Download as pdf or txt
Download as pdf or txt
You are on page 1of 209

B. A.

(Programme) Semester-VI Mathematics

DISCIPLINE SPECIFIC ELECTIVE (DSE)


Option-2: Differential Equations
Study Material : Unit 1-3

SCHOOL OF OPEN LEARNING


University of Delhi

Department of Mathematics
Editor: Prof. Suman Kumar Verma
Graduate Course

Discipline Specific Elective Course (DSE)


Option-2: Differential Equations
Study Material : Unit 1-3
Contents
Unit-1 : Ordinary Differential Equations
Lesson-1 : Introduction and Definition
Lesson-2 : Linear and Exact Differential Equation
Lesson-3 : Linear Differential Equations with Constant Coefficients-I
Lesson-4 : Linear Differential Equations with Constant Coefficients
(Non-Homogeneous)-I
Lesson-5 : Linear Differential Equations with Constant Coefficients
(Non-Homogeneous)-II
Lesson-6 : Linear Differential Equations – Definitions Fundamental theorem
Lesson-7 : Linear Differential Equation of Second order with variable Coefficients
Lesson-8 : Simultaneous Differential Equation
Lesson-9 : Total Differential Equations

Unit-2 : Linear Partial Differential Equation


Lesson-10 : Partial Differential Equation
Unit-3 : Non-Linear Partial Differential Equation
Lesson-11 : Non-Linear Partial Differential Equation-I
Lesson-12 : Non-Linear Partial Differential Equation-II

Edited by:
Prof. Suman Kumar Verma

SCHOOL OF OPEN LEARNING


UNIVERSITY OF DELHI
5, Cavalry Lane, Delhi-110007
Lesson 1
INTRODUCTION AND DEFINITIONS

1. Introduction
An equation which involves differentials or derivatives is called a differential equation.
For example
(y2 – x2)dx – 2xy dy = 0 ...(1)
2
 dy  dy
 dx  xy + dx (4 x – 3 y ) – 12 xy = 0
2 2
...(2)
 
1/2
d 2 y   dy  
2

a 2 = 1 +    ...(3)
dx   dx  

d3y d2y dy
3
+ cos x 2 – 2sin x – y cos x = sin 2 x ...(4)
dx dx dx

d3y d2y
– 3 + 4y = 0 ...(5)
dx 3 dx 2

 dy d 2 y d 3 y dny 
f  x, y, , 2  3 =0 ...(6)
 dx dx dx dx n 

The examples from (1) to (6) are differential equations in contrast with other types of
equations, for example, algebratic equations or trigonometrical equations, which do not
contain differentials or differential coefficients.
A differential equation is said to be ordinary if all the differential coefficients have
reference to a single independent variable. The differential equations given in (2), (3), (4),
(5) and (6) involve only one independent variable x and therefore are ordinary differential
equations. The equation (1) in which either x or y may be taken as the independent
variable is also an illustration of an ordinary differential equation.
The order of a differential equation is the order of the highest derivative appearing in
it. The differential equation (1) and (2) are of the first order.
The differential equation (3) is of the 2nd order.
d3y
The differential equation (4) and (5) are of third order as is the highest order
dx 3
derivative occuring in the equations.
The differential equation (6) is of nth order.

1
Degree of differential equation: The degree of a differential equation is the degree of
the highest order derivative when the differential coefficients are free from radicals and
fractions. The differential equations (4) and (5) are of first degree and the differential
equations (2) and (3) are of the 2nd degree, since the degree of the highest derivative, i.e.,
dy
of in equation (2) is 2. The differential equation (3) when the differential coefficients
dx
2
 d2 y   dy 
are free from radicals, can be written as a2  2 
=1+  
 dx   dx 

d2y
so that the degree of the highest derivative is two
dx 2
 equation (3) is of second degree.

dy y
Again the equation y + 2x = can be written, when the differential coefficient
dx dy / dx
2
 dy  dy
are free from fractions, as y   + 2 x = y so that the degree of the differential
 dx  dx
dy
equation is two, as the highest degree of is two.
dx

2. Solution or Primitive of a Differential Equation


Any relation between the dependent and independent variables, which when
substituted in the differential equation reduces it to an indentity, is called a solution or a
primitive of the differential equation. Consider the differential equation
d2y
= –y ...(1)
d2x
The solution or primitive of the differential equation (1) is
y = A cos (x + ) ...(2)
where A and  are arbitrary constants.
For differentiating (2) with respect to x twice successively, we get,
dy
= – A sin( x + )
dx

d2y
= – A cos( x + ) ...(3)
dx 2

when the relations (2) and (3) are substituted in the differential equation (1), these reduce
it an identity.
A solution of a differential equation (d.e.) does not contain derivatives or
differentials.

2
3. Formation of Differential Equations
Consider a family of parabolas
y2 = ax where a is an abitrary constant ...(1)
Differentiating (1) w.r.t.x, we get
dy
2y =a ...(2)
dx

Eliminating ‘a’ from (1) and (2), we obtain the differential equation
dy
y 2 = 2 xy ...(3)
dx

which is an ordinary differential equation of the first order and first degree. Equation (1)
is called the general solution of the differential equation (3) and contains one arbitrary
constant a.
Take another family or curve represented by
y = A cos (x + ) ...(4)
where A and  are arbitrary constants.
Differentiating (4) twice successively w.r.t.x, we get
dy
= – A sin( x + ) ...(5)
dx

d2y
= – A cos( x + ) ...(6)
dx 2

Eliminating A and  from (4) and (6), we have ordinary differential equation of the 2nd
order viz.,
d2y
= –y ...(7)
dx 2
The relation (4) is the general solution or primitive of the d.e. (7) and it contains two
arbitrary constants A and . Thus we see that the general solution of an ordinary
differential equation of the first order contains one aribtary constant and that of the
equation of 2nd order contains two arbitrary constants.
In order to obtain a differential equation whose solution is
f(x, y, c1, c2 ......... cn) = 0 ...(1)
where c1...cn are arbitary constants, we have to eliminate the n constants for which require
(n + 1) equations. The given equation along with n more obtained by successively
differentiating it n times provides the required (n + 1) equations. The differential equation

3
thus obtained is clearly of the nth order. Its solution viz., (1) contains n arbitrary
constants.
The solution of differential equation which contains a number of arbitrary constants
equal to the order of the equation is called the general solution or the complete integral or
the complete primitive.
Particular Integral: Any solution derived from the complete primitive by giving
particular values of these constants is called a Particular Integral.
Example, the general solution of the differential equation
d2y
+y=0
dx 2

is given by
y = c1 cosx + c2 sinx
where c1, c2 are arbitrary constants. It can be seen that y = 2 cosx + 3 sinx is a particular
solution by giving values of 2 and 3 to c1 and c2 respectively.
y = cosx taking c1 = 1, c2 = 0
y = sinx taking c1 = 0, c2 = 1
are other particular solutions.
Graphical representation. We shall now give example of a method of sketching
rapidly the general form of the family of curve representing the complete primitive of
dy
= f ( x , y)
dx

where f(x, y) is a function of x and y having a perfectly definite finite value for every pair
of finite values of x and y.
The curves of the family are called the characteristics of the equation.
dy
Example : = x ( y – 1)
dx

d2y dy
Here 2
= y –1+ x = ( x 2 + 1)( y – 1).
dx dx

Now a curve has its concavity upwards when the second differential coefficient is
positive. Hence the characteristics will be concave up above y = 1, and concave down
dy
below this line. The maximum or minimum points lie on x = 0, since = 0 there. The
dx
characteristics near y = 1, which is a member of a family are flatter than those further
from it. See fig-1.

4
The most general solution of an ordinary differential equation of nth order must
contain n and only n independent arbitrary constants. The general solution may not
include all possible solutions. There may be a solution which cannot be derived from the
general solution by giving particular values to arbitrary constants. Such a solution is
called a singular solution.
Solved Example 1. By the elimination of the constants a and b obtain the differential
equation, of which
xy = aex + be–x + x2
is a solution.
Solution: The given equation is
xy = aex + be–x + x2 ...(1)
In equation (1) there are two arbitrary constants a and b. Therefore in order to form,
its differential equation, we differentiate it twice in succession w.r.t. x.
We have after the first differentiation
dy
x + y = ae x – be – x + 2 x ...(2)
dx

Differentiating (2), we get


d2y dy
x 2
+ 2 = ae x + be – x + 2
dx dx

d2y dy
or x 2
+ 2 – 2 = ae x + be – x ...(3)
dx dx

From the three equations (1), (2) and (3), two arbitrary constants a and b are to be
eliminated to form the differential equation.

5
From (3) and (1), we get
d2y dy
x 2
+ 2 – 2 = xy – x 2
dx dx

d2y dy
or x + 2 – xy = 2 – x 2 ...(4)
dx 2 dx
The equation (4) is the required differential equation whose solution is given by the
equation (1).
Solved Example 2. Determine the differential equation whose general solution is
given by
y = cx + c – c2 ...(1)
Solution : Since there is only arbitrary constant c in the general solution (1),
therefore, we have to differentiate (1) once w.r.t. x and obtain
dy
=c ...(2)
dx
dy
To eliminate c replace c by as obtained from (2) in equation (1), and we have the
dx
required differential equation of (1) in the form
3
dy dy  dy 
y=x + –  ...(3)
dx dx  dx 
dy ,
Let p= then (3) may be written as
dx

y = px + p – p3.
Solved Example 3. Show that
A
v= +B ...(1)
r

is a solution of the differential equation


d 2 v 2 dv
+  =0 ...(2)
dr 2 r dr
Solution: Differentiating (1) w.r.t.r. twice in succession, we obtain
dv A
=– 2 ...(3)
dr r

d 2v 2A
2
=– 3 ...(4)
dr r

6
Now from (3) and (4), we get
d 2 v 2 dv 2 A 2 A
+ = – 3 =0
dr 2 r dr r 3 r
It follows that (1) is the general solution of (2).
4. Methods of Solving Differential Equations
Though every d.e. is not solvable, there are certain standard forms of differential
equations which can be solved. The following methods are used to solve ordinary
differential equations of the first order and of the first degree:
(I) Method of separating the variables.
(II) Method of solving homogeneous equations in x and y.
(III) Method of solving non-homogeneous equations of the first degree in x and y.
(IV) Method of solving linear equations.
(V) Method of solving exact equations.
These standard methods will be discussed in the next few articles.
1. Variables Separable (Method of Separating Variables)
The general form of such a differential equation is
f(x)dx + (y)dy = 0 ...(1)
Notice that in this of d.e., of the first order, there is a group of terms containing x only
multiplied by dx added to another group of terms containing y only multiplied by dy. Its
solution which is immediately obtained by integration is

 f ( x )dx +  ( y)dy = c ...(2)

The solution (2) is the general solution of the d.e., of the type (1) with an arbitrary
constant c.
If the equation is not in the above form, some times one can see at a glance how to put
it in the form, or at it is commonly expressed, to separate the variables. See Ex. 4 below
Solved Example 4. Solve
sec2x tan y dx + sec2y tan x dy = 0 ...(1)
Solution: The d.e. (1) can be written by dividing L.H.S. and R.H.S. by tan x tan y in
the form
sec2 x sec2 y
dx + dy = 0 ...(2)
tan x tan y

which is of the form

7
f(x)dx + (y)dy = 0,
sec2 x , sec2 y ,
where f ( x) = ( y) =
tan x tan y

Integrating (2), we get the general solution of (1) as


sec2 x sec2 y
 tan x dx +  tan y dy = c = log A(say),
where A is an arbitrary constant of integration.
or log(tan x) + log(tan y) = log A
or log(tan x tan y) = log A
i.e. tan x tan y = A, is the required solution.
Solved Example 5. Solve the d.e.
(ex + 1)ydy = (y + 1)exdx ...(1)
Solution: Dividing throughout by (y + 1) (ex + 1), equation (1) reduces to
ydy e x dx
=
y + 1 ex + 1

e x dx ydy
i.e. – =0
e +1 y +1
x

e x dx ydy
Integrating, we get e x
+1
–
y +1
=c

e x dx ( y + 1) – 1
or e x
+1
–
y +1
dy = c

e x dx  1 
or  e x + 1 –  1 – y + 1  dy = c
or log (ex + 1) – y + log(y + 1) = log c (say)
or log (ex + 1) (y + 1) = log (cey) [ log ey = y log e = y]

i.e., (ex + 1) (x + 1) = cey


is the required solution.
Solved Example 6. Solve the differential equation
dy
= ex + y + x 2ey .
dx

8
Solution: The given differential equation can be written as
Separating the variable, we have e–y dy = (ex + x2)dx
Integrating, we get the general solution in the form
1 3
–e– y = e x + x +c
3

1 3
or e x + e– y + x + c = 0 is the required solution.
3

dy  dy 
Solved Example 7. Solve y – x = a  y2 + 
dx  dx 

Solution : We have
dy
y – ay 2 = ( x + a)
dx

dx dy
 =
x + a y(1 – ay)

Integrating both sides, we get


dx 1 a 
 x + a =   y + 1 – ay  dy + c
 log (x + a) = log y – log (1 – ay) + c
cy
 log( x + a) = log
1 – ay

 (x + a) (1 – ay) = cy is the required solution.


dy  dy 
Solved Example 7. Solve y – x = a  y2 + 
dx  dx 
Solution: We have
dy
y – ay 2 = ( x + a)
dx

dx dy
 =
x + a y(1 – ay)

Integrating both sides, we get


dx 1 a 
 x + a =   y + 1 – ay  dy + c
 log(x + a) = log y – log (1 – ay) + c

9
cy
 log(x + a) = log
1 – ay

 (x + a) (1 – ay) = cy is the required solution.


Solved Example 8. Solve
dy
1 + x 2 + y 2 + x 2 y 2 ) + xy =0
dx

Solution: Given differential equation can be written as


dy
(1 + x 2 ) + y 2 (1 + x 2 ) = – xy
dx

dy
 (1 + x 2 )(1 + y 2 ) = – xy
dx

Separating the variables, we get

1 + x2 y
dx = – dy
x 1 + y2

Integrating, we get the general solution in the form

1 + x2 y
 x
dx = –1
1 + y2
dy + c ...(1)

Let 1 + x2 = t2  2x dx = 2t dt
t
 dx = dt
x

Equation (1) can be written as


t2
x 2
dt = – 1 + y 2 + c

t2
  t 2 – 1dt = – 1 + y + c
2

t2 – 1 + 1
  t 2 – 1 dt = – 1 + y + c
2

 1 
1 + t 2 – 1  dt = – 1 + y + c
2

 

dt
 t+ = – 1 + y2 + c
(t – 1)(t + 1)

10
1 t –1
 t + log = – 1 + y2 + c
2 t +1

1 1 + x2 – 1
 1 + x 2 + log = – 1 + y2 + c
2 1+ x +1
2

1 1 + x2 – 1
 1 + x 2 + 1 + y 2 + log =c
2 1 + x2 + 1

is the required solution.


dy
Solved Example 9. Solve = cos( x + y)
dx
dz dy
Solution: Put x + y = z  =1+ .
dx dx

The given differential equation becomes.


dz
– 1 = cos z
dx

dz
 = 1 + cos z
dx

dz
 = dz (Variable z and x are separable)
1 + cos z

Integrating both sides, we get


dz 1 z
 2 cos 2
z/2
= x + c   sec 2 dz = x + c
2 2

1 z x+y
 2 tan = x + c  tan = x+c
2 2 2

II. Method of solving Homogeneous Equations


dy f ( x, y)
A differential equation of the form = where f(x, y) and (x, y) are
dx ( x, y)
homogenous functions x, y and of the same degree is said to be homogeneous. Such an
equation can be solved by putting y = vx, thereby changing the dependent variable from y
to v. The equation is then reduced to form in which the variable are separable and hence
can be solved.
Since f(x, y) and (x, y) are homogeneous functions of the same degree say n, we can
write
 y
f ( x, y) = x n f1   = x n f1 (v)
x

11
 y
and ( x, y) = x n f2   = x n f2 (v)
x
y
where v= i.e., y = vx
x

Again y = vx gives on differentiation


dy dv
=v+x .
dx dx

The given differential equation, therefore, becomes;


dv f1 (v)
v+x =
dx f2 (v)

dv f1 (v)
x = –v
dx f2 (v)

f2 (v)dv dx
or =
f1 (v) – vf2 (v) x

so that the variable v and x are now separable. The required solution is
f2 (v)dv dx
 f (v) – vf (v) = 
1 2 x
+c

where c is an arbitrary constant.


The method is illustrated in the following examples:
Solved Example 10. Solve
dy y2
=
dx xy – x 2

Solution: It is a homogeneous equation of the 2nd degree in x and y.


Differentiating (2) with respect to x, we get
dy dy
=x +v
dx dx

The given differential equation (1) is now reduced to


dy v2 x 2 v2
x +v= 2 =
dx x (v – 1) v – 1

dv v2 v
or x = –v=
dx (v – 1) v –1

Separating the variables, we have

12
v –1 dx
dv =
v x

Integrating it, we get the solution of (3) as


(v – 1) dx
 v
dv = 
x
+c

dv
or  dv –  v
= log x + c

or v – log v = log x + c
or v = log vx + c
y
or = log y + c [ y = vx ]
x

or y = x (log y + c).
Solved Example 11. Solve
y  y 
x sin dy =  y sin – x  dx
x  x 
Solution: We have
y
y sin –x
dy x
=
dx y
x sin
x
Let y = vx, we get
dv vx sin v – x
v+x =
dx x sin v

dv v sin v – 1
 v+x =
dx sin v

dv v sin v – 1
 x = –v
dx sin v

dv 1
 x =–
dx sin v

dx
 = – sin vdv. (This is a differential equation with separated variables.)
x

Integrating, we get
dx
 x
= –  sin vdv + c

13
 log x = cos v + c
y
 log x = cos + c is the required solution.
x

III. Method of solving non-homogeneous equation of the first degree in x and y.


This differential equation is of the form,
dy ax + by + c
= ...(1)
dx Ax + By + D

The numerator ax + by + c and the denominator Ax + By + D are not homogeneous


expressions in x and y but the d.e. can be reduced to a homogeneous differential equation
of first degree by putting
x = X + h
...(2)
y – Y + k 

where h and k are suitable constants to be found.


