(Module-I (MA107-Math-II, by R.

Singh) 1

Department of Mathematics,
Birla Institute of Technology Mesra, Ranchi
(MA107-Math-II by R. Singh)

Higher-Order Linear Differential(LD) Equations

A linear differential equation of order n is defined by

dn y dn−1 y dn−2 y
+ a1 (x) n−1 + a2 (x) n−2 + . . . + an (x)y = b(x), (1)
dxn dx dx

where a1 (x), a2 (x), . . . , an (x) and b(x) are real-valued function.

Higher-Order LD Equation with Constant Coefficients

A linear differential equation with constant coefficients of order n is given by

dn y dn−1 y dn−2 y
L(y) ≡ + a1 + a2 + . . . + an y = b(x), (2)
dxn dxn−1 dxn−2

d 2 d2
where a1 , a2 , . . . , an are constants. Denote D = , D = 2 , .... The LDE (2) can be represented
dx dx
as

(Dn + a1 Dn−1 + a2 Dn−2 + ... + an I)y = b(x), (3)

Homogeneous Equation: If b(x) = 0, the above equation is called homogeneous.

Non-Homogeneous Equation: If b(x) ̸= 0, then it is called a non-homogeneous equation.

Solution: A solution of (3) is a function ϕ(x) having n derivative such that L[ϕ(x)] = b(x).
If y1 , y2 , . . . , yn are n L.I. solutions of the homogeneous equation L(y) = 0 then

yc = c1 y1 + c2 y2 + . . . cn yn
(Module-I (MA107-Math-II, by R. Singh) 2

where c1 , c2 , . . . , cn are arbitrary constants, is called complementary function (C.F.)

Let yp be a particular solution or particular integral (P.I) of the nonhomogeneous equation L(y) =
b(x).

→ C.F yc involves n arbitrary constants

→ P.I yp does not involve any arbitrary constant.

The general solution of (3) is written as

y(x) = yc (x) + yp (x) (4)

Finding Solution of Homogeneous LDEs Ly = 0

Consider

Ly ≡ (Dn + a1 Dn−1 + a2 Dn−2 + ... + an )y = 0 (5)

Step-1 Substitute y = emx into (5), we obtain

(mn + a1 mn−1 + a2 mn−2 + ... + an )emx = 0 (6)

Since emx ̸= 0, so y = emx is a solution of (5) if and only if m is a solution of

mn + a1 mn−1 + a2 mn−2 + ... + an = 0 (7)

Equation (7) is the auxiliary equation (A.E.) or characteristic equation (C.E.) of (5).

A.E. has real and distinct roots:

If m1 , m2 , m3 , ..., mn be real and distinct roots then y1 = em1 x , y2 = em2 x , . . . , yn = emn x are L.I
solutions of (5), the C.F (or General Sol. ) is

y = c1 em1 x + c2 em2 x + c3 em3 x + ... + cn emn x ,

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

(Module-I (MA107-Math-II, by R. Singh) 3

Example 0.1. Solve (D3 + 6D2 + 11D + 6)y = 0.

The A.E. is m3 + 6m2 + 11m + 6 = 0. The roots are m = −1, −2, −3. Hence the required solution

is y = c1 e−x + c2 e−2x + c3 e−3x .

A.E. has repeated real roots:

Let m1 = m2 be repeated roots of the A.E. Then, we have n − 1 L.I solutions. It can be shown that
a simple choice y = xem1 x is also a solution that is independent of the rest n − 1 solutions. Thus,
the general solution of (5) is given by

y = (c1 + c2 x)em1 x + c3 em3 x + ... + cn emn x

The above idea can be further extended by taking solutions y1 = xem1 x , y2 = x2 em1 x , . . . , yl =

xl−1 em1 x . . . if the root m1 is repeating l−times.

Example 0.2. Solve (D4 + 2D3 − 3D2 − 4D + 4)y = 0.

The A.E. of the given equation is (m4 + 2m3 − 3m2 − 4m + 4) = 0. The roots of the A.E. are

m = 1, 1, −2, −2. The required solution is y = (c1 + c2 x)ex + (c3 + c4 x)e−2x .

Example 0.3. Solve (D3 + 3D2 + 3D + 1)y = 0.

The A.E. and its root are (m + 1)3 = 0 and m = −1, −1, −1. Therefore, the required solution is
y = (c1 + c2 x + c3 x2 )e−x .

