hence the complete solution is

φ(x) = ψ + ϕp = c1 e−x + c2 cos x + c3 sin x + 1

34
Lecture Notes on Differential Equations (MAT-CC-513/MATF-CC-513) Dr Dilip Kumar

φ0 (x) = −c1 e−x − c2 sin x + c3 cos x


φ00 (x) = c1 e−x − c2 cos x − c3 sin x
Using the initial conditions, we get

c1 + c2 = 0
−c1 + c3 = 1
c1 − c2 = 0

one can get c1 = c2 = 0, c3 = 1. Hence the complete solution is φ(x) = ψ(x) + ϕp (x) =
sin x + 1.

8 Linear Equations with Variable Coefficients


A general linear homogeneous differential equation with variable coefficients is in the
form
L(y) = y (n) + a1 (x)y (n−1) + a2 (x)y (n−2) + · · · + an (x)y = 0 (33)
where a1 (x), a2 (x), . . . , an (x) are certain continuous function defined on an interval I.

8.1 Reduction of Order of Homogeneous Differential Equations


Consider the second order differential equation of the form

L(y) = y 00 + a1 (x)y 0 + a2 (x)y = 0. (34)

We have to reduce the order of equation (34) if one of the solutions of this equation is
given.
Let φ1 be given and let φ2 = uφ1 , where u is a function which is to be found out.
Then
φ2 = uφ1
φ02 = uφ01 + u0 φ1
φ002 = uφ001 + 2u0 φ01 + u00 φ1
Consider

L(φ2 ) = L(uφ1 )
= (u00 φ1 + 2u0 φ01 + uφ001 ) + a1 (x)(u0 φ1 + uφ01 ) + a2 (x)(uφ1 ) = 0
= (uφ001 + a1 (x)uφ01 + a2 (x)uφ1 + u00 φ1 + 2u0 φ01 + a1 (x)u0 φ1 = 0
= uL(φ1 ) + u00 φ1 + 2u0 φ01 + a1 (x)u0 φ1 = 0.

Since φ1 is a solution, we have L(φ1 ) = 0 and we get,

u00 φ1 + 2u0 φ01 + a1 (x)u0 φ1 = 0. (35)

35
Lecture Notes on Differential Equations (MAT-CC-513/MATF-CC-513) Dr Dilip Kumar

Now set u0 = v. then (35) reduces to

v 0 φ1 + 2vφ01 + a1 (x)vφ1 = 0.

Multiplying this equation throughout by φ1 , we get

v 0 φ21 + 2vφ01 φ1 + a1 (x)vφ21 = 0.

It can be written as
(vφ21 )0 = −a1 (x)vφ21 .
That is
(vφ21 )0
= −a1 (x).
vφ21
For a fixed x0 , integrate the above equation from x0 to x to get
Z x
2
ln(vφ1 ) = − a1 (t)dt.
x0

That is Rx
− a1 (t)dt
vφ21 = e x0
.
Hence,
1 − Rxx a1 (t)dt
v= e 0 .
φ21
But we know that v = u0 . Therefore,
1 − Rxx a1 (t)dt
u0 = e 0 .
φ21

So, for some fixed x0 , integrating the above equation from x0 to x, we get
Z x
1 − Rxx a1 (t)dt
u= 2
e 0 dt.
x0 φ1

So, the second solution is given by


Z x
1 − Rxx a1 (t)dt
φ2 = φ1 e 0 dt.
x0 φ21

Example 28. Find the solution of L(y) = x2 y 00 − 7xy 0 − 15y = 0 where one of the
solution is given as φ1 = x3 , (x < 0).

Proof.
L(y) = x2 y 00 − 7xy 0 − 15y = 0

36
Lecture Notes on Differential Equations (MAT-CC-513/MATF-CC-513) Dr Dilip Kumar

L(y) = y 00 − x7 y 0 − x152 y = 0.
Here a1 = − x7 and φ1 = x3 . We know that
Z
1 − R a1 (x)dx
u= e dx
φ2
Z 1 R
1 7 dx
= e x dx
x6
Z
1 7 ln(x)
= e dx
x6
Z 7
x2
Z
x
= dx = xdx =
x6 2

2 5
Therefore, φ2 = uφ1 = x2 x3 = x2 .
5
Hence the solution is φ = c1 φ1 + c2 φ2 = c1 x3 + c2 x2 .