From (2), we have
dx = dX, dy = dY
dy dY
 =
dx dX

The equation (1) is then transformed to


dY aX + bY + (ah + bk + c)
= ...(3)
dX AX + BY + ( Ah + Bk + D)

Let h and k be so choosen that they satisfy the linear equations


ah + bk + c = 0
On solving, we get Ah + Bk + D = 0
bD – cB 
h=
aB – bA 
 ...(4)
cA – aD 
k=
aB – bA 
Assuming that
a b
aB – bA  0, i.e.,  ...(5)
A B
the differential equation (3) is reduced to
dY aX + bY
= ...(6)
dX AX + BY

14
The d.e. (6) is homogeneous equation of degree one in X an Y and is therefore
solvable by putting Y = VX.
Let the solution of the equation (6) be
F(X, Y, c1) = 0
c1 being the arbitrary constant of integration. By substituting the values of X and Y in
terms of x, y from (2), we have the solution of the differential equation (1) as
F(x – h, y – k, c1) = 0,
where h, k are given by relation (4).
Special Case: If the condition (5) is not satisfied
a b
i.e., if = =  (suppose)
A B

i.e., if a = A , b = B
then the differential equation (1) is reduced to
dy ( Ax + By) + c
= ...(7)
dx Ax + By + D

Here put Ax + By = u ...(8)


On differentiating (8) with respect to x, we get
dy du
A+B =
dx dx
 du 
 – A
dy  dx 
 =
dx B
Now the equation (7) is reduced with the help of the substitution (8) to form
1  du  u + c
 – A =
B  dx  u+D
du  u + c 
or = B + A
dx  u+D 
which is a differential equation with variables separable and hence can be solved.
Solved Example 12. Solve
dy 2 x – 3y – 5
=
dx 3x + 4 y + 1

Solution: The given equation is non-homogeneous. Putting


x = X + h, y = Y + k we have

15
dy dY
=
dx dX
Equation (1) reduces to
dY 2 X – 3Y + 2 h – 3k – 5
=
dX 3 X + 4Y + 3h + 4 k + 1
Let us choose h, k so that
2h – 3k – 5 = 0 
3h + 4k + 1 = 0 

Solving for h, k we get


h = 1, k=–1
The given equation becomes
dY 2 X – 3Y
=
dX 3 X + 4Y
which is homogeneous and of first degree in X, Y.
Putting Y = VX, we get
dV dY 2 – 3V
V+X = =
dX dX 3 + 4V

X
dV 2 – 3V
= –V =
2 – 6V – 4V 2
=
( )
dV 2 + 3V – 1
dX 3 + 4V 3 + 4V 3 + 4V

or
(3 + 4V ) dV2dX
+=0
2V + 3V – 1
2
X
Now the variables are separable.
Integrating, we get,
(4V + 3)dV dX
 2V 2
+ 3V – 1
+2 
X
= log c (say)

i.e. log (2V2 + 3V – 1) + log X2 = log c


 (2V2 + 3V – 1) X2 = c
 2V2X2 + 3VX2 – X2 = c
 2Y2 + 3XY – X2 = c ( Y = VX)
 2(y + 1)2 + 3(x – 1) (y + 1) – (x – 1)2 = c

16
 2(y2 + 2y + 1) + 3(xy – y + x – 1) – (x2 – 2x + 1) = c
 X = x – h = x – 1
 Y = y – k = y +1 
 

i.e., 2y2 – x2 + 3xy + 5x + y – 2 = c


is the solution.
Solved Example 9. Solve
(2x + y + 1) dx + (4x + 2y – 1) dy = 0.
Solution: Here
dy (2 x + y + 1)
= –
dx (4 x + 2 y – 1)

(2 x + y) + 1
= – ...(1)
(2 x + 2 y) – 1

In order to solve the differential equation (1), put


2x + y = u ...(2)
Differentiating (2), with respect to x, we get
dy du
2+ =
dx dx
dy du
or = –2
dx dx
By the substitution (2), the equation (1) is transformed to
du u +1 ,
–2=– ...(3)
dx 2u – 1
du u +1 3(u – 1)
or =– +2=
dx 2u – 1 (2u – 1)

Separating the variables, we get


2u – 1
du = dx
3(u – 1)

On integrating, we get the solution of the equation (3) as


(2u – 1)
 (u – 1) du = 3  dx + c
2(u – 1) + 1
or 3x + c =  du
u –1

17
du
= 2  du +  u –1
= 2u + log (u – 1)
By replacing u by 2x + y from (2), we get the general solution of the differential equation
(1) as
3x + c = 2 (2x + y) + log (2x + y – 1)
or x + 2y + log (2x + y – 1) = c.

Exercise 1
Solve the following differential equations
1. y(1 + x) dx + x (1 + y) dy = 0.
2. (x2 – y2)dy + (y2 + x2)dx = 0.
3. (xy2 + x)dx + (yx2 + y)dy = 0.
4. cosec x. log y dy + x2y2dx = 0
dy
5. (3x 2
–x ) dx – y+3=0

dy  dy 
6. y – x = 3 1 – x 2  .
dx  dx 

7. cos y log (sec x + tan x) dx = cos x log (sec y + tan y) dx.


dx
8. = cos( x + y).
dy

[Hint : Put x + y = z]
dx
9. ( x + y + 1) = 1.
dy

[Hint : Put x + y + 1 = z so as to change the dependent variable from y to z. By the


substitution the equation is reduced to a form where the variables, z and x are
sparable.]

Exercise 2
Solve the following differential equations:
[Hint : All the equations are homogeneous, put y = vx].
dy
1. xy = x 2 + y2 .
dx

18
dy
2. ( x 2 + y 2 ) = xy.
dx
3. xy2 dy = (x3 + y3)dx.
4. y2 dx + (xy + x2)dy = 0.

5. x 2 dy + y( x + y)dx + 0.

dy x 2 + 2 xy – y2
6. + = 0.
dx y2 – 2 xy + x 2

7. xdy – ydx = x 2 + y 2 dx.

Exercise 3
Solve the following differential equations:
dy x + 2 y – 1 .
1. =
dx 2 x + y + 1

2. (x + y) (dx – dy) = dx + dy.


dy x + y – 1 
[Hint : The equation can be written as =
dx x + y + 1 

3. (2x + y + 1)dx + (4x + 2y – 1)dy = 0.


[Hint : Put 2x + y = z]
4. (2x + 4y + 3)dy = (2y + x + 1)dx.
dy
5. = (4 x + y + 1)2 .
dx
[Hint : Put 4x + y + 1 = z and change the dependent variable from y to z]

19
Lesson 2
LINEAR AND EXACT DIFFERENTIAL EQUATIONS

IV. Method of Solving Linear Differential Equations


A differential equation is said to be linear when the dependent variable and its
derivatives appear only in the first degree and are not multiplied together.
The differential equation of the form
dy
+ Py = Q ...(1)
dx
dx
or + Px = Q ...(2)
dy

where in (1) P and Q are functions of x (or constants) only, and in (2) P and Q are
functions of y (or constants) only is called a linear differential equation of the first order.
By multiplying the differential equation (1) by a integrating factor R (a function of x)
the equation (1) becomes
dy
R + RPy = RQ ...(3)
dx
Let R be such that the L.H.S. of equation (3) viz.,
dy
R + PRy
dx
d dy dR
= [ Ry] = R + y
dx dx dx
dR
So that PRy = y
dx
dR
i.e., = PR
dx
Separating variables and integrating, we get
dR
 R 
= Pdx

or log R =  Pdx ...(A)

the constant of integration being taken to be zero as a particular solution of (A) is


required.

20
R = e
Pdx
 ...(4)
The differential equation (3) is now equivalent to
d
( Ry) = RQ
dx
where R, a function of x only is given by (4). Separating the variables we get
d(yR) = QR dx
and on integration we have general solution of equation (3) and therefore that of equation
(1) in the form

yR =  QR dx + A

where A is an arbitrary constant of integration.


Substituting for R from (4) we get

ye  =  Qe 
Pdx Pdx
dx + A ...(5)

Similarly the solution of equation (2) can be obtained in the same manner by
multiplying equation (2), the integrating factor e 
P dy
and can be written as in the case of
the solution (5) in the form

xe  =  Qe 
P dy P dy
dy + B ...(6)

Solved Example 1. Solve


dy
cos2x + y = tan x ...(1)
dx
Solution. Dividing throughout by cos2 x, (1) reduces to
dy
+ y sec2x = tan x sec2 x
dx
which is in linear form.

= e
Pdx
Integrating factor

= e
2
sec xd x

= etan x
The required solution is
ye tan x =  tan x sec 2 xe tan x dx + c

To evaluate the integral on the right, put tan x = t so that sec2 x dx = dt,

21
 R.H.S. =  tet dt + c

= tet –  et  1  dt + c

= tet – et + c
= tan xetanx – etanx + c
 The required solution is

yetanx = tan x etan – etanx + c


or y = tan x – 1 + ce–tanx
where c is an arbitrary constant.
Solved Example 2. Solve
dy
x log x + y = 2 log x
dx
dy 1 2
Solution: + y=
dx x log x x
which is a linear equation.
1
 x log x dx
Integrating Factor (I.F.) = e = elog(log x ) = log x
 The solution is
2
y log x =  log x dx + c
x
 y log x = (log x)2 + c is the required solution.
Solved Example 3. Solve
dy
( x + y + 1) =1
dx
dy 1 .
Solution: We have =
dx x + y + 1
dy
This equation is not expressible as + Py = Q.
dx
dx
Now = x + y +1
dy
dx
 + (–1) x = y + 1
dy
which is a linear equation.

I.F. = e  = e – y
–1 dy

22
 Solution of equation (1) is

xe–y =  ( y + 1)e – y dy + c

=  ( ye – y + e – y )dy + c

= – ye – y +  e – y dy – e – y + c

= –ye–y – e–y – e–y + c


= –(y + 2) e–y + c'
 Required solution is
x + y + 2 = cey.
Solved Example 4. Solve
dy
( x + 2 y2 ) =y
dx
Solution: The d.e. can be written by changing the independent variable from x to y in
the form
dx
y = x + 2 y2
dy

dx x
or = + 2y
dy y

dx x
i.e. – = 2y ...(1)
dy y

dx 1
which is a linear equation of the form + P' x = Q' where P ' = – and Q' = 2y,
dy y
P' and Q' being functions of y only.
The integrating factor in this case is
1
 – dy 1
e  = e y = e – log y = elog1/ y =
Pdy

y
1
By multiplying the d.e. (1) by the integrating factor , we get the solution of
y
equation (1) in the form.
1 1
x = c +  2 y dy = c + 2 dy
y y
= c + 2y
or x = cy + 2y2

23
Equations Reducible to the Linear form
Sometimes equations which are not linear can be reduced to the linear form.
The differential equation of the type
dy
+ Py = Qy n ...(1)
dx
dy
or + P ' x = Q' xn ...(2)
dx
where P and Q are functions of x in (1), and P' and Q' are functions of y in (2) can be
reduced to the linear equation of the form discussed in the previous article.
The equation (1) can be written as
dy
y– n + Py – n +1 = Q ...(3)
dx
Now put y–n+1 = z ...(4)
dz dy
So that = (1 – n) y – n
dy dx

dy 1 dz
or y–n =
dx 1 – n dx
Therefore, the given d.e. (3) is transformed to
1 dz
= Pz = Q
1 – n dx
dz
or + (1 – n) Pz = Q(1 – n) ...(5)
dx
The equation (1) has been transformed to the d.e. (5) which is a linear equation with z
as dependent variable and can be solved by the method as given in IV. Similarly equation
(2) can be reduced to a linear form.

Solved Example 5. Solve


dy
x + y = x 3 y4
dx
Solution: Dividing throughout by xy4, we have
1 dy 1 1
4
+ 3  = x2
y dx y x

Putting y–3 = v, we get on differentiating w.r.t.x

24
dy dv
–3 y –4 =
dx dx
dy 1 dv
or y –4 =–
dx 3 dx
The equation transforms to
1 dv 1
– + v = x2
3 dx x

dv  3 
or –   v = –3 x 2 ,
dx  x 
3
1
which is in linear form with v as the dependent variable. I.F. = e x = e–3log x = . The
– dx

x
solution is
1 1
v 3
=  –3 x 2 3 dx + c
x x
v 3
i.e., 3
=  – dx + c
x x
v
 = –3log x + c
x3
 v = –3x3 log x + cx3
Replacing v by y–3, we get
1
y –3 = –3x 3 log x + cx 3 3
= –3x 3 log x + cx 3 as the required solution.
y
Solved Example 6. Solve
(x2y3 + xy)dy = dx ...(1)
Solution: The differential equation (1) can be written as
dx
– xy = x 2 y3 ...(2)
dy

which is of the form


dx
+ P ' x = Q' xn
dy

where P' = –y, Q' = y3 and n = 2.


The d.e. (2), on dividing by x2, is written in the form

25
dx
x –2 – x –1 y = y3 ...(3)
dy

Put x–1 = z ...(4)


1 dx dz
 –  =
x 2 dy dy

dx dz
i.e., x –2 =–
dy dy

Now the differential equation (3) is reduced to


dz
+ yz = – y3 ...(5)
dy
which is a linear differential equation and can be solved. The integrating factor is
y2
e = e 2
ydy

 The general solution of the d.e. (5) is given by


y2 y2
ze 2
= –  y e dy + c
3 2

Now consider the integral


y2

 y e 2 dy = 1 (say)
3

y2
Here put = u so that y dy = du
2
y2
Then I =  y e y dy is transformed to  2ue u du
2 2

Integrating by parts, we get


I = 2 ueu du = 2 ueu –  eu du  = 2[ueu – eu ]
 
y2 y2
= y 2 e 2 – 2e 2
 the solution of the equation (5) is given by
y2 y2 y2
z e 2 = c – I = c – y 2 e 2 + 2e 2
– y2
or z = ce 2
– y2 + 2
Putting z = x–1 by (4)
we get the solution of differential equation (1) as
2
–y
1
= ce 2 – y2 + 2
x

26
1 dy x
Solved Example 7. Solve + = xy –1/2
y dx 1 – x 2

Solution: The given differential equation can be written as


1 dy 1/2 1
+y =x
y1/2 dx 1 – x2

1 –1/2 dy dz
Put y +1/2 = z  y =
2 dx dx
1 dy +2dz
 =
y1/2 dx dx

dz x
Then 2 + z=x
dx 1 – x 2
dz x x
+ 2
z=
dx 2(1 – x ) 2
x 1 2x
 2(1– x 2 )dx  1– x 2 dx
I.F. = e = e4
1 1 1
– log(1– x 2 ) =
log (1– x 2 )1/4 (1– x 2 )4
=e 4
=e

Solution of given equation is


1 xdx
z 2 1/4
= +C
(1 – x ) 2(1 – x 2 )1/4
y1/2 dt
2 1/4
= –  1/4 + C (putting the value of z)
(1 – x ) yt
where 1 – x2 = t
1
xdx = – dt
2
y1/2 1 t
Hence =– +C
2 1/ 4
(1 – x ) 4 1
– +1
4

y 1
 2 1/4
= –  t 3/4 + C
(1 – x ) 3

3 y
 = –(1 – x 2 )3/4 + C '
(1 – x )1/4

 3 y = –(1 – x 2 ) + C '(1 – x 2 )1/4

 3 y + 1 – x 2 = C '(1 – x 2 )1/4

27
Exercise
Solve the following differential equations:
dy
1. + y sec x = tan x
dx
dy
2. (1 + x ) + y = 1
dx
dy
3. ( x + 2 y3 ) = y
dx
dx 1
[Hint : Rewrite the equation as – x = 2 y2 , which is linear]
dy y
dy
4. x + 2 y = x 2 log x
dx
dv
5. (1 – x 2 ) + 2 xy = x 1 – x
dx
 1  dy
6. y 2 +  x –  = 0
 y  dx
dx x 1
[Hint: Rewrite the equation as + 2 = 3 , which is a linear equation].
dy y y
dy
7. ( xy – x 2 ) = y2
dx
1 dx 1 1 1 1
[Hint: Rewrite the equation as 2
–  = – 2 put = z ].
x dy x y y x
dy
8. + x sin 2 y = x 2 cos2 y
dx
[Hint: Divide the equation by cos2y and then put tan y = z].
dy 1 e y
9. + =
dx x x 2
[Hint: Divide the equation by ey and put e–y = z].
dy y 2
10. x + = y.
dx x
11. (x2y2 + xy)dx = dy.
V. Method of Solving Exact Differential Equations
A differential equation is said to be exact if it can be derived from its primitive of
solution by direct differentiation without any further operation of elimination or reduction
or in other words, an exact differential equation is formed by equating an exact
differential to zero. For example, consider the differential equation
x dx + y dy = 0 ...(1)
The primitive or solution of (1) is