A.E. has complex roots

If m1 = α + iβ and m2 = α − iβ, then the solutions em1 x , em2 x , . . . , emn x are L.I. and the general
solution is given by

y = c′1 em1 x + c′2 em2 x +c3 em3 x + ... + cn emn x .

| {z }

The above solution can be simplified as

y = c′1 eαx (cos βx + i sin βx) + c′2 eαx (cos βx − i sin βx) +c3 em3 x + ... + cn emn x .
| {z }
(Module-I (MA107-Math-II, by R. Singh) 4

Defining new constants c1 = c′1 + c′2 and c2 = i(c1 − c2 ), the general solution becomes

y = eαx (c1 cos βx + c2 sin βx) +c3 em3 x + ... + cn emn x .

| {z }

Similar to case II, the solution for repeated complex roots can be found; see the example below.

Example 0.4. Solve (D2 + 1)2 y = 0.

The A.E. and its roots are (m2 + 1)2 = 0, m = ±i, ±i. This is the case of repeated complex root, so
case II and case III can be combined to give the desired solution as

y = (c1 + c2 x) cos x + (c3 + c4 x) sin x

Example 0.5. Solve (D3 − 8)y = 0.

√
The A.E. of the given equation is (m3 − 8) = 0. Its root are m = 2, −1 ± i 3. The required solution
is
√ √
y = c1 e2x + e−x (c2 cos 3x + c3 sin 3x) .
| {z }

Example 0.6. Solve (D2 − 2D + 5)2 y = 0.

The auxiliary equation is (m2 − 2m + 5)2 = 0. Its roots are m = 1 ± 2i, 1 ± 2i Hence the required
solution is y = ex [(c1 + c2 x) cos 2x + (c3 + c4 x) sin 2x] .

Example 0.7. Solve (D2 + D + 1)2 (D − 2)y = 0.

√ √
1 3 1 3
The A.E. is (m + m + 1) (m − 2) = 0. Its roots are m = − ± i
2 2
,− ±i , 2. Hence, the desired
2 2 2 2
solution is [ √ √ ]
3 3
y = c1 e2x + e− 2 x (c2 + c3 x) cos
1
x + (c4 + c5 x) sin x
2 2

Example 0.8. Solve (D2 + 1)3 (D2 + D + 1)2 y = 0.

√ √
1 3 1 3
The A.E. is (m + 1) (m + m + 1) = 0. The roots are m = ±i, ±i, ±i, − ± i
2 3 2 2
,− ± i .
2 2 2 2
Therefore, the desired solution is
[ √ √ ]
− 12 x 3 3
y(x) = (c1 + c2 x + c3 x2 ) cos x + (c4 + c5 x + c6 x2 ) sin x +e (c7 + c8 x) cos x + (c9 + c10 x) sin x
| {z } 2 2
(Module-I (MA107-Math-II, by R. Singh) 5

Linear Combination (L.C.) of Functions

If f1 (x), f2 (x), ..., fn (x) are any n functions, then an L.C. of these functions is

f (x) = c1 f1 (x) + c2 f2 (x) + ... + cn fn (x),

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

Linear Independence and Dependence

The functions f1 (x), f2 (x), ..., fn (x) defined on I are said to be L.I. if the equation

c1 f1 (x) + c2 f2 (x) + ... + cn fn (x) = 0, x∈I

in n unknowns c1 , c2 , ..., cn has only the trivial solution c1 = c2 = ... = cn = 0.

Otherwise, these functions are said to be linear dependence (L.D.) on I.

Example 0.9. Show that the functions f1 = x, f2 = x2 , f3 = x3 are L.I. on any interval I.

A very elegant test of the L.I. and L.D. of the given function is an application of Wronskians.

Wronskian

The Wronskian of f1 (x), f2 (x), ..., fn (x) functions are defined as

f1 f2 ... fn
f1′ f2′ ... fn′
W (x) = W (f1 , f2 , ..., fn ) = .. .. .. .
..
. . . .
(n−1) (n−1)
f1 f2 . . . fn(n−1)

The Wronskian of the n functions exists if all the functions f1 , f2 , ..., fn are differential n − 1 times
on I. If any one or more functions are not differential, the Wronskian does not exist.
(Module-I (MA107-Math-II, by R. Singh) 6

Theorem 1. Let y1 , y2 , ..., yn are n solutions of

L(y) = y (n) + a1 (x)y (n−1) + a2 (x)y (n−2) + . . . + an (x)y = 0, (8)

where a1 , a2 , . . . , an are continuous on I. Then

(i) y1 , y2 , ..., yn are L.I. on I ⇔ W (x) = W (y1 , y2 , ..., yn ) ̸= 0 ∀ x ∈ I.
(ii) W (x0 ) = 0 for fixed x0 ∈ I ⇒ W (x) = 0, ∀ x ∈ I and y1 , y2 , ..., yn are L.D.