Theorem 8 (Formula for Wronskian). Let φ1 and φ2 be two linearly independent solu-
tions of
L(y) = y 00 + a1 (x)y 0 + a2 (x)y = 0
on an interval I containing a point x0 , then
Rx
− a1 (t)dt
W (φ1 , φ2 )(x) = W (φ1 , φ2 )(x0 )e x0
, ∀x∈I

Proof. We know that

φ1 (x) φ2 (x)
W (φ1 , φ2 )(x) =
φ01 (x) φ02 (x)
= φ1 (x)φ02 (x) − φ2 (x)φ01 (x).

and

W 0 = φ1 (x)φ002 (x) + φ01 (x)φ02 (x) − φ2 (x)φ001 (x) − φ02 (x)φ01 (x)
= φ1 (x)φ002 (x) − φ2 (x)φ001 (x) (36)

Since φ1 and φ2 are solution s of L(φ) = 0, we get

L(φ1 ) = φ001 + a1 (x)φ01 + a2 (x)φ1 = 0


L(φ2 ) = φ002 + a1 (x)φ02 + a2 (x)φ2 = 0.

Hence,

φ001 = −a1 (x)φ01 − a2 (x)φ1


φ002 = −a1 (x)φ02 − a2 (x)φ2 .

37
Lecture Notes on Differential Equations (MAT-CC-513/MATF-CC-513) Dr Dilip Kumar

Now substituting these two equations in equation (36), we get

W 0 = φ1 (x)(−a1 (x)φ02 − a2 (x)φ2 ) − φ2 (x)(−a1 (x)φ01 − a2 (x)φ1 )


= −a1 (x)φ1 (x)φ02 − a2 (x)φ1 (x)φ2 + a1 (x)φ2 (x)φ01 + a2 (x)φ1 φ2 (x)
= −a1 (x)φ1 (x)φ02 + a1 (x)φ2 (x)φ01
= −a1 (x)(φ1 (x)φ02 − φ2 (x)φ01 )
= −a1 (x)W
0
W + a1 (x)W = 0.

The above equation is a first order linear differential equation with variable coefficient,
so that the solution is in the form
Rx
− a1 (t)dt
W (φ1 , φ2 )(x) = ce x0
.

Putting x = x0 in the above equation we get,

W (φ1 , φ2 )(x0 ) = c.

Therefore, Rx
− a1 (t)dt
W (φ1 , φ2 )(x) = W (φ1 , φ2 )(x0 )e x0
.

Theorem 9. Let φ1 , φ2 , φ3 , . . . , φn be n linearly independent solutions of

L(y) = y (n) + a1 (x)y (n−1) + a2 (x)y (n−2) + · · · + an (x)y = 0

on an interval I containing a point x0 , where a1 (x), a2 (x), . . . , an (x) are certain contin-
uous function defined on I, then
Rx
− a1 (t)dt
W (φ1 , φ2 , φ3 , . . . , φn )(x) = W (φ1 , φ2 , φ3 , . . . , φn )(x0 )e x0
, ∀ x ∈ I.

Proof. We know that the Wronskian is the determinant

φ1 φ2 ... φn
φ01 φ02 ... φ0n
W (φ1 , φ2 , . . . , φn ) = .. .. ..
. . ··· .
(n−1) (n−1) (n−1)
φ1 φ2 . . . φn

and the derivative of the Wronskian is sum of n determinants W1 , W2 , . . . , Wn , where

38
Lecture Notes on Differential Equations (MAT-CC-513/MATF-CC-513) Dr Dilip Kumar

each Wi is different from W in its ith row and is just the derivative of ith row of W . So