28
x2 + y2 = c ...(2)
By taking the differential of both sides of equation (2) we get differential equation
x dx + y dy = 0
The constant c has been eliminated directly without further operation. Therefore the
differential equation (1) is exact.
Consider another example of an exact differential equation viz.,
(a2 – 2xy – y2)dx – (x + y)2 dy = 0 ...(3)
The primitive or general solution of the d.e. (3) is
y3
a2x – x2y –xy2 – =c ...(4)
3
Taking the differential of both sides of the primitive (4), we get
a2dx – (2xy dx + x2 dy) – (y2 dx + 2xy dy) – y2 dy = 0
or (a2 – 2xy – y2)dx – (x + y)2 dy = 0
which is the same as the given d.e. (3). The arbitrary constant c of the equation (4) has
been eliminated by direct differentiation without any further operation.
Theorem 1. The necessary and sufficient condition for the ordinary differential
equation of the first order
M dx + N dy = 0
to be axact, is that
M N
=
y x

Proof. The proof for sufficiency being beyound the scope of study, we shall only
show that the condition is necessary.
In order that M dx + n dy be an exact differential, it must have been derived by
differentiating some function u of x and y and performing no other operation.
i.e. du = M dx + N dy
u u
But du = dx + dy,
x y

so that on comparison, we get


u u
= M, = N
y y

29
M  2u N  2 u
 = and =
y yx x xy

 2u  2u
Hence assuming that = , we obtain
yx xy

M N
=
y x

 The necessary condition for the differential equation M dx + N dy = 0 to be exact is

M N
= ...(1)
y x

Rule for finding the Solution of an Exact Differential Equation


 M N 
If M dx + N dy = 0 is exact i.e., = then the following procedure is to be
 y x 
adopted.
(1) Integrate M w.r.t.x, regarding y as a constant.
(2) Integrate w.r.t.y the terms in N which do not involve x.
The sum of the two integrals equated to an arbitrary constant is the required solution.
Solved Example 5. Solve
(ax + hy + g)dx + (hx + by + f)dy = 0 ...(1)
Solution: The differential equation (1) is of the form
M dx + N dy = 0
M = ax + hy + g 
where ...(2)
N = hx + by + f 

M N
 =h=
y x

The condition of an exact differential equation of the first order is satisfied. Hence the
differential equation (1) is exact.
Integrating M w.r.t.x regarding y as constant, we obtain by (2)

 M dx =  (ax + hy + g)dx ...(3)

We find from (2) that the terms in N free from x are (by + f). Integrating (by + f)
w.r.t.y, we have
by 2
 (by + f )dy = 2
+ fy ...(4)

30
 By the rule of solving exact equations stated above the general solution of the exact
differential equation (8) is by (3) and (4).
1 2  1 2 
 2 ax + (hy + g) x  +  2 by + fy  = constant
   
i.e., ax2 + by2 + 2hxy + 2gx + 2fy = c
c, being an arbitrary constant.
Integrating Factors
Sometimes an equation, which is not exact, can be made exact on multiplication by
some suitable function of x and y. Such a function is called an integrating factor. For
example, on multiplying the linear differential equation
dy
+ Py = Q ...(1)
dx

by the integrating factor e  the equation (1) becomes


P dx

 e  Pdx  dy + Py  e  P dx = Qe  P dx
 
  dx

d   P dx   P dx
or  ye  = Qe
dx  

d   P dx 
= Qe dx 
dx  

d   P dx 
–  Qe dx  = 0
P dx
or  ye
dx    

d
or [u( x, y)] = 0
dx

u( x, y) = ye  –  Qe 
P dx P dx
where dx

i.e., the linear equation (1) is made exact after being multiplied by the integrating factor
e
P dx

Solved Example 6. Solve


y dx – x dy + log x dx = 0
The given equation can be written as (y + log x) dx – x dy = 0
Solution: The equation is of the form
M dx + N dy = 0 where M = y + log x and N = –x.

31
M N M N
= 1, = –1, 
N x N x
 The equation is not exact.

1
Multiplying the given equation by we have
x2

 y + log x  1
 x2  dx – x dy = 0
 
It can be easily seen that the above equation is exact.
 It has for its solution

 y + log x 
  x2  dx = c

Integrating w.r.t.x, regarding y as constant, we get
y log x
– +  2 dx = c
x x

y  1 1
or – + log x  –  +  2 dx = c
x  x x

y log x 1
– – – =c
x x x
or cx + y + 1 + log x = 0
is the required solution.

Rules for finding integrating factors


 1 
Rule I. When Mx + Ny  0, and the differential equation is homogeneous,  
 Mx + Ny 
is an integrating factor of M dx + N dy = 0.
Example : (x2y – 2xy2) dx – (x3 – 3x2y) dy = 0 ...(1)
Here, M = x2y – 2xy2
N = 3x2y – x3
Mx + Ny = x3y – 2x2y2 + 3x2y2 – x3y
= x2y2  0,
and homogeneous in x and y.
Therefore according to Rule I, integrating factor (I.F.) is

32
1 1
= 2 2
Mx + Ny x y
 1 
Multiplying equation (1) by I.F.  2 2  , we get the exact differential equation
x y 
1 2  x 3
 –  dx –  2 –  dy = 0
y x y y
ydx – xdy 2 3
 2
– dx + dy = 0
y x y
x
 d   dx – 2d (log x ) + 3d (log y) = 0
 y
x
Integrating d – 2 log x + 3 log y = c
y

x  y3 
or + log  2 =c
y x 

Rule II. When Mx – Ny  0, and the differential equation is of the form


f1(xy) y dx + f2 (xy) x dy = 0,
1
is an integrating factor.
Mx – Ny

Example : y (xy + 2x2y2)dx + x (xy – x2y2)dy = 0. ...(1)


Here, Mx – Ny = x2y2 + 2x3y3 – x2y2 + x3y3
= 3x3y3  0,
and it is of the form f1(xy) y dx + f2(xy) x dy = 0
where f1(xy) = xy + 2x2y2 and f2(xy) = xy – x2y2,
1
so its integrating factor is
3x 3 y 3

 1 
Multiplying equation (1) by I.F.  3 3  , we have
 3x y 

ydx + xdy 2 1
2 2
+ dx – dy = 0
x y x y

d ( xy)
+ 2d (ln x) – d (ln y) = 0
x 2 y2

Integrating,

33
1 x2
– + ln = c
xy y

x2 1
ln = +c
y x4

 M N 
 y – x 
Rule II. When   is a function of a x alone, say f(x), then e  –( x ) dx is an
N
integrating factor.
Example : (x2 + y2) dx – 2xy dy = 0
M = x2 + y2, N = – 2xy
M N
= 2 y, = –2 y
y x

 M N 
 y – x 
  = +4 y = – 2 = f ( x )
N –2 xy x

= e
f ( x ) dx
I.F. Integrating factor
2 1
1
= e  x = e –2 ln x = e x 2 = 2
– dx ln

x
1 ,
Multiplying by equation (1) by I.F. we get
x2

y 2 dx – 2 xy dy
dx + =0
x2
Integrating
y2
x– = c,
x
x2 – y2 = cx
 N M 
 x – y 
Rule IV. When   is a function of y alone, say f (y), then e  f1 ( y ) dy is an
1
M
integrating factor.
Example: (3x2y4 + 2xy) dx + (2x3y3 – x2) dy = 0. ...(1)
M = 3x2y4 + 2xy, N = 2x3y3 – x2
M N
= 12 x 2 y3 + 2 x, = 6 x 2 y3 – 2 x
y x

34
 N M 
 x – y 
 = 6 x y – 2 x – 12 x y – 2 x = –2 x(3xy + 2)
2 3 2 3 3

M 3 x y + 2 xy
2 4
xy(3xy + 2)
3

–2 x(3xy3 + 2) 2
= = – = f1 ( y)(say)
xy(3xy + 2)
3
y
2 1
 – y dy ln 1
I.F. = e = ey 2
=
y2
  – 2y dy 
Multiplying equation (1) by integrating factor  e  we have
 
 
(3 x 2 y 4 + 2 xy) 2 x 3 y3 – x 2
dx + dy = 0
y2 y2
2 xydx – x 2 y
3 x 2 y2 dx + 2 x 3 ydy + =0
y2
x2
Integrating x 3 y2 + =c
y

Exercise
Solve the following equations:
1. (x3 + 3xy2) dx + (y3 + 3x2y) dy = 0.
2. ey dx + (1 + x ey) dy = 0.
3. (xy2 + x) dx + (yx2 + y) dy = 0.
4. (x2 – ay) dx = (ax – y2) dy.
5. (ey + 1) cos x dx + ey sin x dy = 0.
6. (a2 – 2xy – y2) dx – (x + y)2 dy = 0.
7. (y – 2x3) dx – x (1 – xy) dy = 0.
1
[Hint: To make this equation exact, multiply by ]
x2
8. x2y dx – (x3 + y3) dy = 0.
9. (x2 + y2 + 2x) dx + 2y dy = 0.
10. (y4 + 2y) dx + (xy3 + 2y4 – 4x) dy = 0.

35
Lesson 3
LINEAR DIFFERENTIAL EQUATIONS WITH
CONSTANT COEFFICIENTS-I

Ordinary differential equations of an order higher that the first are to be considered now.
In this lesson and the next we will study a single class of these equations called ordinary
linear differential equations.
1. Definition
Ordinary linear differential equations are those in which the dependent variable and its
derivatives appear only in the first degree and are not multiplied together, their
coefficients all being constants or functions of x. Thus the general linear differential
equation of the nth order is of the form
dny d n –1 y d n –2 y dy
n
+ p1 n –1
+ p2 n –2
+ ... + pn –1 + pn y = Q ...(1)
dx dx dx dx
where p1, p2, p3 ... pn – 1, pn and Q are constants or functions of x and the co-efficient of
dny
the highest order derivative viz., is unity. If the co efficient of the highest order
dx n
derivative is not units, then the equation can be divide by this co-efficient and then the
differential equation will be in the form (1).
The linear differential equation of the first order
dy
+ p1 y = Q
dx
is obtained from (1) by putting n = 1 and this equation has already been solved in an
earlier lesson.
2. Linearly Independent Functions
A set of n functions f1(x), f2(x), ...fn(x) are said to be linearly dependent if we can find
constants b1, b2, ...bn; not all zero such that
b1f1(x) + b2f2(x) + .........+ bnfn(x) = 0
b2 b bn
or f1 ( x) = – f 2 ( x ) – 3 f3 ( x ) – fn ( x)
b1 b1 b1

where b1  0 and the function f1(x) is expressed in terms of the other remaining (n – 1)
functions.
On the other hand a number of n functions
f1(x), f2(x) ............, fn(x)

36
are said to be linearly independent if no set of constant b1, b2, b3 ............ bn (which are not
all zeros) exists, such that
b1f1(x), b2f2(x) + b3(x) ............. + bnfn(x) = 0.
In this case a function say f1(x) cannot be expressed in terms of the remaining
functions.
3. Let us first consider the general solution of the equation
dny d n –1 y d n –2 y dy
+ p1 + p2 + + pn –1 + pn y = 0 ...(1)
dx n dx n –1 dx n –2 dx

where p1, p2 .......... pn–1, pn are constants or functions of x.


The differential equation (1) is the same as the equation (1) of $1 with the right-hand
side Q = 0.
Theorem 1. If y = f1(x), y = f2(x), ..... y = fn are solutions of the d.e. (1) and if the n
functions
f1(x), f2(x), f3(x), ........... fn(x)
are also independent of one another, then general solution of the equation (1) is
y = c1f1(x) + c2f2(x) + ........... + cnfn(x)
where c1, c2, ....... cn are arbitrary constants.
dny d n –1 y d n –2 y dy
Proof: + p1 + p2 + + pn –1 + pn y = 0 ...(1)
dx n dx n –1 dx n –2 dx
is the given differential equation.
If y = f1(x) is a solution of the equation (1) then y = c1f1(x), c1 being an arbitrary
constant, as also a solution. For, since y = f1(x) is supposed to be a solution of the
equation (1), on substituting f1(x) for y left hand side of (1), we get
dn d n –1 y d
( c f
1 1 ) + p1 (c1 f1 ) + + pn –1 + (c1 f1 ) + pn (c1 f1 )
dx n dx n –1 dx

dn f d n –1 f df1 
 c1  n1 + p1 n –11 + + pn –1 + pn f1 
 dx dx dx 

 c1  0 = 0 by the relation (2)

Thus y = c1f1(x) is also a solution of (1), when y = f1(x) is a solution.


Similarly if y = f2(x), y = f3(x), .......... y = fn(x) be solution of (1)
then y = c2f1(x), y = c3f3(x), .......... y = cnfn(x),
where c2, c3, cn are arbitrary constants, are all solutions of the equation (1).

37
Since y = c1f1(x), c2f2(x) ........., y = cnfn(x) are the solutions of the d.e. (1) we conclude
that
y = c1f1(x) + c2f2(x) + ......... + cnfn(x)
r =n
=  cr fr ( x ) ...(3)
r =1

is also a solution of the equation (1), for by substituting


n
y =  crfr (x) in the L.H.S. of (1), we get
r =1

dn d n –1
[c1 f1 + c2 f2 + + cn f n ] + p1 [c1 f1 + c2 f2 + cn f n ]
dx n dx n –1
d
+ + pn –1 [c1 f1 + c2 f2 + + cn f n ] + pn [c1 f1 + c2 f2 + + cn f n ]
dx

dn f d n –1 f d n –2 f2 
= c1  n1 + p1 n –11 + p2 + + pn f1 
 dx dx dx n –2 

dn f d n –1 f2 
+ c2  n2 + p1 + + pn f2  +
 dx dx n –1 

 d n fn –1 d n –1 fn –1 d n –2 f n –1 
+ cn –1  n
+ p1 + p2 + + pn fn –1 
 dx dx n –1 dx n –2 

 d n fn d n –1 fn 
+ cn  n
+ p1 + + pn fn 
 dx dx n –1 

=0 ...(4)
 y = c1f1 + c2f2 + c3f3 + ... cnfn
satisfies the d.e. (1) Moreover the solution (3) is the general solution of the nth order
differential equation (1) since it contains n arbitrary constants and the n functions f1(x) .
f2(x) ... fn(x) are supposed to be linearly independent.
In these lessons we shall only deal with linear differential equations with constant co-
efficients and homogeneous linear equations the discussion of the general equations is
beyond the present course of study.
4. Let us suppose that a particular solution y = (x) involving no arbitrary constant, of the
differential equation
dny d n –1 d n –2 y dy
+ p1 + p2 + + pn –1 + pn y = Q
dx n dx n –1 dx n –2 dy

38
where p1, p2 ... pn are constants and Q is either a constant or functions of x can be guessed
or obtained in any manner.
Theorem 2. The general solution of the differential equation
dny d n –1 y d n –2 y dy
+ p1 + p2 + + pn –1 + pn y = Q ...(1)
dx n dx n –1 dx n –2 dy

dny d n –1
i.e., + p1 + + pn y – Q = 0 ...(1')
dx n dx n –1
is y = [c1f1(x) + c2f2(x) + ... + cnfn(x)] + f(x) ...(2)
= u + (x) ...(2')
in which c1, c2 ..., cn are the n arbitrary constants and (x) is any particular solution of the
equation (1) involving no arbitrary constant, and f1(x), f2(x), f3(x), ... fn(x), are the
independent solutions of the d.e.,
dny d n –1 d n –2 y dy
n
+ p1 n –1 + p2 n –2 + + pn –1 + pn y = 0 ...(3)
dx dx dx dx
Proof : In 3, it has been proved already that
r =n
c1f1(x) + c2f2(x) + ... + cnfn(x) = c
r =1
f ( x)
r r

is the general solution of the differential equation (3).


Since y = (x) is supposed to be a particular solution of the d.e., (2) it therefore
satisfies the equation (1) i.e.,
d n d n –1 d n –2  d
+ p1 + p2 + pn –1 + pn  = Q ...(4)
dx n dx n –1 dx n –2 dx
Now it can be easily verified that (2) is a solution of the differential equation (1), for
substituting
r =n
y =  cr fr ( x ) + ( x )
r =1

in left hand side of (1'), we get


dn
[c1 f1 + c2 f 2 + + c n f n + ]
dx n

dn
+ [c1 f1 + c2 f2 + + cn f n + ]
dx n

............................................