Theorem 2. (Abel’s formula) Let y1 , y2 , ..., yn are n solutions of

L(y) = y (n) + a1 (x)y (n−1) + a2 (x)y (n−2) + . . . + an (x)y = 0, (9)

( ∫ x )
on I and let x0 be any point in I. Then W (y1 , y2 , ..., yn ) = W (x0 ) exp − a1 (t)dt .
x0

Example 0.10. Show that y1 = e−x cos(x) and y2 = e−x sin(x) are L.I. solutions of Ly = y ′′ +2y ′ +2y =
0.

Solution: Clearly L[y1 ] = 0 and L[y2 ] = 0. Verify? To show y1 , y2 are L.I., we must show that

W (x) = W (y1 , y2 ) ̸= 0, ∀ x. Abel’s formula, it suffices to show that W (x0 ) ̸= 0 for some x0 ∈ R.
Taking x0 = 0, we see that
1 0
W (0) = = −1 ̸= 0.
−1 1

Example 0.11. Show that y1 = x and y1 = x−1 are L.I. solutions of Ly = x2 y ′′ + xy ′ − y = 0, x > 0.

Solution: Clearly L[y1 ] = 0 and L[y2 ] = 0. Verify? To show y1 , y2 are L.I., we must show that
W (x) = W (y1 , y2 ) ̸= 0, ∀ x,

x x−1
W (x) = = −2x−1 ̸= 0.
−2
1 −x
(Module-I (MA107-Math-II, by R. Singh) 7

Method of Reduction

If y1 ̸= 0 is known solution of

L(y) = y ′′ + a1 (x)y ′ + a2 (x)y = 0 (10)

Then, we can find a second solution as y2 = u(x)y1 (x).

∫ ∫
e− a1 dx
y2 = uy1 = y1 dx
y2
| {z1 }
∫
It can be verified that the Wronskian of y1 , y2 is equal to W (y1 , y2 ) = e− a1 dx
̸= 0.
1
Example 0.12. LDE x2 y ′′ + 4xy ′ + 2y = 0, x > 0, it’s one solution is y1 = . Find the second
x
solution, y2 .

u 4 1 ∫
Solution: The second solution y2 is given by y2 = uy1 = , a1 = , v = 2 e− a1 dx =
∫ ∫ x x y1
∫ 1 1 1 1 1 1
2 − x4 dx
xe = 2 and u = v(x)dx = 2
dx = − . Hence y2 = uy1 = (− ) = − 2
x x x x x x

Determination of Particular Integral (P.I.)

Consider

(Dn + a1 Dn−1 + a2 Dn−2 + ... + an I)y = f (D)y = b(x) (11)

Operating 1/f (D) on both the sides of (11), we obtain a P.I

1 1
yp = b(x) = n n−1
b(x) (12)
f (D) D + a1 D + a2 Dn−2 + ... + an

If the operator f (D) can be expressed as

f (D) = (D − α1 )(D − α2 ) . . . (D − αn )
(Module-I (MA107-Math-II, by R. Singh) 8

Theorem 3. If b(x) is a function of x and α is a constant, then

∫
1
yp = b(x) = eαx b(x)e−αx dx.
D−α

Rule I: b(x) = eαx

1 αx 1 αx
(i) If f (α) ̸= 0, yp = e = e ,
f (D) f (α)

(ii) If f (r−1) (α) = 0, f (r) (α) ̸= 0, then f (D) = (D − α)r g(D) where g(α) ̸= 0. Then P.I

1 αx xr αx
yp = e = e
(D − α)r r!

Example 0.13. Find the general solution of (D2 − 3D + 2)y = e3x .