W 0 = W1 + W2 + . . . + Wn
φ01 φ02 ... φ0n φ1 φ2 ... φn
φ01 φ02 ... φ0n φ001 φ002 ... φ00n
= φ001 φ002 ... φ00n + φ001 φ002 ... φ00n
.. .. .. .. .. ..
. . ··· . . . ··· .
(n−1) (n−1) (n−1) (n−1) (n−1) (n−1)
φ1 φ2 . . . φn φ1 φ2 . . . φn
φ1 φ2 ... φn
φ01 φ02 ... φ0n
00
+ · · · + φ1 φ002 ... φ00n .
.. .. ..
. . ··· .
(n) (n) (n)
φ1 φ2 . . . φn
Since the value of the determinant containing two identical rows or columns is zero,
hence, the above equation becomes
φ1 φ2 ... φn
φ01 φ02 ... φ0n
00
W 0 = φ1 φ002 ... φ00n . (37)
.. .. ..
. . ··· .
(n) (n) (n)
φ1 φ2 . . . φn

Since φ1 , φ2 , . . . , φn are solutions of L(y) = 0, we get


(n) (n−1) (n−2)
L(φ1 ) = φ1 + a1 (x)φ1 + a2 (x)φ1 + · · · + an (x)φ1 = 0
(n) (n−1) (n−2)
L(φ2 ) = φ2 + a1 (x)φ2 + a2 (x)φ2 + · · · + an (x)φ2 = 0
..
.
L(φn ) = φ(n) (n−1)
n + a1 (x)φn + a2 (x)φ(n−2)
n + · · · + an (x)φn = 0.

Hence
n−1
X
(n) (n−1) (n−2) (j)
φ1 = −a1 (x)φ1 − a2 (x)φ1 − · · · − an (x)φ1 = − an−j φ1
j=0
n−1
X
(n) (n−1) (n−2) (j)
φ2 = −a1 (x)φ2 − a2 (x)φ2 − · · · − an (x)φ2 = − an−j φ2
j=0
.. ..
. .
n−1
X
φ(n)
n = −a1 (x)φ(n−1)
n − a2 (x)φ(n−2)
n − · · · − an (x)φn = − an−j φ(j)
n
j=0

39
Lecture Notes on Differential Equations (MAT-CC-513/MATF-CC-513) Dr Dilip Kumar

By substituting these expressions in the equation (37) we get


φ1 φ2 ... φn
φ01 φ02 ... φ0n
W0 = φ001 φ002 ... φ00n .
.. .. ..
. . ··· .
Pn−1 (j) Pn−1 (j) (j)
. . . − n−1
P
− j=0 an−j φ1 − j=0 an−j φ2 j=0 an−j φn

Since the value of the determinant is unchanged when we add any row with a constant
multiple of any other row. So, multiplying the first row of the above determinant by
an , second row by an−1 , third row by an−2 , . . ., (n − 1)th row by a2 and add all these
rows to the last row, then we get
φ1 φ2 ... φn
φ01 φ02 ... φ0n
W0 = φ001 φ002 ... φ00n
.. .. ..
. . ··· .
(n−1) (n−1) (n−1)
−a1 φ1 −a1 φ2 . . . −a1 φn
= −a1 W.
Therefore W 0 + a1 W = 0. The above equation is a first order linear differential equation
with variable coefficient, so that the solution is in the form
Rx
− a1 (t)dt
W (φ1 , φ2 , . . . , φn )(x) = ce x0
.
Putting x = x0 in the above equation we get,
W (φ1 , φ2 , . . . , φn )(x0 ) = c.
Therefore, Rx
− a1 (t)dt
W (φ1 , φ2 , . . . , φn )(x) = W (φ1 , φ2 , . . . , φn )(x0 )e x0
.

Theorem 10. Let φ be any solution of L(y) = y 00 + a1 (x)y 0 + a2 (x)y = 0, where a1 and
a2 are certain functions defined on some interval I. Then φf forms another solution
for L(y) = 0 whenever f 0 satisfies (vφ2 )0 + a1 (x)(vφ2 ) = 0, (v = f 0 ) and then prove that
φ and φf are linearly independent.
Proof. Given
L(y) = y 00 + a1 (x)y 0 + a2 (x)y = 0 (38)
and φ forms one solution for equation (38). Now, we have to prove φf forms another
solution of equation (38). To prove this, consider
L(φf ) = (φf )00 + a1 (x)(φf )0 + a2 (x)φf = 0
φf 00 + 2φ0 f 0 + φ00 f + a1 (x)(φf 0 + φ0 f ) + a2 (x)φf = 0
φ00 f + a1 (x)φ0 f + a2 (x)φf + φf 00 + 2φ0 f 0 + +a1 (x)φf 0 = 0.