39
d
+ pn –1 [c1 f1 + c2 f2 + + cn f n + ]
dx
+ pn [c1f1 + c2f2 + ... + cnfn + f] – Q
dn
= [ c1 f1 + c2 f2 + ... + cn f n ]
dx n
d n –1
+ p1 [c1 f1 + c2 f2 + cn f n ] +
dx n –1
d
+ pn –1 [c1 f1 + c2 f2 + cn f n ]
dx
= pn[c1f1 + c2f2 + ... + cnfn]
 d n d n –1 
+  n + p1 + + pn   – Q
 dx dx n –1 
= 0 by relations (4) of 3 and 4.
 (2) is the general solution or primitive of the equation (1) of (1'), since it contains n
arbitrary constants and the solution f1(x), f2(x), ... fn(x) of the d.e. (3) are linearly
independent. Hence the general solution (2) of the differential equation (1) or (1') consists
of two parts viz., (i) the general solution of the d.e. (2), containing n arbitrary constants,
and (ii) any particularly integral (involving no constant) of the equation (1) or (1').
The part of the solution (2), viz.,
u = c1f1(x) + c2f2(x) + ... + cnfn(x)
is called the complementary function and the part
v = (x)
is called the particular integral. The general or complete solution is the sum of the
complementary function and the particular integral.
6. Symbolic Operator
Let us write
dy 
Dy for
dx 

2 d2y 
d y for
dx 2 

3 d3y  ...(1)
d y for
dx 3 

..................... 
dr y 
D r y for 
dx r 

40
d d2 d'
so that D stands for the operator , D 2 for 2 ..., D' for .
dx dx dx '
The index of D indicates the number of times the operation of differentiation is to be
carried out. Hence we write,
dny d n –1 y d n –2 y dy
+ p1 + p2 + + pn –1 + pn y
dx n dx n –1 dx n –2 dx
as (Dny + p1Dn–1y + p2Dn–2y + ... + pn–1Dy + pny)
or (Dn + p1Dn–1 + ... + pn–1D + pn)y
or f(D)y ...(2)
where f (D)y denotes the operator
Dn + p1Dn–1 + p2Dn–2 + ... + pn–1D + pn
dn d n –1 d n –2 d
i.e., + p1 + p2 + + pn –1 + pn
dx n dx n –1 dx n –2 dx
We have (D – m2)y = Dy – m2y
dy
= – m2 y
dx
The notation (D – m1) (D – m2) y is defined to denote that (D – m2) operates first on y
and then (D – m1) operates on (Dy – m2y)
If m1 and m2 are constants then
(D – m1) (D – m2)y = (D – m1) (Dy – m2y)
 dy 
= ( D – m1 )  – m2 y 
 dx 

 dy   dy 
= D  – m2 y  – m1  – m2 y 
 dx   dx 

d  dy   dy 
=  – m2 y  – m1  – m2 y 
dx  dx   dx 

d2y dy dy
= 2
– m2 – m1 + m1m2 y
dx dx dx
= [D2 – (m1 + m2) D + m1m2]y
Again if the order of two operational factors (D – m1) and (D – m2) is changed then
 dy 
(D – m2 )(D – m1 ) y = ( D – m2 )  – m1 y 
 dx 

41
d  dy   dy 
=  – m1 y  – m2  – m1 y 
dx  dx   dx 

d2y dy dy
= – m1 – m2 + m1 m2 y
dx 2 dx dx

= D2 y – (m1 + m2 )Dy + m1m2 y


= [D2 – (m1 + m2 )D + m1m2 ]y
which is the same as (4).
 (D – m1) (D – m2) y = (D – m2) (D – ml)y
= [D2 – (m1 + m2)D + m1m2]y.
Therefore the order of the operational factors (D – m1) and (D – m2) is immaterial and
m1 and m2 are constants.
Similarly if m1, m2 and m3 are constants (D – m3) (D – m2) (D – m1) y denotes that
(D – m3) operates on (D – m2) (D – m1)y.
 (D – m3) (D – m2) (D – m1)y
d2y dy 
= ( D – m3 )  2 – (m1 + m2 ) + m1m2 y  by result (4)
 dx dx 

d d2y dy 
=  2 – (m1 + m2 ) + m1 m2 y 
dx  dx dx 

d2y dy 
– m3  2 – (m1 + m2 ) + m1 m2 y 
 dx dx 

d2y d2y dy
= 3
– (m1 + m2 ) 2 + m1m2
dx dx dx

d2y dy
– m3 2
+ m3 (m1 + m2 ) – m1m2 m3 y
dx dx

d3y d2y dy
= 3
– ( m1 + m2 + m3 ) 2
+ (m1m2 + m2 m3 + m3 m1 ) – m1m2 m3 y
dx dx dx

= D3 y – (m1 + m2 + m3 )D2 y + (m1m2 + m2 m3 + m3m1 )Dy – m1m2 m3 y

= [ D3 – (m1 + m2 + m3 )D2 + (m1m2 + m2 m3 + m3m1 )D – m1m2 m3 ]y ...(5)

It can be easily verified that in whatever order the three operational factors (D – m1),
(D – m2), (D – m3) are considered, the final result is the same as (5).
In general if,

42
f(D)y = (Dn + p1Dn–1 + p2Dn–2 + ... pn–1D + pn)y
= (D – 1) (D – 2) ... (D – n)y
assuming that the polynominal in D is factorizable into n real linear factors. Then it can
be seen that order of the factors in the R.H.S. is immaterial. Since D stands for the
d
operator we write
dx
f(m) = mn + p1mn–1 + p2mn–2 + ... + pn–1m + pn
There 1, 2 ... an are the roots of the equation f(m) = 0
f(m) = 0 is called the auxiliary equation of the given diff. equation f(D)y = 0
For example, if the given diff. equation is (D2 – 5D + 6) y = 0
the auxiliary equation will be m2 – 5m + 6 = 0 which can be factorized as
(m – 3) (m – 2) = 0 and hence has the roots 3 and 2.
6. Solution of the Linear Differential Equation with Constant Co-efficients
Let us find the general solution of the differential equation
dny d n –1 y d n –2 dy
+ p1 + p2 + + pn –1 + pn y = 0 ...(1)
dx n dx n –2 dx n –2 dx

where p1, p2, .. pn–1, pn are all constants.


The differential equation (1) may be written in the symbolic form as
[f(D)]y = [Dn + p1Dn–1 + p2Dn–2 + ... pn–1D + pn]y
= (D – mn) (D – mn–1) ... (D – m2) (D – m1)y
=0 ...(2)
where m1, m2, m3, ... mn–1, mn are the n roots of the auxiliary equation
f(m) = mn + p1mn–1 + ... + pn–1m + pn ...(3)
=0
The solution of the ordinary d.e. of the first order
(D – m1)y = 0 ...(4)
is also a solution of equation (2) and therefore of the equation (1), because from (2), we
get
f(D)y = (D – mn) (D – mn–1) ... (D – m2) 0
= 0 [as (D – m1)y = 0 by (4)]

43
Now since the operational factors of the equation (2) can be put in any order,
therefore the solution each of the following (n – 1) ordinary differential equation of the
first order viz.,
( D – m2 ) y = 0 

( D – m3 ) y = 0 
......................  ...(5)

( D – mn –1 ) y = 0 
( D – mn ) y = 0 

is also a solution of the equation (2) and therefore of the equation (1).
The differential equation (4) gives
(D – m1)y = 0
dy
i.e., – m1 y = 0
dx
Separating the variables and integrating, we have
dy
 y 
= m1dx

where the arbitrary constant has been omitted.


 log y = m1x
i.e., y = em x 1
...(6)
Similarly each of the (n – 1) differential equations of the first order included in (5)
can be solved and we get the (n – 1) solution, viz.

y = e m2 x 

y = e m3 x 
............  ...(7)

y = e mn–1 x 

y = e mn–1 x 

The arbitrary constant omitted in each case.


 From (6) and (7) it follows that

y = em1x , em2 x , em3 x , em4 x , emn–1 x , emn x ...(8)

are the solutions of the differential equation (1).


If the n roots m1, m2, m3, ... mn of the auxiliary equation (3) are all real and distinct
then the set of n solution (8) is linearly independent and therefore form a set of n

44
independent solution. Then the general solution of the differential equation (1) is given by
(3) of 3, as
y = c1em1 x + c2 em2 x + + cn –1emn–1 x + cn emn x ...(9)

where c1, c2, c3, ... cn–1,cn and n arbitrary constants.


Solved Example 1. Solve
d2y dy
2 2
+ 9 – 18 y = 0
dx dx

d2y dy
Solution: 2 2
+ 9 – 18 y = 0
dx dx
Now dividing the d.e. (1) by 2, we get
d 2 y 9 dy
= – 9y = 0
dx 2 2 dx

 2 9 
or  D + 2 D – 9 y = 0 ...(2)
 
i.e., f(D) y = 0
(1) is linear differential equation of 2nd order with constant co-efficients and therefore
of the type of equation (1) of 6.
The auxiliary equation of the d.e. (2) is the algebraic equation
9
f ( m)  m 2 + m – 9 = 0
2
or 2m2 + 9m – 18 = 0
i.e., (2m – 3) (m + 6) = 0 ...(3)
3
where m1 = –6 and m2 = .
2
The general solution of the d.e. (1) by (9) of 6, is
y = c1e–6 x + c2e3 x /2

c1, c2, being two arbitrary constants.


7. Case of the Auxiliary Equation having Equal Roots
Case of two equation roots of the auxiliary equation
If all the n roots of the auxiliary equation (3) of  5 viz.
f(m) = 0 ...(1)

45
be not all distinct, let two of them be equal say m1 = m2 then the solution (9) of the d.e.
(1) of 6 can be written as
y = c1em1 x + c2 em1 x + c3em3 x + + cn emn x

= (c1 + c2 )em x + c3em x +


1 3
+ cn emn x

= c2 em x + c3em x +
1 3
+ cn emn x ...(A)

where c1 + c2 is replaced by a single arbitrary constant c. Therefore this solution has


(n – 1) arbitrary constants viz., c, c3, c4, ... cn whereas the order of the differential
equation (1) of 6 is n.
Therefore (A) cannot represent the general solution.
The original differential equation is
f(D)y = (D – mn) (D – mn–1) ... (D – m2) (D – m1)y
= (D – mn) (D – mn–1) ... (D – m1)2y ( m2 = m1)
The differential equation (1) of 6 is satisfied by the solution of differential equation
of the second order viz.
(D – m1)2y = 0
Therefore the general solution of the equation (2) is also a solution of the equation (1)
of 6. But since the d.e. (2) is of the second order, therefore its general solution or
primitive solution contain two arbitrary constants. The general solution of (2) is given by
(D – m1)2y = 0
or (D – m1) (D – m1)y = 0
Put (D – m1)y = v
The differential equation (3) is reduced to
(D – m1)v = 0
dv
i.e., – m1 v = 0
dx
Separating the variables and integrating, we get
dv
 v 
= m1dx + log c2

c2 being the arbitrary constant


or logv = m1x + logc2

46
v
or log = m1 x
c2

v
i.e., = e m1 x
c2

i.e., v = c2 em1x

Now from (4), we get (D – m1) y = v = c2 em x 1

m1 x
i.e., ( D – m1 ) y = v = c2 e
dy
i.e., – m1 y = c2 e m1 x ...(5)
dx
which is linear differential equation of the first order. The integrating factor for this d.e. is

e = e
Pdx – m1dx
+ e – m1 x
 The general solution of the equation (5) is given by
ye – m1 x = c1 +  c2 e m1 x  e – m1 x dx

= c1 + c2  dx

= c1 + c2x
c1 being another arbitrary constant
 y = em1 x [c1 + c2 x ] ...(6)

which contains two arbitrary constants c1 and c2.


Now the general solution of the equation (1) consists of
(i) the general solution of the linear equation of (n – 2)th order viz.,
(D – mn) (D – mn–1) ... (D – m4) (D – m3)y = 0 ...(7)
where mn, mn–1, ... m4 and m3 are supposed to be distinct, and
(ii) the general solution of the d.e. (2) of the 2nd order.
Now the general solution of the equation (7) is
y = c3em3 x + c4 em4 x + cn –1emn–1 x + cnemn x ...(8)
which involves (n – 2) arbitrary constants viz., c3, c4, ..., cn.
So from (6) and (8), we get the complete primitive of the d.e. (1) in the case when two
roots of the auxiliary equation are equal as
y = (c1 + c2 x)em1 x + [c3em3 x + c4 em4 x + + cne mn x ] ...(9)

involving n arbitrary constants c1, c2, c3, ... cn.

47
In a similar manner it can be shown that if r roots of the auxiliary equation f(m) = 0 of
the d.e., (1), of 6 are equal say m1 = m2 = ... = mr then the general solution of the d.e. (1)
of 6 is

y = [(c1 + c2 x + c3 x 2 + cr x r –1 )emn x ]

+ [cr +1em r +1 x
+ cr +2 emr+2 x + cn emn x ] ...(B)

Note here that the part of the complementary function corresponding to r roots of the
auxiliary equation being equal (each being equal to m1) is

[c1 + c2 x + + cr x r –1 ]emn x

Solved Example 2. Solve


d4y d3y d2y dy
4
– 3
– 9 2
– 11 – 4 y = 0
dx dx dx dx
Solution: The differential equation is of the form
(D4 – D3 – 9D2 – 11D – 4)y = 0
or f(D)y = 0
where f(D) = D4 – D3 – 9D2 – 11D – 4
The auxiliary equation is
f(m) = m4 – m3 – 9m2 – 11m – 4 = 0
or (m – 4) (m3 + 3m2 + 3m + 1) = 0
or (m – 4) (m + 1)3 = 0
One root of the auxiliary equation is 4 and the other three roots equal, each being
(–1). The complete primitive of the given differential equation by (B) of the last article is
y = (c1 + c2x + c3x2)e–x + c4e4x
c1, c2, c3, and c4 being arbitrary constants.
8. Case of the Auxiliary Equation having Complex Roots
Let the auxiliary equation
f(m) = 0
of the differential equation (1) or (2) of 6 viz f(D)y = 0, have complex roots. But for
equations with real co-efficients, complex roots occur in pairs. Suppose
m1 = a + ib
m2 = a – ib

48
are the complex the roots of the equation (1) and other roots m3, m4, mn are suppose to be
real and distinct.
Therefore the general solution of d.e. f(D)y = 0 by (9) of 6 is
y = [c1e( a+ib) x + c2 e( a – ib) x ] + c3em3 x + c4 em4 x + cn –1emn–1 x + cnemn x ]

But [c1e(a+ib) x + c2e(a–ib) x ]

= eax [c1eibx + c2e– ibx ]

= eax [c1 (cos bx + i sin bx) + c2 (cos bx – i sin bx)]

= eax [(c1 + c2 ) cos bx + (c1 – c2 ) i sin bx)]


 y = eax[A cos bx + B sin bx]
+ c3 e m x + c4 e m x +
3 4
+ cn –1e m1 n–1 x + cn e mn x  ...(2)

where A = (c1 + c2) and B = [i(c1 – c2)] . c3, c4, ... cn are n arbitrary constants.
Now put
A = R cos  and B = –R sin 
–B
 A2 + B2 = R2, tan  = ...(3)
A
The complete primitive (2) may also be written in the form
y = Re ax  cos  cos bx – sin  sin bx  + c3e m x + 3
+ cn e mn x 

= Re ax cos(bx + ) + [c3em x + 3
+ cn emn x ]

involving n arbitrary constants R,  c3, c4 .. cn.


Solved Example 3. Solve
d3y
– 8y = 0
dx 3
Solution: The given d.e. is
(D3 – 8)y – f(D)y = 0
where f(D) = D3 – 8
The auxiliary equation is
f(m) = m3 – 8 = 0
or (m – 2) (m2 + 2m + 4) = 0

–2  4 – 16
 m = 2 and m = = –1  i 3
2

49
(
The three roots of the auxiliary equation are 2 and –1  i 3 )
The complete primitive by (2) of the last article, is therefore, given by
y = e– x  A cos 3x + B sin 3x  + Ce2 x ,
 
A, B and C being three arbitrary constants.
Solved Example 4. Solve the differential equation
(D2 + a2) (D – b) y = 0 ...(1)
Solution: This differential equation is equivalent to
f(D)y = 0
where f(D) = (D2 + a2) (D – b),
Its auxiliary equation is
(m2 + a2) (m – b) = 0
 m2 = – a2 i.e., m = + ia ...(2)
and m = b.
 The root of the auxiliary equation are + ia and b.
 The general solution of the given differential equation is
y = [c1eiax + c2e–iax] + c3ebx
= c1 (cos ax + i sin ax) + c2 (cos ax – i sin ax) + c3ebx
= (c1 + c2) cos ax + i (c1 – c2) sin ax + c3ebx
= A cos ax + B sin ax + c3ebx
= R cos (ax + a) + c3ebx
by putting A = R cos; – B = R sin .
Solved Example 5. Solve the differential equation
d4y d3y d2y dy
4
– 2 3
+ 5 2
– 8 + 4y = 0
dx dx dx dx
Solution: We have
(D4 – 2D3 + 5D2 – 8D + 4) y = 0
The auxiliary equation is
m4 – 2m3 + 5m2 – 8m + 4 = 0
i.e., (m – 1)2 (m2 + 4) = 0

50
So that roots of auxiliary equation are 1, 1, 2i, – 2i.
The root 1 being repeated twice.
The general solution of the given differential equation is
y = (c1 + c2x)ex + c3e2ix + c4e–2ix
= (c1 + c3x)ex + c3(cos 2x + i sin 2x) + c4 (cos 2x – i sin 2x)
= (c1 + c3x)ex + A cos 2x + B sin 2x
by putting c3 + c4 = A, and i (c3 – c4) = B.
9. Auxiliary Equation having equal Complex Roots
(a) Case of equal pairs of Imaginary Roots
Let two pairs of imaginary roots of the auxiliary equation (1) of 6, viz.
f(m) = 0
be equal say
m1 = m2 = a + ib
m3 = m4 = a – ib
The other roots m5, m6 ..., mn are supposed to be distinct.
With the help of the solution (9) of 7, we get solution of the d.e. (1) 6 as
y = [(c1 + c3x)e(a+ib)x + (c3 + c4x)e(a – ib)x].
+ [c5em5x + c6em6x + ... + cnemnx] ...(1)
Now the first part of the R.H.S.
= eax [(c1 + c3x) (cosbx + i sin bx) + (c3 + c4x) (cosbx – sin bx)
= eax [(c1 + c3) cos bx + i (c1 – c3) sin bx]
+ x [(c2 + c4) cos bx + i (c2 – c4) sin bx]
 y = eax[(A1 cos bx + B1 sin bx) + x(A2 cos bx + B2 sin bx)] + (c5em5x + ... + cnemnx)
where A1 = c1 + c3, B1 = i (c1 – c3),
A2 = c2 + c4 and B3 = i (c2 – c4),
or y = eax[(A1 + x2A) cos bx + (B1 + xB2) sin bx]
+ [c5em5x + ... + cnemnx] ...(2)
If substitution in (3) 8 are used such that
A1 = R1 cos 1, B1 = –R1 sin 1
A1 = R2 cos 2, B2 = – R2 sin 2
The solution (1) takes the form

51
y = [R1eax cos (bx + 1) + R2 xeax cos (bx + 2)]
+ c5e m x +
5
+ cn e mn x  ...(3)

Therefore the complete primitive of the d.e. (1) 6 may be expressed in one the three
forms (1), (2) and (3). It may be noted that corresponding to the repeated roots a + ib, any
one of these forms contain four arbitrary constants.
Solved Example 6. Solve
d2y dy
2
– 4 + 5y = 0 ...(1)
dx dx
dy
given that y = 1 and = 2, when x = 0
dx
Solution: The differential equation is (in symbolic notation)
(D2 – 4D + 5)y = 0
or [f(D)] y = 0
where f(D) = D2 – 4D + 5.
The auxiliary equation is
f(m) = m2 – 4m + 5 = 0 ...(2)
4  16 – 20
 m= =2i
2
The roots of the auxiliary equation are 2 + i.
 The complete primitive of the d.e. (1), as by (2) of 8, is
y = e2x[A cos x + B sin x] ...(3)
From the given conditions we shall determine the values of the arbitrary constants A
and B. We have
dy
= 2e2 x [ A cos x + B sin x ] + e2 x [– A sin x + B cos x ] ...(4)
dx
By the given condition y = 1 when x = 0, we get from (3)
1=A
dy
Again by the other given condition viz. that = 2 when x = 0, we get from (4)
dx
2 = 2A + B
From (5) A=1
B=0
Substituting these values for A and B from (5) and (6), the solution (3) is given by
y = e2x cos x.