Solution: The A.E. is m2 − 3m + 2 = 0 ⇒ (m − 1)(m − 2) = 0 ⇒ m = 1, 2. The

C.F. is given as

C.F. = c1 ex + c2 e2x

The particular integral is

1 1 1
yp = e3x = 2 e3x = e3x .
D2 − 3D + 2 3 − 3.3 + 2 2

1
The general solution is y = c1 ex + c2 e2x + e3x .
| {z } 2
yc |{z}
yp

Example 0.14. Solve (4D2 − 12D + 9)y = 144e3x/2

Solution: The A.E. is 4m2 − 12m + 9 = 0 ⇒ m = 3/2, 3/2. The complementary function
is
yc = (c1 + c2 x)e3x/2
(Module-I (MA107-Math-II, by R. Singh) 9

144 3x/2 144 1 3x/2 x2 3x/2

The particular integral is yp = e = e = 36 e . The required
(2D − 3)2 4 (D − 3/2)2 2!
solution is
x2
y = (c1 + c2 x)e3x/2 + 36 e3x/2 .
| {z } | 2!{z }

Rule II: b(x) = cos αx or b(x) = sin αx

If ϕ(−α2 ) ̸= 0,

1 1 1 1
yp = 2
sin(αx) = 2
sin(αx), yp = 2
cos(αx) = cos(αx),
ϕ(D ) ϕ(−α ) ϕ(D ) ϕ(−α2 )
| {z } | {z }

If ϕ(−α2 ) = 0

( ) ( )
1 1 iαx 1 iαx
sin(αx) = Im e = Im e provided f (iα) ̸= 0
f (D) f (D) f (iα)

Similarly,
( )
1 1 iαx
cos(αx) = Re e provided f (iα) ̸= 0
f (D) f (iα)

Example 0.15. Solve (D2 + 1)y = cos 2x.

Solution: The C.E. (A.E.) is given by m2 + 1 = 0 ⇒ m = ±i.

yc = c1 cos x + c2 sin x.

The particular integral is given by

1 1 1
yp = cos 2x = cos 2x = − cos 2x.
D2 +1 2
(−2 + 1) 3

The required solution is:

1
y = (c1 cos x + c2 sin x) − cos 2x .
| {z } | 3 {z }
(Module-I (MA107-Math-II, by R. Singh) 10

Example 0.16. Solve the differential equation (D2 − 4D + 3)y = sin x.

Solution: The roots of the A.E. m2 − 4m + 3 = 0 are m = 1, 3. The C.F. is

yc = c1 ex + c2 e3x

The particular integral is

1
yp = sin x
D2 − 4D + 3

Replacing D2 by −1, we get

1 1 1 1 1 + 2D
yp = sin x = sin x = sin x
2 − 4D 2 1 − 2D 2 1 − 4D2

Again, by replacing D2 with −1, we obtain

1 1
yp = (1 + 2D) sin x = (sin x + 2 cos x)
10 10

Hence the complete solution is

1
y = c1 ex + c2 e3x + (sin x + 2 cos x),
10

where c1 and c2 are arbitrary constants.

Rule III: b(x) is a polynomial of degree l

Take out the lowest degree term from f (D), to reduce it in the form

[1 ± f (D)]n .

Take it to the numerator, i.e., [1 ± f (D)]−n , and expand it in ascending powers of D with the help
of Binomial series:

α(α − 1) 2 α(α − 1)(α − 2) 3

(1 + x)α = 1 + αx + x + x + ...
2! 3!
(Module-I (MA107-Math-II, by R. Singh) 11

Note that in the expansion, we do not need to consider terms with power more than l since (l + 1)th
and higher order derivatives of the polynomial of degree l will be zero.

Example 0.17. Solve (D2 + D)y = x2 + 2x + 4

Solution: The A.E. is m2 + m = 0 ⇒ m = 0, −1. The complementary function is

yc = c1 + c2 e−x .

The particular integral is

1 1 1
yp = [x2 + 2x + 4] = [x2 + 2x + 4]
D2 +D D (1 + D)

Taking 1 + D into the numerator and expanding this into an infinite series, we get

1 1
yp = (1 − D + D2 − D3 + ...)(x2 + 2x + 4) = (x2 + 2x + 4 − 2x − 2 + 2)
D D

1
Operating on each term, we obtain
D

1 2 x3
yp = (x + 4) = + 4x
D 3

The desired general solution is

x3
y = c1 + c2 e−x + + 4x.
3

Rule IV: b(x) = eαx V , where V is any function of x

Using the shift property of the operator, we can easily prove that

1 αx 1
e V = eαx V.
f (D) f (D + α)

Example 0.18. Solve (D2 − 2D + 1)y = x2 ex .