40
Lecture Notes on Differential Equations (MAT-CC-513/MATF-CC-513) Dr Dilip Kumar

Since φ is a solution, we have

φ00 f + a1 (x)φ0 f + a2 (x)φf = 0.

Hence
φf 00 + 2φ0 f 0 + a1 (x)φf 0 = 0.
Now, let f 0 = v, to get
φv 0 + 2φ0 v + a1 (x)φv = 0.
Multiplying throughout the equation by φ to have

φ2 v 0 + 2φ0 φv + a1 (x)φ2 v = 0.

(φ2 v)0 + a1 (x)φ2 v = 0. (39)


Hence φf forms another solution for equation (38).
From equation (39), one can get
Rx
2 − a1 (t)dt
φ v=e x0

Next, we have to prove φ and φf are linearly independent, so consider

W (φ, φf )(x) = φ(φf )0 − (φf )φ0


= φ(φf 0 + φ0 f ) − φφ0 f
= φ2 f 0 + φφ0 f − φφ0 f
Rx
− a1 (t)dt
= φ2 f 0 = φ2 v = e x0
6= 0

From this we conclude that the solutions φ and φf are linearly independent.
Example 29. If one solution of L(y) = x2 y 00 − 2y = 0, (x > 0) is φ1 = x2 , then find
all the solutions of
x2 y 00 − 2y = 2x − 1.
Proof. Consider the given equation

L(y) = x2 y 00 − 2y = 0.

That is
2
L(y) = y 00 − y = 0.
x2
Here a1 (x) = 0. Therefore,
Z
1 − R a1 (x)dx
φ2 (x) = φ1 (x) e dx
φ21
Z
2 1 1 1
=x 4
dx = −x2 3 = − .
x 3x 3x

41
Lecture Notes on Differential Equations (MAT-CC-513/MATF-CC-513) Dr Dilip Kumar

To find the particular solution of


x2 y 00 − 2y = 2x − 1
we write it as
2 2x − 1
2
y 00 −
y= .
x x2
Here b(x) = 2x−1
x2
. Let ϕp = u1 φ1 + u2 φ2 be the particular solution. Then using the
method of variation of parameters, we have
Z
b(x)φ2
u1 = − dx
W (x)
Z
b(x)φ1
u2 = dx.
W (x)
Here
1
x2 − 3x 1 2
W (x) = 1 = + = 1.
2x 3x2 3 3
Therefore,
Z (2x−1) 1
(− 3x )
Z
b(x)φ2 x2
u1 = − dx = − dx
W (x) 1
Z  
(2x − 1)
Z
1 1 2 1
= dx = − dx
3 x3 3 x2 x3
 
1 2 1
= − + 2
3 x 2x
Z
b(x)φ1
u2 = dx
W (x)
(2x−1) 2
x
Z
x2
= dx
1
Z
= (2x − 1)dx = (x2 − x).

Hence the particular solution is


ϕp = u1 φ1 + u2 φ2
 
1 2 1 1
= − + 2 x2 + (x2 − x)(− )
3 x 2x 3x
2x 1 x 1 1
= − + − + = −x + .
3 6 3 3 2
Hence the complete solution is given by
1 1
ϕ(x) = c1 φ1 + c2 φ2 + ϕp = c1 x2 − c2 −x+ .
3x 2

42
Lecture Notes on Differential Equations (MAT-CC-513/MATF-CC-513) Dr Dilip Kumar

8.2 Transformation of Equations with Variable Coefficients into Constant


Coefficients
Euler’s Differential Equation (Cauchy’s Differential Equation)
The equations of the form
L(y) = a0 xn y (n) + a1 xn−1 y (n−1) + a2 xn−2 y (n−2) + · · · + an y = b(x) (40)
where a1 , a2 , . . . , an are constants. This was first discovered by Leonard Euler and hence
are called Euler’s Differential Equation. Sometimes these are also called Cauchy’s Dif-
ferential Equation.
To solve these type of equations we reduce the given equations having variable coeffi-
cients into constant coefficients.