52
Exercises
Solve the following differential equations:
d2y
1. –  2 y = 0,  being a given constant.
dx 2
d2y
2. +  2 y = 0,  being a given constant.
dx 2
d4y
3. – a 4 y = 0,  a being a given constant.
dx 4
d3y d2y dy
4. 3
– 2 2
– 4 + 8 y = 0.
dx dx dx
d3y d2y
5. – 3 + 4 y = 0.
dx 3 dx 2
d2y dy
6. + 4 + 13 y = 0.
dx 2 dx
d4y d2y
7. – 2 + y = 0.
dx 4 dx 2
d3y d2y dy
8. 3
– 2 2
– 4 + 8 y = 0.
dx dx dx
d4y d2y
9. 4
+ 2 2 + y = 0.
dx dx
d2y dx
10. 9 2
+ 18 – 16 x = 0.
dx dy
[Hint : Here x is the dependent variable and y the independent variable. Auxiliary
equation is 9m2 + 18m – 16 = 0]
d3y
11. + y = 0.
dx 3
d2y
12. Solve + x = 0,   0; given that
dx 2
dx  .
x = a, = 0, when t =
dt 
d 4 y d 3 y 2 d 2 y dy
13. + + – + 3y = 0
dx 4 dx 3 dx 2 dx

53
Lesson 4
LINEAR DIFFERENTIAL EQUATIONS WITH CONSTANT
COEFICIENTS (Non-Homogeneous)-I

10. Now consider the differential equation

dny d n –1 y d n –2 y dy
+ p1 + p2 + pn –1 + pn y = Q ...(1)
dx n dx n –1 dx n –2 dx
where p1, p2, ... pn are all constants and Q is a function of x only.
It was pointed out in 4 of Lesson no. 5, that the general solution of the equation
(1) consists of two parts viz, (i) the complementary function (C.F.), i.e., the general
solution of the differential equation (1) when Q = 0 and (ii) a particular solution, i.e., a
particular integral (P.I.) of the d.e. (1) involving no arbitrary constant.
When Q = 0, the differential equation is reduced to the equation (1) to 6. Methods to
find the general solution of the equation (1) 6 were discussed in detail in the subsequent
articles 7, 8 and 9. Therefore the complementary function of equation (1) may be
supposed to be already found out. In the next few articles methods to find a particular
integral of the equation (1) will be taken up so that the complete integral of the equation
(1) may be obtained.
11. Inverse Operators and Particular Integral
Definition. If Q be a function of x then
1
Q ...(1)
( D – m1 )

stands for the function of x, such that when (D – m1) operates on n, the result is the
function Q
1
i.e. ( D – m1 )  Q = Q.
( D – m1 )

Let by such a function, i.e., let


1
y= Q, ...(2)
( D – m1 )

m1 being a constant, then according to the definition


(D – m1)y = Q
i.e. Dy – m1y = Q

54
dy
i.e., – m1 y = Q ...(3)
dx
1
i.e., the function y = Q must satisfy the linear differential equation (3) of the
( D – m1 )

first order whose solution can be found out. The integrating factor = e  1 = e – m1 x
– m dx

 The general solution of equation (3) is


ye– m1 x = c +  Qe– m1 x dx,

c being an arbitrary constant.

 ycem1x + em1x  Qe– m1x dx,

If we put c = 0, then a particular integral of the linear equation (3) of the first order is
obtained in the form
1
y= Q = em1 x  Qe – m1 x dx ...(4)
( D – m1 )

In particular if m1 = 0,then from (4)


1
y= Q =  Qdx ...(5)
D
1
Let the symbol Q, where m1 and m2 are constants, denote a
( D – m2 )( D – m1 )
1 1
function which is obtained after operates on the function Q.
( D – m2 ) ( D – m1 )

1
Let y= Q ...(6)
( D – m2 )( D – m1 )

1
Then, by definition (D – m2)y = Q.
( D – m1 )

Then again, by definition


(D – m1) (D – m2)y = Q ...(7)
1
i.e., the function y = Q satisfies the linear differential equation (7) of
( D – m2 )( D – m1 )
the 2nd order.
A particular integral of the linear equation (7) of the 2nd order is given is given by (4),
in the form

55
1
y= Q
( D – m2 )( D – m1 )

1  1 
=  Q
( D – m2 )  ( D – m1 ) 

=
1
( D – m2 )
( )
em1 x  Qe – m1 x dx by (4)

( )
= em2 x  em1x  Qe – m1x dx e – m2 x dx, on again using (4)

= em2 x  e( m1 – m2 ) x  (Qe– m1x dx)dx

It was proved in 5, that in the d.e. (6), the order of the factors is immaterial, if m1
and m2 are constants.
i.e., (D – m1) (D – m2) y – (D – m2) (D – m1) y = Q
1 1
 y= Q= Q,
( D – m2 )( D – m1 ) ( D – m1 )( D – m2 )

If m1 and m2 are constants.


1 1
Hence the order of factors and in the result (7) is also immaterial.
( D – m2 ) ( D – m1 )

12. To find the particular integral of the equation


f(D)y = Q ...(1)
1
We define Q to the function such that when f(D) operates on it the result is Q.
f ( D)

1
 The symbol Q stands for a solution of the differential equation (1).
f ( D)

Let f(D) = (D – m1) (D – m2) ... (D – mn)


1 1
 P.I. = Q = Q
f ( D) ( D – m1 )( D – m2 ) ( D – mn )

1  1 
=  Q
(D – m1 )(D – m2 ) (D – mn –1 )  D – mn 

1
= emn x  Qe– mn x dx
( D – m1 )( D – m2 ) ( D – mn –1 )

56
We write Q = e m x  Qe – m x dx
n n

1
P.I. = Q.
( D – m1 )( D – m2 ) ( D – mn )

1
= emn–1 x  Qe– mn–1 x dx
( D – m1 )( D – m2 ) ( D – mn –1 )

Proceeding in this manner we will after a finite number of steps obtain the particular
integral.
1 1
Theorem 3. ( AQ) = A (Q)
f (D) f (D)

where A is a given constant, and Q is any function of x, and


f(D) = Dn + p1 Dn–1 + p2 Dn–2 + ... + pn–1 D + n,
p1, p2, p3, ..., pn being constants.
1
Proof: Let y = AQ
f ( D)

then by definition
f(D) y = AQ
= Q where A  0 where A  0
(on dividing both sides by the constant A)
1 n
or ( D + p1 D n –1 + p2 D n –2 + + pn –1 D + pn ) y = Q
A

y y y y


or Dn   + p1 Dn –1   + + pn –1 D   + pn   = Q
 A  A  A  A
y
or f ( D) =Q
A
y 1
 = Q
A f (D)

1
i.e., y= A Q.
f ( D)

12. Alternative Method for finding the particular integral


1
To find Q where Q is a function of x.
f ( D)

Let f(D) = (D – m1) (D – m2) ... (D – mn)

57
By resolving into partial fractions,
1 A1 A2 An
= + + +
f ( D) D – m1 D – m2 D – mn

 Particular integral
1
= Q
f ( D)

 A1 A2 An 
=  + + + Q
 D – m1 D – m2 D – mn 

= A1e m x  Qe – m x dx + A2 e m x  Qe – m x dx
1 1 2 2

+ + An e mn x  Qe – mn x dx

Note: The method given above is a general one which can be used to obtain the
particular integral in any given case. Shorter methods depending upon the form of the
function Q will be discussed in the following section.
As we require only a particular integral we may not introduce arbitrary constants in
carrying out the integrations in the method explained above.
13. The Complete Primitive of Linear Equation with Constant co-efficients
The differential equation is given by
dny d n –1 y d n –2 y dy
+ p1 + p2 + + pn –1 + pn y = Q ...(1)
dx n dx n –1 dx n –2 dx
or f(D)y = Q
where p1, p2, p3, ... pn–1, pn are given constants and Q is a function x. The complete
integral of the d.e. (1) is by 4 and 10, the sum of the Complementary Function (C.F.)
and the Particular Integral (P.I) and is therefore given by
1
y = [c1 f1 ( x) + c2 f2 ( x) + + cn fn ( x)] + Q ...(2)
f (D)

where the C.F. is the general solution of the equation (1) with Q = 0 and can be found out
1
by articles 6, 7, 8 and 9, the P.I., Q can be obtained by 12.
f ( D)

d 2 y dy
Solved Example 1. Solve + – 2y = 2x ...(1)
dx 2 dx

58
Solution : The given differential equation can be written as
(D2 + D – 2)y = 2x.
or f(D)y = Q ...(2)
where f(D) = D2 + D – 2, Q = 2x
The complementary function is the general solution of the d.e.
f(D)y = 0
whose auxiliary equation is
f(m) = m2 + m – 2 = 0
f(m) = 0  (m + 2) (m – 1) = 0
The auxiliary equation has two roots
m = – 2 and m = 1
The complementary function is given by
C.F. = c1e–2x + c2ex
c1, c2 being arbitrary constants.
The particular integral is given by
1
P.I. = Q
f ( D)

1
= (2 x )
D +D–2
2

1
=2 x
(D + 2)(D – 1)

2 1 1 
= – x
3  D + 2 D – 1 

2 1 2 1
= x– x
3 D –1 3 D+2
2 x 2
= e  xe – x dx – e –2 x  xe2 x dx
3 3 by 12

2 x
= e – xe – x +  e – x dx 
3  

2  e2 x e2 x 
– e –2 x  x  – dx 
3  2 2 

59
2 x 2 1 1 
= e [– xe– x – e– x ] – e2 x  xe2 x – e2 x 
3 3 2 4 
2 2 1 1
= – x– – x+
3 3 3 6
1  1
= –x – = – x + 
2 2  
 The complete solution is given by
 1
y = c1e–2 x + c2 e x –  x + 
2  
d2y dy
Solved Example 2. Solve 2
– 5 + 6 y = e4 x .
dx dx
Solution: The differential equation (1) is of the form
(D2 – 5D + 6)y = e4x
Its auxiliary equation is
5(m) = m2 – 5m + 6 = 0
or (m – 3) (m – 2) = 0
 m = 3, 2
Hence the complementary function is
C.F. = c1e2x + c2e3x
c1 and c2 being arbitrary constants.
The particular integral is given by
1 1
P.I. = Q= 2 e4 x
f (D) D – 5D + 6
1
= e4 x
(D – 2)(D – 3)
 1 1 
= –  e4 x
D–3 D–2
1 4x 1
= e – e4 x
D–3 D–2
= e3 x  e –3 x dx – e2 x  e –2 x e 4 x dx

= e3 x  e x dx – e2 x  e2 x dx

e2 x
= e3 x e x – e 2 x
2

60
1 1
= e3 x – e 4 x = e 4 x ...(3)
2 2
 The complete solution of the d.e. (1), by (2) and (3), is
1
y = c1e2 x + c2 e3 x + e 4 x
2
1
14. We shall now discuss some shorter method to find the particular integral Q
f ( D)
when the R.H.S. of the d.e. (1) of 13 takes the following special forms
(i) eax where a is any constant.
(ii) xm where m is a positive integer.
(iii) sin (ax + b) and cos (ax + b)
where a and b are given constants.
(iv) eaxu, where u is any function of x and a is a given constant.
(v) x u, where u is any function of x.
1 ax
15. Particular Integral e
f ( D)
1 ax 1 ax
Theorem 4. e = e ...(1)
f (D) f (a)
provided the constant a is such that
f(a)  0 ...(2)
where f (D) = Dn + p1Dn–1 + p2Dn–2 ... + pn–1D + pn
p1, p2, p3 ..., pn being constants.
Proof: By successive differentiation, we have
d ax 
De ax = (e ) = ae ax 
dx

d 2 ax 2 ax 
D e = 2 (e ) = a e 
2 ax

dx

d 2 ax 3 ax 
D e = 3 (e ) = a e 
3 ax
...(3)
dx

....................................... 
dn 
D n e ax = n (e ax ) = a n e ax 
dx 


 f(D) eax = (Dn + p1Dn–1 + p2 Dn –1 + + pn –1D + pn )eax

= (Dn eax + p1Dn –1eax + p2 Dn –2eax + + pn –1Deax + pneax

61
= [an + p1an –1 + p2 an –2 + + pn –1a + pn ]eax

of f(D)eax = f(a) eax ...(4)


By definition of inverse operators, we get
1
eax = f (a)eax
f (D)

1 ax
= f (a) e ...(5)
f (D)

1
By bringing out the constant quantity f(a) outside the operator 1
f ( D)

Since f(a)  0, therefore from (5) by dividing both sides by f(a), we get
1 ax 1 ax
e = e ...(6)
f ( D) f (a)

The condition (2) viz., f(a)  0 implies that a is not a root of the equation
f(m) = 0
Case : If ‘a’ is a root of the equation
f(m) = 0
which is the auxiliary equation corresponding to the d.e. (1) of 13, viz., f(D) y = Q, then
the formula (1) is not valid.
To find the particular integral in this case the following method used.
Let a be a single root of the auxiliary equation (6), then
f(m) = (m – a) 1(m) ...(7)
where f1(a)  0
 f(D) = (D – a) 1(D)
1 1 1
  eax =  eax
f (D) ( D – a) 1 ( D)

1 1
=   eax [by (1) Now 1(a)  0]
( D – a) 1 (a)

1 1
=  eax
1 (a) D – a

1 ax – ax ax
1 (a) 
= e e e dx

62
 1 
 D – m Q = e  Qe dx 
mx – mx

 
1
= (a)eax  dx
1

1 ax 1
 e = xeax
f (D) 1 (a)

where  (a) is given by (7).


In a similar way let ‘a’ be double root of the auxiliary equation (6), then
f(m) = (m – a)2 2(m)
where 2(a)  0
Here f(D) = (D – a)2 2(D)
1 ax 1 1
 e =   eax
f ( D) ( D – a) 2 ( D)
2

1 1
=   eax
( D – a) 2 (a)
2

1 1
=  eax
2 (a) ( D – a)2

1 1  1 
=  eax
2 (a) ( D – a)  D – a 

1 1 1 – ax – ax ax
=  ( xeax ) = e  e ( xe )dx
2 (a) D – a 2 (a)

1 ax
2 (a) 
= e xdx

1 ax 1 x2
i.e., = e = eax ...(10)
f ( D) 2 (a) 2

[where 2(a) is given by (9)]


Continuing as before let us suppose that a is a triple root of the equation (6), then
(m) = (m – a)3 3(m) ...(11)
where 3(a)  0
Here f(D) = (D – a)3 3(D)

63
1 ax 1 1
 e = eax
f ( D) ( D – a) 3 ( D)
3

1 1 ax
= e
( D – a) 3 (a)
3

1 1 ax
= e
3 (a) 3 (a)

1 1
= eax
3 (a) ( D – a)3

1 1  1 ax 
=
3 (a) ( D – a)3  ( D – a)2 e 
 

1 1  ax x 2 
= e 
3 (a) ( D – a)3  2

1 ax – ax  ax x 2 
3 ( a) 
= e e e  dx
 2 

1 eax 2
(a) 2 
= x dx

1 eax x 3
=
3 (a) 2 3

1 ax 1 x 3 ax 1 x 3 ax
 e = e = e ...(12)
f ( D) 3 (a) 3 3 3

where 3(a) is given by (11).


Generalizing, we prove in a similar way that if the auxiliary equation (6) has a root ,
repeated r times, then
f(m) = (m – a)r r(m) ...(13)
[where r(a) = 0]
1 ax 1 xr
and then e =   eax
f ( D) r (a) r

where r(a) is given by (13).