(Module-I (MA107-Math-II, by R. Singh) 12

Solution: The characteristic equation and its roots are m2 − 2m + 1 = 0, and m = 1, 1. Thus,
the complementary function is
yc = (c1 + c2 x)ex

The particular integral is

1 1
yp = x2 ex = x2 ex
D2 − 2D + 1 (D − 1)2

Using shift property, we get

( ) ( )
x 1 2 x 1 2 x 1 1 2 x1 x3 x4
yp = e x = e x = e x =e = ex .
(D + 1 − 1)2 D2 D D D 3 12

x4
The required solution is y = (c1 + c2 x)ex + ex .
12

Rule V: b(x) = xV , where V is any function of x

1 1 f ′ (D)
yp = (xV ) = x V − V
f (D) f (D) [f (D)]2

Example 0.19. Solve f (D)y = (D2 + 9)y = x sin x.

Solution: The roots of the characteristic equations are ±3i. Hence, the complementary function
is
yc = c1 cos 3x + c2 sin 3x

The particular integral is

1
yp = x sin x
D2 +9

Using Rule V, we get

1 1 f ′ (D)
yp = (xV ) = x V − V
f (D) f (D) f (D)2
1 2D
yp = x sin x − sin x
D2 +9 (D + 9)2
2
(Module-I (MA107-Math-II, by R. Singh) 13

This can now be evaluated as

1 2D 1 2D 1 1
yp =x sin x − sin x = x sin x − sin x = x sin x − cos x
8 (D2 + 9)2 8 64 8 32

The required general solution is

1 1
y = c1 cos 3x + c2 sin 3x + x sin x − cos x
8 32

Rule VI: b(x) = xm sin αx or xm cos αx

In this case, Rule IV or Rule V can be applied.

1 [ 1 ]
1. yp = xm sin αx = Imag. Part xm eiαx
f (D) f (D)
1 [ 1 ]
2. yp = xm cos αx = Re. Part xm eiαx .
f (D) f (D)

Example 0.20. Find a particular integral of (D2 + 1)y = x2 sin 2x.

Solution: The particular integral is

1 1
yp = x2 sin 2x = Im 2 x2 e2ix
D2 +1 D +1

Applying Rule IV, we get the particular integral as

( ) ( )
2ix 1 2 2ix 1 2
yp = Im e x = Im e x
(D + 2i)2 + 1 D2 + 4Di − 3
( [ ( )]−1 )
e2ix 4iD D2
= Im 1− + x2
−3 3 3
(Module-I (MA107-Math-II, by R. Singh) 14

Using the Binomial expansion, we get

( [ ( ) ( )2 ] )
e2ix 4iD D2 4iD D2
= Im 1+ + + + ... x2
−3 3 3 3 3
( 2ix [ ] )
e 4iD D2 16D2
= Im 1+ + − + ... x 2
−3 3 3 9
( [ ])
1 8ix 26
= Im − (cos 2x + i sin 2x) x − 2
−
3 3 9

Collecting the imaginary part we have

[( ) ]
1 26 8
yp = − x −
2
sin 2x + x cos 2x .
3 9 3

Method of Variation of Parameters

We learn another general method, called the method of variation of parameters, of finding the
particular integral of the non-homogeneous differential equation. Consider a second-order LDE of

the form

L(y) = y ′′ + a1 (x)y ′ + a2 (x)y = b(x) (13)

where a1 (x), a2 (x), b(x) are functions of x or constants.

If y1 (x) and y2 (x) are two L.I. solutions of L(y) = 0. Then, its general solution is

yc (x) = c1 y1 (x) + c2 y1 (x) (14)

The method of variation of parameters relies on finding a particular solution

yp (x) = y1 (x) u1 (x) +y2 (x) u2 (x) (15)

| {z } | {z }
(Module-I (MA107-Math-II, by R. Singh) 15

of (13), where the unknowns u1 (x) and u2 (x) are computed where

∫ ∫
y2 b(x) y1 b(x)
u1 (x) = − dx, u2 (x) = dx, (16)
W W

y1 y2
W = = y1 y2′ − y1′ y2 ̸= 0 is Wronskian of y1 and y2 . Hence, the P.I. is given by
y1′ y2′

( ∫ ) (∫ )
y2 b(x) y1 b(x)
yp (x) = y1 (x) − dx +y2 (x) dx .
W W
| {z } | {z }
u1 u2

Hence, the general solution of the given differential equation is

y = yc + yp

Example 0.21. Solve the differential equation y ′′ + n2 y = sec nx.

Solution: Comparing with y ′′ + a1 (x)y ′ + a2 (x)y = b(x), we have a1 (x) = 0, a2 (x) = n2 , and

b(x) = sec nx. The A.E. of the corresponding homogeneous equation is: m2 + n2 = 0, so that
m = ±in. The C.F. is

yc = c1 cos nx + c2 sin nx.