Example 30. Find the solution of the equation


L(y) = x2 y 00 + xy 0 + 4y = 1. (41)
Proof. Given that L(y) = x2 y 00 + xy 0 + 4y = 1.
For this problem we consider the transformation x = et . Here dx
dt
= et or dt
dx
= e−t .
Now
dy dt dy −t
y0 = = e
dt dx dt
dy
xy 0 =
dt
2
 
00 −t d y −t dy dt
y = e 2
−e
dt dt dx
 2 
−2t d y dy
=e −
dt2 dt
 2 
d y dy
x2 y 00 = − .
dt2 dt
Using these derivatives in the equation (41), we get
 2 
d y dy dy
2
− + + 4y = 1
dt dt dt
or
d2 y
+ 4y = 1 (a)
dt2
The corresponding homogeneous equation is given by y 00 + 4y = 0.
The characteristic polynomial is given by r2 + 4 = 0.
Hence r1 = 2i and r2 = −2i are the roots.
Therefore, the solution of the homogeneous equation is given by
ψ(t) = c1 cos 2t + c2 sin 2t

43
Lecture Notes on Differential Equations (MAT-CC-513/MATF-CC-513) Dr Dilip Kumar

which is the complementary solution.


To find the particular solution, consider the function y = a as a solution where a is the
undetermined coefficients.
Substituting y = a in (a),

4a = 1
1
a=
4
Therefore the particular solution is
1
ψp =
4
The general solution is given by
1
y(t) = c1 cos 2t + c2 sin 2t + .
4
Since x = et or t = ln x, therefore, the required solution becomes,
1
y(x) = c1 cos(2 ln x) + c2 sin(2 ln x) + .
4

Example 31. Find the solution of the equation

L(y) = x2 y 00 − 2xy 0 + 2y = 0. (42)

Proof. Given that L(y) = x2 y 00 − 2xy 0 + 2y = 0.


For this problem we consider the transformation x = et . Here dx
dt
= et or dt
dx
= e−t .
Now
dy dt dy −t
y0 = = e
dt dx dt
dy
xy 0 =
dt
2
 
00 −t d y −t dy dt
y = e 2
−e
dt dt dx
 2 
−2t d y dy
=e −
dt2 dt
 2 
d y dy
x2 y 00 = − .
dt2 dt

Using these derivatives in the equation (42), we get


 2 
d y dy dy
2
− − 2 + 2y = 0
dt dt dt

44
Lecture Notes on Differential Equations (MAT-CC-513/MATF-CC-513) Dr Dilip Kumar

or
d2 y dy
− 3 + 2y = 0 (b)
dt2 dt
The characteristic polynomial is given by r2 − 3r + 2 = 0.
Hence r1 = 2 and r2 = 1 are the roots.
Therefore, the general solution is given by

y(t) = c1 e2t + c2 et .

Since x = et , t = ln x, therefore, the required solution becomes,

y(x) = c1 x2 + c2 x.

Example 32. Find the solution of the equation

L(y) = xy 00 − (x + 1)y 0 + y = 0.

Proof. Given that L(y) = xy 00 − (x + 1)y 0 + y = 0.

xy 00 − xy 0 − y 0 + y = 0
x(y 00 − y 0 ) = (y 0 − y).

Now, let the transformation be u = y 0 − y. By substituting this in the above equation,


we get
xu0 = u
or
u0 1
= .
u x
Integrating on both sides, we get

ln u = ln x + ln c1

That is,
u = c1 x.
Since u = y 0 − y, we have
y 0 − y = c1 x.
Multiplying both sides of the equation by e−x , we get

e−x y 0 − e−x y = c1 xe−x .

This can be written in the form


d −x
(e y) = c1 xe−x .
dx

45
Lecture Notes on Differential Equations (MAT-CC-513/MATF-CC-513) Dr Dilip Kumar

On integration, we get
Z
−x −x
e y = −c1 xe + c1 e−x dx + c2

e−x y = −c1 xe−x − c1 e−x + c2


y = −c1 (x + 1) + c2 ex

Which is the required solution.

46

You might also like