Solved Example 3. Solve

64
d2y dy
2
+ 5 + 6 y = e2 x
dx dx
Solution: The differential equation can be written in the form
f(D) y = e2x
where f(D) = D2 + 5D + 6
 its auxiliary equation is given by
f(m) = m2 + 5m + 6 = 0
or (m + 2) (m + 3) = 0
 The roots of the auxiliary equation are m = –2, = –3.
The complementary function of the equation (1) is given by
C.F. = c1e–2x + c2e–3x
The Particular Integral is given by
1 2x 1
P.I. = e = 2 e2 x
f (D) D + 5D + 6

1 1
= e2 x = e2 x
2 + 5.2 + 6
2
20

 1 ax 1 
 e = (a)eax provided f (a)  0 
 f ( D) f 

The complete primitive of the given differential equation is given by


y = c1e–2x + c2e2x
Solved Example 4. Solve
(D2 – 3D + 2) = ex + e2x
Solution: The auxiliary equation is
m2 – 3m + 2 = 0
 m2 – 2m – m + 2 = 0
 m(m – 2) – (m – 2) = 0
 (m – 1) (m – 2) = 0
 m = 1, 2
 C.F. = c1ex + c2e2x

65
1
P.I. = (e x + e2 x )
(D – 1)(D – 2)
1 1
= ex + e2 x
(D – 1)(D – 2) ( D – 1)( D – 2)
1 1
= ex + e2 x
(D – 1)(1 – 2) (2 – 1)( D – 2)
1 1
= – ex + e2
(D – 1) ( D – 2)
= – e x  e – x e x dx + e2 x  e –2 x e 2 x dx

= – e x  dx + e2 x  dx

= –xex + xe2x
Hence the required solution is y = c1ex + c2e2x – xex +xe2x
Solved Example 5. Solve
d3y
3
– y = (e x + 1) 2
dx
Solution: The differential equation may be written in the form
(D3 – 1) y = (ex + 1)2 = e2x + 2ex + 1 ...(1)
Hence f(D) = D3 – 1 ...(2)
Its auxiliary equation is
m3 – 1 = 0
or (m – 1) (m2 + m + 1) = 0

–1  1 – 4 1 3
 m = 1 and m = = – i 
2 2 2
 The auxiliary equation has a real root 1 and a pair of conjugate complex roots

1 3
– i
2 2
 3 3 
C.F. = c1e x + e – x /2  c2 cos x + c3 sin x ...(3)
 2 2 

The Particular Integral of the d.e. (1), is given by


1
P.I. = [e2 x + 2e x + 1]
f (D)

66
1
= (e 2 x + 2 e x + e 0 )
3
D –1
1 1 1
= e2 x + 3 2e x + 3 e0 x
D3 – 1 D –1 D –1
1 1 1
= e2 x + 2 3
3
ex + 3 e0 x
D –1 D –1 D –1
The first and the third terms of (4) can be evaluated by the formula in (1) or the last
article and are equal to
1 1 1 2x 1 0x
= e2 x + 3
3
2e 0 x = 3 e + e
D –1 D –1 2 –1 0 –1
1 2x
= e –1
7
1
The formula (1) of the last article is not applicable to the middle term viz., 2. 3
ex
D –1
since 1 is root of the auxiliary equation (2). Now
1 1 1
ex =  2 ex
3
D –1 D –1 D + D +1
1 1
=  ex
D –1 1+1+1
1 1 x
=  e
D –1 3
1 1
=  ex
3 D –1
1 1 1
= e x  e – x e x dx = e x  dx = xe x
3 3 3
1 2
 2 3
e x = xe x ...(6)
D –1 3
 The particular integral from (5) and (6), is given by
1  2
P.I. =  e x – 1  + xe x ...(7)
7  3
From (3) and (7) the complete primitive of the d.e. (1) is
 3 3  1 2x 2 x
y = c1e x + e – x /2  c2 cos x + c3 sin x  + e + xe – 1
 2 2  7 3

67
1
16. Particular Integral x m where m is positive integer
f ( D)

1
The general method consists in expanding in ascending integral power of D,
f ( D)
regarding D as an algebric quantity and then operate each term of the expansion on xm.
The sum of the results of these operations is the required value. Since the (m+1)th and
higher order derivative of xm are zero it is sufficient to retain terms only upto the mth
1
degree in D in the expansion of . The following examples will illustrate the
f ( D)
procedure.
1
Example 1. Evaluate (x3 )
D+3

1 1
Solution: We have =
D+3  D
3 1 + 
 3
–1
1 D
=  1+ 
3 3

1  D (–1)(–2) D 2 
= 1 – +  
3 3 1.2 9 

(–1)(–2)(–3) D 3 
= +  + higher powers of D 
1.2.3 27 

1  D D3 D3 
+ 1 – + – 
3 3 9 27 

1 1  D D2 D3  3
 x3 = 1 – + – x
D+3 3 3 9 27 

1 3 1 1 1
= x – D( x 3 ) + D 2 ( x 3 ) – D3 ( x 3 )
3 9 27 81
1 3 1 1 1
= x –  3x 2 +  6 x –  6
3 9 27 81
1 3 1 2 2 2
= x – x + x–
3 3 9 27

 d 3
=  D( x 3 ) = ( x ) = 3x 2
 dx

68
d2 3
D2(x3) = ( x ) = 3.2 x = 6 x
dx 2

d3 3 
D3(x3) = 3
( x ) = 3.2.1 = 6 
dx 

Here we have retained terms only upto D3 as we are operating on x3.


1
Example 2. Evaluate (x2 )
D – 3D + 2
2

Solution : We have
1
D – 3D + 2
2

1
=
 3 D2 
2 1 – D – 
 2 2 
–1
1 3 D2 
= 1 –  D + 
2 2 2 

1 3 
2
D2   3 D2 
= 1 +  D – +
  D –  
2 2 2  2 2  

1 3 D2 9 2 
= 1 + D – + D + higher powers D 
2 2 2 4 

1
Now (x2 )
D – 3D + 2
2

1 3 D2 9 2  2
= 1 + D – + D x
2 2 2 4 

1 3 7 
= 1 + D + D2  x 2
2 2 4 

1 2 3 7 
=  x + 2 x +  2
2 2 4 

1 2 3 7
= x + x+
2 2 4
Example 3. Solve the differential equation
d3y d2y dy
3
+ 3 2
+2 = x2
dx dx dx

69
Solution: The differential equation can be written in the form
f(D) y = x2
where f(D) y = D3 + 3D2 + 2D.
 Its auxiliary equation is given by
f(m) = m3 + 3m2 + 2m = 0
i.e., m(m2 + 3m + 2) = 0
i.e., m(m + 2) (m + 1) = 0
 The roots of the auxiliary equation are 0, –2, –1. Hence the complementary
function of the given equation is
C.F. = c1 + c2e–2x + c3e–x
The particular integral is given by
1
P.I. = x2
D + 3D + 2 D
3 2

1
= x2
 D 2 + 3D 
2D  
 2 
x1
1  D 2 + 3D  2
=   x
2D  2 

1   D2 + 3D   D2 + 3D  
2

= 1 –  +   x2
2D   2   2  

1  3D 7 2  2
= 1– + D + x
2 D  2 4 
[Here we have retained terms only upto D4 as we are operating on x2]
1  2 3 7 
=  x –  2 x +  2
2D  2 4 
1  2 7
=  x – 3x + 
2D  2
1  2 7
= 
2  x – 3x +  dx
2
1  x3 3x 2 7 
=  – + x
2 3 2 2 
x
= (2 x 2 – 9 x + 21)
12

70
 1 2 
 Note : D x =  x dx and so on  and so on
2

 
 The complete primitive of the given differential equation is
y = C.F. + P.I.
x
= c1 + c2 e –2 x + c3e – x + (2 x 2 – 9 x + 21)
12

Exercise
Solve the differential equation
d2y dy
1. 2
+ 8 + 25 y = 50.
dx dx
d2y dy
2. 2
– 2k + k 2 y = ex .
dx dx
d2y dy
3. 2
+ 2 + y = e x /2 + e – x /2 .
dx dx
d3y
4. 3
+ y = 3 + e – x + 5e –2 x /3 .
dx
d2y dy
5. 9 2
+ 12 + 4 y = e –2 x /3 .
dx dx
d2y
6. 9 – 4 y = x.
dx 2
d2y dy
7. 2
+ 2 + y = 2 x + e –2 x .
dx dx
8. (D2 + 2D + 1) y = 2 x + x 2
d2y dy
9. 2
– 7 + 6 y 2 e3 x .
dx dx
dy
Given that y = 1, = 0 when x = 0.
dx

Answers
1. y = e–4x (c1 cos 3x + c2 sin 3x) + 2.
ex .
2. (c1 + c2 x)ekx +
(k – 1)2
4
3. y = (c1 + c2 x )e – x + e x /2 + 4e – x /2 .
9
 3 3  5 xe – x
4. y = c1e – x + e x /2  c2 cos x + c3 sin x  + 3 + e2 x + .
 3 2  9 3

71
1 2 –2 x /3
5. y = (c1 + c2 x )e –2 x /3 + xe
18
1 1
6. y = c1e2 x + c2 e–2 x –  x 2 +  .
4 2 
7. y = (c1 + c2 x)e + 2 x – 4 + e .
–x –2 x

8. y = (c1 + c2 x)e– x + x 2 – 2 x + 2.
7 1 6 x 1 3x
9. y = e x – e – e .
5 15 3

72
Lesson 5
LINEAR DIFFERENTIAL EQUATIONS WITH CONSTANT
COEFFICIENTS (Non-Homogeneous)-II

17. To find the Particular Integral


1 1
sin(ax + b) and cos(ax + b)
f (D) f (D)

Let us first consider the case when f(D) consists only even powers of D and let
f(D) = f(D2)
= (d2)n + p1(D2)n–1 + p2(D2)n–2 + ... pn–1D2 + pn
where p1, p2, p3, ..., pn are constants.
Theorem 5. If f(D) contains only even powers of D and is represented by (D2) as in (1)
then the particular integrals
1 1
sin(ax + b) = sin(ax + b)
f (D) (D2 )

1
= sin(ax + b)
(– a2 )

1 1
and cos(ax + b) = cos(ax + b)
f (D) (D2 )

1
= cos(ax + b)
(– a2 )

Provided f(–a2)  0.
d
Proof: D sin(ax + b) = sin(ax + b) = a cos(ax + b)
dx
D cos(ax + b) = –a sin (ax + b)
Again D sin(ax + b) = D[D sin (ax + b)] = D[a cos (ax + b)]
= –a2 sin (ax + b)
Similarly
D2 cos (ax + b) = –a2 cos (ax + b)
Again D4 sin (ax + b) = [D2 sin (ax + b)]
= D2 [a2D2 sin ax + b)]

73
= –a2D2 sin (ax + b)
= (–a2)2 sin (ax + b)
Similarly D4 cos (ax + b) = (–a2)2 cos (ax + b).
Again D6 sin (ax + b) = D2D4 sin (ax + b)
= (–a2)2D2 sin (ax + b)
= (–a2)2D2 sin (ax + b)
= (–a2)3 sin (ax + b)
and similarly D6 cos (ax + b) = (–a2)3 cos (ax + b).
Generalizing, we have
D2n sin (ax + b) = (–a2)n sin (ax + b) ...(3)
and D2n cos (ax + b) = (–a2)n cos (ax + b)
 Form (1)
(D2) sin (ax + b) = [(D2)n + p1 (D2)n–1 + p2(D2)n–2 + ... + pn–1D2 + pn)] sin (ax + b)
= (D2)n sin (ax + b) + p1(D2)n–1 sin (ax + b) = p2(D2)n–2 sin (ax + b) + ... + pn–1D2 sin
(ax + b) + pn sin (ax + b)
= [(–a2)n + p1(–a2)n–1 + p2(–a2)n–2 + ... + pn–1 (–a2) + pn] sin (ax + b)
with the help of the above relations, we obtain
(D2) sin (ax + b) = (–a2) sin (ax + b) ...(4)
1
Let (–a2)  0, then operating on both sides of (4) with we have
( D 2 )

1 1
(D2 )sin(ax + b) = (– a2 )sin(ax + b)
(D )
2
(D2 )

1
or sin(ax + b) = (–a2) sin(ax + b)
(D2 )

Since (–a2)  0, on dividing both sides by f(–a2), we get


1 1
sin(ax + b) = sin(ax + b)
(D )
2
(D2 )
In particular if b = 0, then
1 1
sin ax = sin ax ...(5)
( D )
2
(– a2 )

where (–a2)  0.

74
Similarly with the help of the result (3), it can be proved that
(D2) cos (ax + b) = [(–a2)n + p1(–a2)n–1 + p2(–a2)n–2 + ... pn–1(–a2)
+ pn] cos (ax + b)
= (–a2) cos (ax + b)
 Assuming (–a2)  0, as before we get
1 1
cos(ax + b) = cos(ax + b)
(D )
2
(– a2 )

In particular if b = 0 then
1 1
cos(ax) = cos(ax)
(D )
2
(– a2 )

where (–a2)  0.
Example 1. Solve
d2y 1
2
– 4 y = 2sin x
dx 2
Solution: The differential equation is
x
(D2 – 4)y = 2 sin
2
1
or f(D)y = 2sin x
2
where f(D) = D2 – 4
 Its auxiliary equation is m2 – 4 = 0
 m =+2
The complementary function of the given d.e. is given by
C.F. = u = c1e2x + c2e–2x
The particular integral is given by
1 x
v = P.I. = 2
2sin
D –4 2
1 x
=2 2
sin
D –4 2
 1 x 8 x
= 2 2
sin = – sin 
 –(1 / 2) – 4 2 17 2

 1 1 
 sin ax = sin ax, If (– a2 )  0 
 ( D )
2
(– a )
2

75
The complete primitive of the given d.e. therefore, is
8 x
y = c1e2 x + c2 e –2 x – sin
17 2
Case: Now we consider the case when f(D) containing both even and odd powers
of D.
Let f(D) = Dn + p1 Dn–1 + p2Dn–2 + ... pn–1D + pn
which can be broken up into two parts containing even and odd powers of D respectively,
so
f(D) = 1(D2) + D2(D2) ...(7)
where 1(D2) and 2(D2) are functions containing even powers of D and therefore
D 2(D2) contains odd powers of D.
Thus 1(D2) contains terms involving even powers of D, and those in D2(D2) are
terms involving odd powers of D. For example:
f(D) = D5 + 3D4 + 4D3 + 6D2 + 5D + 8
= (3D4 + 6D2 + 8) + D (D4 + 4D2 + 5)
= 1(D2) + D2(D2).
so that in this case f(D) = 1(D2) + D2(D2).
1 1
Let sin(ax + b) = sin(ax + b)
f (D) ( D ) + D2 ( D2 )
2

In the denominator of the R.H.S. of (A), we replace 1(D2) and 2(D2) by 1(–a2)
and 2(–a2) respectively and then calculate the function
1
sin (ax + b)
1 (– a ) + D2 (– a2 )
2

1 (– a2 ) – D2 (– a2 )
= sin (ax + b)
[1 (– a2 ) + D2 (– a2 )][1 (– a2 ) – D2 (– a2 )]

1 (– a2 ) – D2 (– a 2 )
= sin (ax + b)
[1 (– a2 )]2 – D 2 [2 (– a 2 )]2

1 (– a2 ) – D2 (– a2 )
= sin (ax + b)
[1 (– a2 )]2 – a2 [2 (– a2 )]2

[By replacing D2 by – a2 in the denominator]

76
1 (– a2 )sin(ax + b) – 2 (– a 2 ) D sin(ax + b)
=
[1 (– a2 )]2 – a2 [ 2 (– a2 )]2

1 (– a2 )sin(ax + b) – a2 (– a2 )cos(ax + b)


=
[1 (– a2 )]2 – a2 [2 (– a2 )]2

 when f(D) = 1(D2) + D 2(D2), we get


1 1
sin(ax + b) = sin(ax + b)
f (D) ( D ) + D2 ( D2 )
2

1
= sin (ax + b)
1 (– a ) + D 2 (– a2 )
2

Similarly it can be shown that


1 1
cos(ax + b) = cos (ax + b)
f (D) 1 (D ) + D 2 (D2 )
2

1
= cos(ax + b)
1 (– a ) + D 2 (– a2 )
2

Solved Example 2. Solve the equation


d 3 y d 2 y dy
+ – – y = cos 2x
dx 3 dx 2 dx
Solution: The differential equation is of the form
[D3 + D2 – D – 1]y = cos 2x
i.e. f(D)y = cos 2x
where f(D) = D3 + D2 – D – 1
= (D + 1) (D2 – 1) = (D – 1) (D + 1)2
The auxiliary equation of the d.e., is
(m – 1) (m + 1)2 = 0
 The auxiliary equation has a single root 1 and a double root –1. The complementary
function of the d.e. is given by
C.F. = u = c1ex + e–x(c2 + c3x)
Now the Particular Integral is
1 1
P.I. = cos2 x = 3  cos2 x
f (D) D + D2 – D + 1

77
1
= cos2 x
(D + 1)(D2 – 1)

1
= cos2 x
(D + 1)(–4 – 1)

1
= – cos2 x
5(D + 1)

(D – 1)
= – cos2 x
5(D2 + 1)

1 (D – 1)
= –  cos2 x
5 (–4 – 1)

1
= ( D – 1)cos2 x
25
1
= [ D cos2 x – cos2 x ]
25
1
= [–2sin 2 x – cos2 x ]
25
The complete primitive of the given differential equation is
2 1
y = c1e x + e – x (c2 + c3 x ) – sin 2 x – cos2 x
25 25
18. The particular integral in exceptional cases
1 1
sin ax and cos ax
f (D) f (D)

The method indicated in the last article to find the particular integrals
1 1
sin ax and cos ax fails if (D2 + a2) is a factor of f(D), i.e., if
f (D) f (D)

f(D) = (D2 + a2)  (D)


where (D) does not contain the factor (D2 + a2).
In this case, we have
1 1
sin ax = sin ax
f ( D) (D + a2 )(D)
2

1  1 
= sin ax 
( D)  D2 + a2  ...(2)

78
1 1
In order to evaluate (2), we will determine first sin ax and then apply
D +a22
( D)
the result.
1
Now sin ax cannot be evaluated with the help of (2) of 17, since the
D + a2
2

denominator (D2 + a2) vanishes after substitution of (–a2) for D2.