Here, we have y1 = cos nx and y2 = sin nx its Wronskian is

cos nx sin nx
W (y1 , y2 ) = = n ̸= 0.
−n sin nx n cos nx

Then, the particular integral of the given equation is

yp = y1 u1 + y2 u2

where
∫ ∫
y2 b(x) sin nx sec nx 1
u1 = − dx = − dx = 2 ln(cos nx)
W n n

and
∫ ∫
y1 b(x) cos nx sec nx x
u2 = dx = dx =
W n n
(Module-I (MA107-Math-II, by R. Singh) 16

Hence, the required solution is

1 x
y = c1 cos nx + c2 sin nx + cos nx 2 ln(cos nx) + sin nx
| {z } n n}
yc | {z
yp

Example 0.22. Find the general solution of the differential equation y ′′ + n2 y = tan nx.

Solution: Here, we have a1 (x) = 0, a2 (x) = n2 , and b(x) = tan nx. The complementary function is
yc = c1 cos nx + c2 sin nx.
Here, we have y1 = cos nx, y2 = sin nx its Wronskian is given

cos nx sin nx
W = = n ̸= 0.
−n sin nx n cos nx

The particular integral is

yp = y1 u1 + y2 u2

where
∫ ∫
y2 b(x) sin nx tan nx 1
u1 = − dx = − dx = 2 [sin nx − ln(sec nx + tan nx)]
W n n

and
∫ ∫
y1 b(x) cos nx tan nx 1
u2 = dx = dx = − 2 cos nx
W n n

The desired general solution is

( )
[sin nx − ln(sec nx + tan nx)] cos nx
y = c1 cos nx + c2 sin nx + cos nx + sin nx −
| {z } n2 n2
| {z }

Example 0.23. Solve the differential equation y ′′ + n2 y = cot nx

Solution: The complementary function is: yc = c1 cos nx + c2 sin nx. In this case, we have y1 =
cos nx, y2 = sin nx, and b(x) = cot nx, and its Wronskian is given

cos nx sin nx
W = = n ̸= 0
−n sin nx n cos nx

Then, the particular integral is

yp = y1 u1 + y2 u2
(Module-I (MA107-Math-II, by R. Singh) 17

where
∫ ∫
y2 b(x) sin nx cot nx 1
u1 = − dx = − dx = − 2 sin nx
W n n

and
∫ ∫
y1 b(x) cos nx cot nx 1 [ ( nx )]
u2 = dx = dx = 2 cos nx + ln tan
W n n 2

The required solution is

( )
sin nx 1 [ ( nx )]
y = c1 cos nx + c2 sin nx + cos nx − + sin nx 2 cos nx + ln tan
| {z } n2 n 2
| {z }

Example 0.24. Using the method of variation of parameters solve y ′′ + y = sec2 x

Solution: The C.F. is yc = c1 cos x + c2 sin x. In this case, two L.I. solutions are: y1 = cos x, y2 =

sin x, and b(x) = sec2 x and

cos x sin x
W = = 1 ̸= 0.
− sin x cos x

Then, the particular integral is

yp = y1 u1 + y2 u2

where
∫ ∫ ∫
y2 b(x)
u1 = − dx = − sin x sec xdx = −
2
sec x tan xdx = − sec x
W

and
∫ ∫ ∫
y1 b(x) 2
u2 = dx = cos x sec xdx = sec xdx = ln(sec x + tan x)
W

The required solution is

y = c1 cos x + c2 sin x + cos x (− sec x) + sin x ln(sec x + tan x) .

| {z } | {z }

Method of Variation of Parameters for higher order

Consider the LDE of order n is given by

L(y) = y (n) + a1 (x)y (n−1) + a2 (x)y (n−2) + . . . + an (x)y = b(x), (17)

(Module-I (MA107-Math-II, by R. Singh) 18

If y1 , y2 , ..., yn are n L.I. solutions of the corresponding homogeneous equation. The general solution
of a homogeneous equation is

yc (x) = c1 y1 + c2 y2 + ... + cn yn . (18)

The method of variation of parameters relies on finding a particular solution

yp = u1 y1 + u2 y2 + ... + un yn . (19)

and u1 , u2 , ..., un to be computed as

∫
b(t)Wm (x)
um = dx, m = 1, 2, ..., n. (20)
W (x)

Here, W (x) = W (y1 , y2 , ..., yn ) is Wronskian; and Wm is the determinant obtained from W by

replacing the mth column with the column (0, 0, ..., 0, 1). Hence, the P.I.