We therefore write
1 1
sin ax = sin ax
D +a
2 2
(D + ia)( D – ia)

1  1 1 
= – sin ax
2ia  D – ia D + ia 

1
by breaking into partial fractions.
D + a2
2

1 1  1 1 
sin ax =  sin ax – sin ax 
D +a
2 2
2ia  D – ia D + ia 

=
2ia
(
1 iax iax
e  e sin ax dx – e – iax  eiax sin ax dx )
 1 
 Q = emx  e– mx Qdx 
 D–m 

1  iax – iax  eiax – eiax  iax  e


iax
– eiax  
2ia     2i dx  dx 
iax
= e e   dx – e e
 2i  

=
4a
(
–1 iax
e  (1 – e(–2iax ) dx – e – iax )  (e2iax – 1)dx )
–1  iax   – iax  e2 
iax
e –2 iax
= e  x +  – e  – x 

4 a   2ia   2ia 

– x – iax 1
=
4a
( ) (
e + e – iax + 2 eiax – e – iax
8ia
)
x 1
= cos ax + 2 sin ax
2a 4a
1
Again in similar way let us evaluate cos ax
D + a2
2

1 1  1 1 
Now cos ax = – cos ax
D +a
2 2
2ia  D – ia D + ia 

79
1  1 1 
=  cos ax – cos ax 
2ia  D – ia D + ia 
1  iax – iax
e e cos ax dx – eiax  eiax cos ax dx 
2ia  
= 

1  iax  eiax + e – iax  ia x  e


iax
+ e – iax  
  2
iax
= e   dx – e e dx 
2ia   2   

1  iax
= e  (1 + e –2 iax )dx – e – iax  (eiax + 1)dx 
4ia  

1  iax  e– iax  iax  e2iax 


= e  x – –e  + x 
4ia   2ia   2ia 

x 1
= sin ax + 2 cos ax
2a 4a
1 x sin ax 1
 cos ax = + 2 cos ax ...(4)
D +a
2 2
2a 4a
Solved Example 3. Solve the differential equation
d 2 y dy
+ + y = sin 2x
dx 2 dx
Solution: The given equation is
(D2 + D + 1)y = sin 2x
Here auxiliary equation is m2 + m + 1 = 0

–1  –3 –1  i 3
It has roots =
2 2
The complementary function of the given d.e. is
 3 3 
e – x /2 c1 cos x + c2 sin x
 2 2 

where c1, c2, are arbitrary constants.


Particular Integral
1
= (sin 2 x )
D + D +1
2

1
= (sin 2 x )
–4 + D + 1

80
 1 1 
 ( D 2 ) sin ax = (– a2 ) sin ax  where f(–a )  0
2
 

1
= (sin 2 x)
D–3
D+3
= (sin 2 x )
D2 – 9
( D + 3) 1
= (sin 2 x ) = – ( D + 3)sin 2 x
–4 – 9 13
2 3
= – cos2 x – sin 2 x
13 13
The complete solution of the given d.e. is,
 3 3 
y = e – x /2 c1 cos x + c2 sin x
 2 2 

1
– (2 cos2 x + 3sin 2 x )
13
Solved Example 4. Solve
(D2 – 1)y = cos3 x
Solution: The auxiliary equation m2 – 1 = 0 has two roots + 1.
The complementary function is c1ex + c2e–x.
1
Particular Integral = 2
(cos3 x ) .
D –1
First we express cos3x in terms of cos 3x and cos x, using the formula
cos3x = 4 cos3x – 3 cos x
1
 cos3x =
4
cos3 x + 3cos x 
1 1 1 
 = 2
(cos3 x) = 2  (cos3 x + 3cos x) 
D –1 D –14 
1 1 3 1
=  2 (cos3 x ) + (cos x )
4 D –1 4 D2 – 1

1 1  3 1 
=   cos3x +  cos x
4  –9 – 1  4  –1 – 1 

1 3
= – cos3 x – cos x
40 8

81
 The complete solution of the d.e. is given by
1 3
y = c1e x + c2 e – x – cos3 x – cos x
40 8
19. The following two theorems gives us formula which will be helpful in finding the
particular integral of the differential equation of the form f(D)y = Q.

Q = e ax u 
where 
Q = xu 
where us is any given function of x and a any given constant. i.e., Q stands for the product
of the two function (i) u and eax (ii) u and x...
Theorem 6. To show that
1 1
(eax u) = eax u
f (D) f (D + a)

where u is any given function of x.


Proof. If v is any function of x, then
d ax
D(eax v) = (e v)
dx
dv
= e ax + ae ax v
dx
= eax (Dv + av)
= eax(D + a)v
i.e., D(eax v) = eax(D + a)v
Again D2(eax v) = D[(Dax + a)v]
= D[eax (D + a)v] by (2)
= D[eaxv1]
where v1 = (D + a)v
D2(eaxv) = D(eaxv1) = eax(D + a)v1 by (2)
= eax (D + a) (D + a)v [ v1 = (D + a)v]
D2(eaxv) = eax(D + a)2v
Again D3(eaxv) = D[D2(eaxv)]
= D[eax(D + a)2 v]
= D[eax[(D + a)2 v]]

82
= eax (D + a) . [(D + a)2 v)] by (2)
 D3(eaxv) = eax (D + a)3v
Similarly D4(eaxv) = eax(D + a)4v
and D5(eaxv) = eax(D + a)5v
In general Dr(eaxv) = eax(D + a)r . v, ...(2)
where r is a positive integer.
In (5) we notice that in Dr(eax v) when the factor eax is brought outside the operator Dr,
D is changed to D + a.
Let f(D) = Dn + p1Dn–1 + p2Dn–2 + p3Dn–3 + ... + pn–1D + pn ...(6)
 f(D) (eax v) = [Dn + p1Dn–1 + p2Dn–2 + ... + pn–1D + pn](eax v)
= Dn(eaxv) + p1Dn–1(eaxv) + p2Dn–2(eaxv) + ... + pn–1D(eaxv) + pn(eaxv) ...(7)
by putting r = 1, 2, 3, ... n is (5) and substituting in (6), we get
f(D) [eaxv] = eax = (D + a)n v + p1eax (D + a)n–1 + ... + pn–1eax (D + a)v + pneaxv
= eax [(D + a)nv + p1(D + a)n–1v + p2(D + a)n–2v + ... + pn–1(D + a)v + pn v]
= eax f(D + a)v
 f(D)(eaxv) = eaxf(D + a)v
i.e. in f(D) (eaxv) when the factor eax is brought outside the operator f(D), is changed
D + a.
Let f(D + a)v = u
By definition of inverse operator we have from (8)
1
v= u
f ( D + a)
The left hand side of (7) is
= f(D) (eaxv)
 1 
= f ( D) eax v u
 f ( D + a) 

The right hand side of (7), as on using (8), is


= eax f(D + a)v
= eaxu
 From (7), we have

83
 1 
f ( D) e ax u  e ax u
 f ( D + a) 

1
Now by applying the operator on both sides of (10), we have
f ( D)

1  1  1
f ( D) e ax u (e ax u)
f ( D)  f ( D + a)  f ( D )

1 1
 (eax u) = eax
f (D) (D + a)

1 1 ,
Hence we conclude that in (eax u) when the factor eax is brought outside the D
f (D) f ( D)
is changed (D + a)
Solved Example 5. Solve
d2y dy
– 2 + 4y = ex cosx
dx dx

Solution: The differential equation can be written is


(D2 – 2D + 4)y = ex cosx
or f(D)y = ex cosx
where f(D) = D2 – 2D + 4
 Its auxiliary equation is

f(m) = m2 – 2m + 4 = 0
2  4 – 16
m= =1 i 3
2

The auxiliary equation has roots 1  i 3.


The complementary function of the differential equation is given by

u = e x c1 cos 3x + c2 sin 3x 

Again the particular integral is given


1
v=  e x cos x
D – 2D + 4
2

1
= ex cos x
(D + 1) – 2(D + 1) +
2

84
 1 1 
 (eax u) = eax u
 f ( D) f ( D + a) 

1
= ex cos x
D + 2D + 1 – 2D – 2 + 4
2

ex
= cos x
D2 + 3

ex  1 1 
= cos x  cos ax = cos ax if (– a 2 )  0 
–1 + 3  ( D )
2
(– a )
2

1 x
= e cos x
2
 The complete solution of the differential equation is
1
y = e x c1 cos 3 x + c2 sin 3 x  + e x cos x
2
Solved Example 6. Solve (D + 2) (D – 1)3 y = ex.
Solution: The auxiliary equation of the differential equation is
(m + 2) (m – 1)3 = 0
 m = –2 and (repeated thrice)
The auxiliary equation has a single root –2 and a triple root 1.
The complementary function is given by
= (c1 + c2x + c3x2)ex + c4e–2x
The particular integral is given by
1
v= ex
(D – 1) (D + 2)
3

1 1 x 
= e
( D – 1)3  3 

1 1  1 ax ax 1 
= ex   e ue u
(D + 1) – 1 f ( D + a) 
3
3  f ( D)

1 x 1
= e  31
3 D
1 x 1 1
= e  2  1
3 D D

85
1 x 1 1 1
= e  2  dx = e x  2 x
3 D 3 D
1 x 1 1
= e   x
3 D D
1 x 1 1 1  x2 
= e   x dx = e x   
3 D 3 D 2 

1 x 2 1
= e  x dx = e x x 3
6 18
The complete solution is given by
1 x 3
y = u + v = (c1 + c2 x + c3 x 2 ) + c4 e –2 x + e x
18
Solved Example 7. Solve (D2 – 5D + 6)y = e2x sin 3x.
Solution: The auxiliary equation of the differential equation is
m2 – 5m + 6 = 0
(m –2) (m – 3) = 0
m = 2, 3
The complementary function is given by
C.F. = C1e2x + C2e3x
The particular integral is given by
1
P.I. = e2 x sin 3 x
D – 5D + 6
2

1
= e2 x sin3x
(D + 2) – 5(D + 2) + 6
2

1
= e2 x 2
sin 3 x
3 –D
1
= – e2 x sin 3 x
D+9
D–9
= – e2 x sin 3 x
(D – 9)( D + 9)
D–9
= – e2 x sin 3 x
D 2 – 81
D–9
= – e2 x sin 3 x
–9 – 81
e2 x
= ( D – 9)sin 3 x
90

86
e2 x
= (3cos3 x – 9sin 3 x )
90
Hence the complete solution is
1 2x
y = C1e2 x + C2 e3 x + e (cos3 x – 3sin 3 x )
30
Theorem 7. To show that
1  1  1
xu =  x – f ( D)  u
f ( D)  f ( D)  f ( D)

were u is any given function of x.


Proof: For any function V,
D(xV) = xDV + V
or D(xV) = xDV + (D)V
Again D2(ixV) = (DxV)
= D[xDV + V] by using (1)
= xD2V DV + DV
= xD2V + 2DV
xD2V + (D2) V
Similarly D3 (xV) = xD3V + (D3) V and so on
Thus for any positive integer n,
Dn(xV) = xDnV + (Dn)V
 If f(D) is a Differential operator
f(D) (xV) = x f(D)V + f(D)V
d
where f(D) = f ( D).
dD
Let V be the function such that
f(D)V = u
1
Then V= u
f (D)

substituting the values from (3) and (4) in (2), we get


 1  1
f ( D)  x u  = xu + f (D) u
 f ( D)  f ( D)

87
 1  f ( D)
xu = f ( D)  x u – u
 f ( D)  f ( D)

 f ( D)  1
= f ( D)  x –  u
 f ( D)  f ( D )

1  f ( D)  1
 xu =  x – u u
f ( D)  f ( D)  ( D)

Solved Example 8. Solve


d2y
+ 4 y = x sin x
dx 2
Solution: The auxiliary equation of the given differential equation is
m2 + 4 = 0
 m2 == –4
 m = + 2i
The C.F. of the given differential equation is
u = c1 cos 2x + c2 sin 2x.
The particular integral is given by:
1  2D  1
 Here f ( D) = D 3 + 4, f ( D)2 D 
V= ( x sin x ) =  x – 2 sin x
D +4
2
 D + 4  D2 4

 2D  1 1 2D 
= x –  sin x =  x sin x – 2 sin x 
 D + 4  –1 + 4
2
3 D +4 

1 2D 1 2 
=  x sin x – sin x   x sin x – cos x 
3 –1 + 4 3 3 

1 2 
 complete solution is y = u + v = c1 cos 2x c2 sin x +  x sin x – cos x  .
3 3 

Exercise
Solve the following equation:
d2y dy
1. – 4 + 4 y = sin 2 x.
dx 2 dx

d2y
2. + 4 y = cos3 x + sin 3 x.
dx 2

88
d 2 y dy
3. + – 2 y = x + sin x.
dx 2 dx

d2y
4. – 4 y = cos2 x
dx 2

d2y
5. + y = xe 2 x .
dx 2

d4y
6. – y = e x cos x .
dx 4
7. (D2 + 9) = ex – cos 2x.
8. (D2 – 2D + 5)y = e2x sin x.
9. (D2 – 3D + 2)y = ex sin 2x
10. (D2 – 7D – 6)y = e2x (1 + x)
Answers
1
1. y = (c1 + c2 x )e2 x + cos2 x.
8
1
2. y = c1 cos2 x + c2 sin 2 x – [cos3 x + sin 3 x ].
5
1 1
3. y = c1e x + c2 e –2 x – (cos x + +3sin x ) – ( x + 1).
10 4
1 1 
4. y = c1e–2 x + c2 e–2 x – 1 + cos2 x  .
8 2 
e2 x
5. yc1 cos x + c2 sin x + (5 x – 4).
25
1
6. y = c1e x + c2 e – x + c3 cos x + c4 sin x – e x cos x.
5
1 x 1
7. y = c1 cos3 x + c2 sin 3 x + e – cos2 x.
10 5
1 1
8. y = e x (c1 cos2 x + c2 sin 2 x) – e2 x cos x + eex sin x.
10 5
ex
9. y = c1e2 x + c2 e x + (2 cos2 x – 4sin 2 x ).
20
ex 1 
10. y = c1e– x + c2 e–2 x + c3e3 x +  6 – x
6  

89
Lesson 6
LINEAR DIFFERENTIAL EQUATIONS – DEFINITION,
FUNDAMENTAL THEOREM

Definition
Let us consider a linear differential equation of order n in the following
form:
dny d n –1 y d n –2 y
a0 ( x ) a1 ( x ) + a2 ( x ) + + a0 ( x ) y = f ( x ) ...(1)
dx n dx n –1 dx n –2

where a0(x), a1(x), ..., an(x) and f(x) are continuous on this interval. (a, b) [which may be
open, closed, semi-open, finite or infinite] and a0(x)  0 on this interval. Recall that a
solution y = y(x) over (a, b) will have the property that it together with y'(x), y''(x), ... y(n–1)
dny
(x) will be continuous on (a, b) and consequently from (1), y(n)(x) = will also be
dx n
continuous on this interval.
We now state the Fundamental Existence theorem for the linear differential
equation (1).
Statement: Let x = x0 be any point of the interval (a, b) and let c0, c1, c2, ..., cn–1 be
any arbitrary set of n real numbers; then there exists a unique solution, y(x), of (1) with
the property that
y(x0) = c0, y'(x0) = c1, y''(x0) = c2, ... y(n–1)(x) = cn–1
Further this solution is defined over the interval (a, b).
Corollary : A solution y (x) of (1) with the property that at a point x = x0 of (a, b)
y(x0) = 0, y'(x0) = 0, ..., y(n–1)(x0) = 0
is identically zero on (a, b).
We omit the proof of this theorem. Here we illustrate it by the following example.
d2y
Example 1. Let – 9 y = 18 be a given differential equation, where y (0) = 2, y' (0)
dx 2
= 6. Find its unique solution.
Here (D2 – 9)y = 18
Its auxiliary equation is
m2 – 9 = 0

90
 m = 3, – 3
 C.F. = c1e3x + c2e–3x
1
Also P.I. = 2
18
D –9
1
3= 2
e0 x = –2
D –9
 The complete solution of (i) is given by
y(x) = c1e3x + c2e–3x – 2
Now y(0) = 2  c1 + c2 – 2 = 2
and y'(0) = 6  3c1 – 3c2 = 6
Solving these, we get, c1 = 3, c2 = 1.
 The unique solution of (1) is given by
y(x) = 3e3x + e–3x – 2
Exercise I
Find the unique solution of the differential equation:
d2y
1. – 4 y = 12; given that y(0) = 4, y'(0) = 1
dx 2
d2y dy
2. 2
– 3 + 2 y = 2e3 x ; given that y(0) = 1, y'(0) = 4
dx dx

Answers
15 2 x 13 –2 x
1. y( x ) = e + e – 3.
4 4
2. y( x) = –ex + e2 x + e3 x .

Assumption. In the remainder of this section, we shall confine our statements of


theorems and their proofs to the important case n = 2 i.e., the second order differential
equation. The student, if he is interested will find no difficulty in extending the theorems
to the general case.
Linear Dependence and Linear Independence
Definition. Two solution y1(x) and y2(x) of a given differential equation
a0(x)y'' + a1(x)y' + a2(x)y = f(x)
are said to be linearly dependent if there exist constants, not both zero, such that
c1y1(x) + c2y2(x)  0 [x  (a, b)].