∑
n ∫
b(t)Wm (t)
yp = ym (x) dt (21)
m=1
W (t)

Example 0.25. Using the method of variation of parameters solve: y ′′′ − 6y ′′ + 11y ′ − 6y = e−x

Solution: The C.F. is given by

yc = c1 ex + c2 e2x + c3 e3x

Also we have y1 = ex , y2 = e2x , y3 = e3x and b = e−x and

ex e2x e3x 1 1 1
W = ex 2e2x 3e3x = e6x 1 2 3 = 2e6x ̸= 0.

ex 4e2x 9e3x 1 4 9

0 e2x e3x ex 0 e3x ex e2x 0

W1 = 0 2e2x 3e3x = e5x , W2 = ex 0 3e3x = −2e4x , W3 = ex 2e2x 0 = e3x

1 4e2x 9e3x ex 1 9e3x ex 4e2x 1

(Module-I (MA107-Math-II, by R. Singh) 19

∫ ∫ ∫
e−x .e5x e−2x e−x (−2e4x ) e−3x e−x e3x e−4x
u1 = dx = − ; u2 = dx = ; u3 = dx = −
2e6x 4 2e6x 3 2e6x 8

The required solution is

( ) ( ) ( )
e−2x e−3x e−4x
x 2x 3x
y = c1 e + c2 e + c3 e + e x
− +e 2x
+e 3x
− .
| {z } 4 3 8
| {z }

Cauchy-Euler LD Equation

An LD equation of the form

(a0 xn Dn + a1 xn−1 Dn−1 + a2 xn−2 Dn−2 + ... + an )y = b(x). (22)

where a0 ̸= 0, a1 , a2 , . . . , an are constants, is called Cauchy-Euler LDE.

We transform the above equation into LDE with constant coefficients by substituting

dz 1
x = ez , or ln x = z, so that = . (23)
dx x

Using the chain rule for differentiation, we obtain

dy dy dz 1 dy
= =
dx dz dx x dz

d
Defining =: D1 , we have
dz

dy dy
x = ⇔ xDy = D1 y
dx dz

Similarly, for the second-order derivative

( ) ( ) ( ) ( )
d2 y d dy d 1 dy 1 dy 1 d dy 1 dy 1 d dy dz
2
= = =− 2 + =− 2 +
dx dx dx dx x dz x dz x dx dz x dz x dz dz dx
1 dy 1 d2 y
=− 2 + 2 2
x dz x dz
(Module-I (MA107-Math-II, by R. Singh) 20

Thus, we have

2
2d y d2 y dy
x = 2− ⇒ x2 D2 y = D1 (D1 − 1)y.
dx2 dz dz

Similarly, x3 D3 y = D1 (D1 − 1)(D1 − 2)y and so on.

In general, we have

xn Dn = D1 (D1 − 1)(D1 − 2) . . . (D1 − n + 1)y

Substituting the above values of x, xD, x2 D2 , . . . , xn Dn into (22), we get

[a0 D1 (D1 − 1) . . . (D1 − n + 1) + . . . + an−2 D1 (D1 − 1) + an−1 D1 + an ] y = b (ez ) (24)

Equation (24) is an L.D. equation with constant coefficients, which can be solved with the methods

discussed. Finally, we obtain the desired solution by replacing z with ln x.

Example 0.26. Solve (x2 D2 + xD − 4)y = 0.

Solution: Substituting x = ez ⇒ ln x = z and ⇒ xD = D1 , x2 D2 = D1 (D1 − 1), the given

equation reduces to

[D1 (D1 − 1) + D1 − 4] y = 0 ⇒ (D12 − 4)y = 0

The roots of the corresponding C.E. are m = 2, −2. The required solution of the transformed
equation is
y(z) = c1 e2z + c2 e−2z

Putting z = ln x, we have the desired solution as y = c1 x2 + c2 x−2 .

Example 0.27. Find the general solution of the differential equation (x2 D2 + 1)y = 3x2 .

Solution: Substituting x = ez , the given equation reduces to

(D1 (D1 − 1) + 1)y = 3e2z ⇒ (D12 − D1 + 1)y = 3e2z

(Module-I (MA107-Math-II, by R. Singh) 21

The characteristic equation of this differential equation is

√
m2 − m + 1 = 0 ⇒ m = (1 ± i 3)/2

The complementary function is

[ ( √ ) ( √ )]
yc = ez/2 c1 cos z 3/2 + c1 sin z 3/2

Substituting z = ln x, we get

√ [ ( √ ) ( √ )]
yc = x c1 cos ln x 3/2 + c1 sin ln x 3/2

The particular integral of the transformed equation is

1 1
yp = 3e2z = 2 3e2z = e2z
D12 − D1 + 1 2 −2+1

Hence, the desired solution of the given differential equation is

√ [ ( √ ) ( √ )]
y= x c1 cos ln x 3/2 + c1 sin ln x 3/2 + x2

Legendre’s Linear Differential Equations

A linear differential equation of the form is

[(a + bx)n a0 Dn + a1 (a + bx)n−1 Dn−1 + a2 (a + bx)n−2 Dn−2 + ... + an ]y = b(x), (25)

where a, b, a1 , a2 , ..., an are constants, is called Legendre’s LDE.

Put

dz b
a + bx = ez , or ln(a + bx) = z, so that = . (26)
dx a + bx
(Module-I (MA107-Math-II, by R. Singh) 22

Compute first-order derivative

dy dy dz b dy
= =
dx dz dx (a + bx) dz

This implies
dy dy
(a + bx) =b ⇔ (a + bx)Dy = bD1 y
dx dz

Second order derivative, we get

( ) ( )
d2 y d dy d b dy
2
= =
dx dx dx dx (a + bx) dz

This can be further simplified to get

( )
d2 y b2 dy b d dy
=− +
dx2 (a + bx)2 dz (a + bx) dx dz
( )
b2 dy b d dy dz
=− +
(a + bx)2 dz (a + bx) dz dz dx

Substituting dz/dx, we obtain

d2 y b2 dy b2 d2 y
= − +
dx2 (a + bx)2 dz (a + bx)2 dz 2

This gives us

( )
d2 y d2 y dy
(a + bx)2 2 =b 2
− ⇔ (a + bx)2 D2 y = b2 D1 (D1 − 1)y
dx dz 2 dz

In general, we have

(a + bx)n Dn = bn D1 (D1 − 1)(D1 − 2) . . . (D1 − n + 1)y

Substituting the above values of (a + bx), (a + bx)D, (a + bx)2 D2 , . . . , (a + bx)n Dn in (25), we get

( )
( ) ez − a
a0 b D1 (D1 − 1)...(D1 − n + 1) + ... + an−2 b D1 (D1 − 1) + an−1 bD1 + an y = b
n 2
b

Example 0.28. Solve the differential equation

d3 y 2
3d y dy 1
(1 + x)4 3
+ 2(1 + x) 2
− (1 + x)2 + (1 + x)y =
dx dx dx (1 + x)
(Module-I (MA107-Math-II, by R. Singh) 23

d
Solution: Using D = and dividing both sides by (x + 1), the given DE can be rewritten as
dx

[ ]
(1 + x)3 D3 + 2(1 + x)2 D2 − (1 + x)D + 1 y = (1 + x)−2 .

This is Legendre’s linear equation which can be solved by substituting

1 + x = ez ⇒ z = ln(1 + x)

This substitution readily implies

(1 + x)D = D1 , (1 + x)2 D2 = D1 (D1 − 1), (1 + x)3 D3 = D1 (D1 − 1)(D1 − 2)

The given differential equation reduces to

[D1 (D1 − 1)(D1 − 2) + 2D1 (D1 − 1) − D1 + 1] y = e−2z

or
( )
D13 − D12 − D1 + 1 y = e−2z

The A.E. of the corresponding homogeneous equation is

m3 − m2 − m + 1 = 0

and its roots are m = 1, 1, −1. Hence the C.F. of the transformed differential equation is

yc = (c1 + c2 z)ez + c3 e−z

The particular integral of the transformed differential equation can be found as

1 1 1
yp = e−2z = e−2z = − e−2z
(D13 − D12 − D1 + 1) −2 − 2 + 2 + 1
3 2 9
(Module-I (MA107-Math-II, by R. Singh) 24

Hence the general solution of the transformed differential equation is

1
y(z) = [c1 + c2 z]ez + c3 e−z − e−2z
9

Replacing z by ln(1 + x), we obtain the desired solution of the given differential equation

c3 1 1
y(x) = (c1 + c2 ln(1 + x)) (1 + x) + − .
(1 + x) 9 (1 + x)2