91
If y1(x) and y2(x) are not linearly dependent, they are said to be linearly independent i.e.,
if then y1(x) and y2(x) are linearly dependent.
Example 2. Show that cos x and sin x – cos x are linearly independent solutions of
y'' + y = 0.
Let y1(x) = cos x and y2(x) = sin x – cos x. It is easy to verify that y1(x) = cos x and y2(x)
are solutions of the given differential equation.
Suppose: c1y1(x) + c2y2(x)  0  x  IR

 c1 cos x + c2 (sin x – cos x)  0


 (c1 – c2) cos x + c2 sin x  0 ...(1)
In particular for x = 0 and x =  , we have from (1),
2

c1 – c2 = 0 and c2 = 0
 c1 = c2 = 0
 y1(x) and y2(x) are linearly independent solutions.
Exercises II
1. Show that sin 2x and cos 2x are linearly independent solutions of the differential
equation: y'' + 4y = 0.
2. Show that ex sin x and ex cos x are linearly independent solutions of the equation:
y" – 2y' + 2y = 0.
3. Show that y1(x) = sin x and y2(x) sin x – cos x are linearly independent solutions of
the equation: y" + y = 0. Determine the constants c1 and c2 satisfying
sin x + 3 cos x  c1y1(x) + c2y2(x).
(Ans. c1 = 4, c2 = –3).
Algebraic Properties of solutions
We now come to the algebraic properties of the solutions of differential equations.
Property I. If y = y1 is a solution of the linear differential equation
[a0(x) Dn + a1(x) Dn–1 + ... + an(x)] y = 0
 n

 or

 a ( x) D
r =0
r
n–r
y = 0  ; then y = cy1 is also a solution, where c is any arbitrary constant

d
and D  .
dx
Property II. If y = y1, y = y2, they y = y2, ... y = ym are any m (m < n) solutions of the
equation:

92
n
or  a ( x)D
r =0
r
n–r
y = 0, then y = c1y1 + c2y2 + ... cmym is also a solution where c1, c2, ..., cm are

any arbitrary constants.


The proofs of these properties are very straightforward and therefore, left to the reader.
Property III. There exist two linearly independent solutions y1(x) and y2(x) of the
differential eq.
a0(x) y'' + a1(x) y' + a2(x)y = 0, where a0(x)  0 ...(1)
with the property that every solution y (x) of (1) can be written in the form
y(x) = c1y1(x) + c2y2(x)
where c1 and c2 are suitably chosen constants.
Proof. Recall that we have assumed a0(x), a1(x) and a2(x) are continuous and a0(x)  0 on
the interval (a, b). Let x = x0 be any point of (a1, b) and y1(x), y2(x) be the solution
(guaranted by the Fundamental Existence theorem), with the properties:
y1 ( x0 ) = 1, y1 ( x0 ) = 0 
 ...(2)
y2 ( x0 ) = 0, y2 ( x0 ) = 1

Suppose there exists constants c1 and c2 such that


c1y1(x) + c2y2(x)  0 ...(3)
 c1y'1(x) + c2y'2(x)  0 on (a, b) ...(4)
 (3) and (4) hold for x = x0
 c1 y1 ( x0 ) + c2 y2 ( x 0 ) = 0 
 ...(5)
c1 y1 ( x0 ) + c2 y2 ( x 0 ) = 0 

By (2) and (5), we get c1 = 0 = c2


 y1(x) and y2(x) are linearly independent solutions of (1).
Let (x) be any arbitrary solutions of (1). We shall write it in the form
y(x) = c1y1(x) + c2y2(x)
by choosing suitable constants c1 and c2
Let Y(x) = y(x) – y(x0) y1(x) – y' (x0) y2(x) ...(6)
Note that Y(x) is a solution of (1) by Property II.
Now Y(x0) = y(x0) – y(x0) – y1(x0) – y'(x0) y2(x0) = 0, using (2)
 Y(x)  0, as x0 is any point of (a, b)
 (6)  y(x) = y(x0)y1(x) (x) + y' (x0)y2(x)

93
Here c1 = y(x) and c2 = y' (x0) and the proof of Property III is complete.
Definition: Wronskian of two solution
Let y1(x) and y2(x) be two solutions of the differential equation
a0(x)y'' + a1(x)y' + a2(x) y = 0 ...(1)
then
y1 ( x ) y2 ( x )
w(x) =
y1 ( x ) y2 ( x )

= y1(x)y'2(x) – y2(x)y'1(x)
is known as Wronskian of these solutions.
It may be noted that we can similarly define the Wronskian of any n solution of a given
linear differential equation of order n.
Example 3. We have seen in Example 2 that y1(x) = cos x and y2(x) = sin x – cos x are
solutions of the equation y'' + y = 0.
Now their Wronskian,
cos x sin x – cos x
w(x) =
– sin x cos x + sin x

= cos x (cos x + sin x) + sin x (sin x – cos x)


= cos2x + sin2x + 1  0
This illustrates the relation stated in the following theorem between the Wronskian of the
solutions and their linear independence.
Property 1. Two solutions y1(x) and y2(x) of the equation
a0(x)y'' + a1(x)y' + a2(x) y = 0 ...(1)
are linearly dependent if and only if their Wronskian is identically zero.
Proof. Sufficient part
Suppose w(x) = 0 and prove that the solution are linearly dependent, we have
y1 ( x ) y2 ( x )
0 ...(2)
y1 ( x ) y2 ( x )

It follows that if x = x0 is any point of (a, b),


y1 ( x0 ) y2 ( x0 )
0 ...(3)
y1 ( x0 ) y2 ( x0 )

There then exist constant c1 and c2, not both zero, such that

94
c1 y1 ( x0 ) + c2 y2 ( x0 )
=0 ...(4)
c1 y1 ( x0 ) + c2 y2 ( x0 )

Consider the solution c1y1(x) + c2y2(x). According to equation (4) this solution and its
derivative (c1y'1(x) + c2y'(x)) are zero when x = x0; hence
c1y1(x0) + c2y2(x0)  0 ...(5)
The proof of the sufficiency is complete.
Necessary Part
We assume that there exist constants c1 and c2, not both zero, such that
c1y1(x0) + c2y2(x0)  0 ...(6)
Now we have to prove that the wronskian w(x) = 0. Differentiate eqn. (9) w.r.t. ‘x’, we
have
c1y'1(x) + c2y'2(x)  0
Let x = –x0 be an orbitrary point of (a, b). The identities (6) and (7) hold, in particular,
when x = x0. There thus exist constants c1 and c2, not both zero, such that equation (4)
 y1 y2 
hold. It follows that the determinant of the coefficients of c1 and c2  = 0  must be
 y1 y2 
zero, that is, equation (3) must hold.
Since x = x0 was an arbitrary point of (a, b) and (2) must hold throughout (a, b).
Hence the proof.
Property 2. The Wronskian of two solutions of the equation
a0(x)y'' + a1(x)y' + a2(x)y = 0 ...(1)
is either identically zero or never zero on (a, b).
Proof. Let us assume that the function a0(x), a1(x) and a2(x) of equation (1) are
continuous in the interval (a, b) [may be open or closed] with a0(x)  0.
We note that
y1 ( x ) y2 ( x )
w(x) =
y1 ( x ) y2 ( x )

y1 ( x ) y2 ( x ) y1 ( x ) y2 ( x )
w'(x) = +
y1 ( x ) y2 ( x ) y1( x ) y2( x )

= y1(x)y"2(x) – y2(x)y"1(x)

95
 a1 a   a a 
= y1 ( x)  – y2 ( x) – 2 y2 ( x)  – y2 ( x)  – 1 y1( x) – 2 y1 ( x) 
 a0 a0   a0 a0 

 a1 a2 
 y( x) = y( x ) – y( x ) 
 a0 a0 
a1
= – [ y1 ( x) y2 ( x) – y( x) y2 ( x)]
a0

a1
w'(x) = – w( x )
a0

 Thus, w(x) is a solution of the first order differential equation


a1
w + w=0
a0

and according to the corollary of the fundamental existence theorem, w(x) = 0 if it is zero
at any point of (a, b).
Hence the proof.
It may be noted that in example 3, preceding this property, the two solutions are linearly
independent (see example 2) and consequently their Wronskian is non-zero.
Exercises III
1. Find the Wronskian of the two solutions of the equation. y" + m2y = 0, where m is
a positive constant.
2. Find the Wronskian of the solution of the equation. y"' – 6y" + 11y' – 6y = 0.
3. Prove that the Wronskian of the function em1x, em2x and em3x is equal to
(m1 – m2) (m2 – m3) (m3 – m1) e(m1 + m2 + m3)x
4. Prove that the Wronskian of the functions xa, xb and xc (x > 0) is equal to
(a – b) (b – c) (c – a) xa+b+c–3,
where a, b, c are three distinct real numbers. Hence show that the given functions are
linearly independent.
Answers
1. m
2. 2e6x

96
Linear Differential Equations with Variable Cofficients
Homogeneous, Euler’s Equations

A linear differential equation of the form


dny d n –1 y d n –2 y dy
xn + p1 x n –1 n –1 + p2 x n –2 n –2 + ... + pn –1 x pn y = Q ...(1)
dx dx dx dx
where p1, p2, p3 ... pn are constants and Q is either a constant or a function of x, is called a
homogeneous linear differential equation.
The equation (1) can be transformed into an equation with constant co-efficients by
changing the independent variable from x to z by the transformation
x = ez ...(2)
or z = log x ...(3)
If this change is made, we have the following relations
dy dy dz 1 dy  dz 1 
1. = =  =
dx dz dx x dz  dx x 

dy dy d
or x = = Dy where D =
dx dz dz

Differentiating again w.r.t. x, we have


dy d2y d 2 y dz 1 d 2 y
1. +x 2 = =
dx dx dz 2 dx x dz 2

dy 2
2 d y d2y
or x +x = 2
dx dx 2 dz

d2y d2 y dy
 x2 2
= 2 –x
dx dz dz

d 2 y dy
= 2–
dz dz

 d 
= ( D2 – D ) y  where D = dz 
 

= D (D – 1)y
Again differentiating w.r.t.x, we get
d2y 3
2 d y  d 3 y d 2 y  dz
2x + x =  3 – 2 
dx 2 dx 3  dz dz  dx

97
1  d3y d2y 
=  – 
x  dz 3 dz 2 

d2y 3
3 d y d3y d3y
or 2x2 + x = =
dx 2 dx 3 dz 3 dz 2

d3y d3y d2 y 2
2 d y
 3x3 = – – 2 x
dx 3 dz 3 dz 2 dx 2
= (D3 – D2)y – 2D (D – 1)y
= (D3 – 3D2 + 2D)y = D(D2 – 3D + 2)
= D(D – 1) (D – 2)y
In general, we get by induction
dny
xn = D(D – 1) (D – 2) ... (D – n+1)y
dx n
where n is a positive integer.
 The given differential equation reduces to
D(D – 1) (D – 2) ... (D – n + 1) y
+ p1D(D – 1) ... (D – n + 2)y
+ p2D(D – 1) ... (D – n + 3)y
..............................................
+ ......... + pn–1Dy + pny = Q'
where Q' is a function of z into which Q is transformed by the transformation z = log x
i.e., x = ex.
The above equation when simplified will reduce to a linear differential of the nth
order with constant co-efficients and therefore can be solved by the methods previously
discussed.
The transformed equation is of the form f(d) y = Q' where
f(D)  D (D – 1) ... (D – n + 1)
+ p1D (D – 1) ... (D – n + 2)
+ p2D (D – 1) ... (D – n + 3)
+ ..................... + pn–1D + pn
The auxiliary equation is given by f(m) = 0, by replacing D by M in the above equation. It
is clear that the auxiliary equation is of degree n, and so has n roots.

98
Case 1. Let m1, m2, m3 ... mn be the n distinct and real roots of the auxiliary equation f(m)
= 0.
Then the complementary function is given by
z
u = c1em1 + c2 em z + c3em z + 2 3
+ cn e mn z

 z = log x by (3)
log x
= c1em1 + c2 em2 log x + cn e mn log x
m1 m2 mn
= celog x + c2 elog x + cn elog x

= c1 x m + c2 x m +
1 2
+ cn x mn

Case 2. Let r roots of the auxiliary equation f(m) = 0 be the same i.e., m1 = m2 = m3 = mn
and let the remaining (n – r) roots viz., mr+1, mr+2 ..., mn be distinct, then the
complementary function in this case is given by
u = (c1 + c2z + c3z3 + ... + crzr–1) em1z
+[cr +1emr+1z + cr + 2 emr+2 z + + cn emn z ]

= c1 + c2 log x + c3 (log x )2 + cr (log x )r –1  e m1 log x

+[cr +1emr+1 log x + cr +2 emr+2 log x + + cn emn log x ]

= [c1 + c2 log x + c3 (log x )2 + cr (log x )r –1 ]x m1

+ [cr +1 x mr+1 + cr +2 x mr+2 + + cn x mn ]

Case 3. Let the auxiliary equation (12) have a pair of conjugate complex roots say
m1 = a + ib, m2 = a – ib and let the remaining roots m3, m4, ... mn be real and distinct then
the complementary function is given by
u = eaz [A cos bz + B sin bz]
+ [c3em3 z + c4 em4 z + + cn emn z ]

= ea log x [A cos b (log x) + B sin b (log x)]


+ [c3em3 log x + c4 em4 log x + + cn emn log x ]

= x a [ A cos(b log x) + B sin(b log x)]

+ [c3 x m3 + c4 x m4 + + cn x mn ]

The particular integral of the transformed differential equation is given by


1
v= Q'
f (D)

99
We shall illustrate the method by the following Examples.
Solved Example 1. Solve
d3y 2
2 d y dy
x3 3
+ 3 x 2
+x +y =0 ...(1)
dx dx dx
Solution: By the substitution x = ez i.e., z = log x, the differential equation (1) is
transformed into the following differential equation.
[D (D – 1) (D – 2) + 3D (D – 1) + D + 1] y = 0 ...(2)
d d d2
where D = x = , D(D – 1) = x 2 2
dx dz dx

d3
D (D – 1) (D – 2) = x 3
dx 3
or f(D) y = 0
where f(D) = D (D – 1) (D – 2) + 3D (D – 1) + D + 1
= D (D2 – 3D + 2) + 3D (D – 1) + D + 1
= D3 + 1
= (D + 1) (D2 – D + 1)
The auxiliary equation of the differential equation (2) is given by
f(m) = (m + 1) (m2 – m + 1) = 0

1 1– 4 1 i 3
 m = –1 and m = =
2 3
The general solution of the given d.e. (1), is given by
z
 3 3 
y = c1e– z + e 2  c2 cos z + c3 sin z
 2 2 

where z = log x
1
log x   3   3 
 y = c1e – kgx + e 2 c2 cos  log x  + c3 sin 
  2
log x  

  2    

c1   3   3 
= + x1/2 c2 cos  log x  + c3 sin  log x  
x  2   2 
     

Solved Example 2. Solve


d3y 2
3 d y dy
x3 3
+ 2 x 2
– x2 + xy = 1 ...(1)
dx dx dx

100
Solution: Dividing both sides by x, we get
d3y 2
2 d y dy 1
x3 3
+ 2 x 2
–x +y = ...(2)
dx dx dx x
On changing the independent variable by putting x = ez
1
so that z = log x and e–z = .
x
The equation (2) is reduced to
[D(D – 1) (D – 2) + 2D(D – 1) – D + 1] + y = e–z
or (D – 1) [D(D – 2) + 2D – 1] y = e–z
or (D – 1) (D2 – 1) y = e–x
or (D – 1)2 (D + 1) y = e–z
2
d d, 2 d
where D = = x D(D – 1) = x etc.
dz dx dx 2
The auxiliary equation of the equation (3) is
(m – 1)2 (m + 1) = 0
 m = –1 and m = 1 (repeated twice)
The auxiliary equation has a double root 1 and a single root –1.
 The complementary function is given by
u = (c1 + c2z) ez + c3e–z, where z = log x
 u = (c1 + c2 log x) elog x + c3e–log x
c3
= (c1 + c2 log x ) x +
x
The particular integral of the equation (3) is given by
1
v= e– z
( D – 1) ( D + 1)
2

1 1
=  2
 e– z
(D – 1) (D – 1)

1 1
=  e– z
D + 1 (–1 – 1) 2

 1 mx 1 mx 
 e = e ; if f (m)  0 
 f ( D) f ( m) 

101
1 1
=   [e – z ]
4 D +1
1 –z 1
= e 1
4 (D – 1 + 1)

 1 ax 1 
 e u = e ax u
 f ( D) f ( D + a) 

1 –z 1 1
 v= e  1 = e – z  dz
4 D 4
1 –z
= e z
4
By the substitution z = log x, we get
1 log x
v=
4 x
The complete solution of the equation (2) or (1) is, from (4) and (5), given by
c3 1 log x
y = u + v = (c1 + c2 log x ) x + + 
x 4 x
Solved Example 3. Solve
d2y dy
x2 2
+ x 4 y = x3
dx dx
Solution: Put x = ez or z = log x
d d
 x = =D
dx dz
d2
x2 = D(D – 1)
dx 2
d
where D  .
dx
The given equation transforms to
[D(D – 1) + D – 4] y = e3z
(D2 – 4) y = e3z
The auxiliary equation m2 – 4 = 0 gives
m =+2
Complementary function is given by
u = c1e2z + c2e–2z

102
c2
i.e. u = c1 x 2 +
x2
( ez = x )
1 1
Particular Integral is given by v=  e3 z = 2 e3 z
D2 – 4 3 –4
1 3z 1 3
= e = x
5 5
The complete solution is given by
c2 1 3
y = u + v = c1 x 2 + + x
x2 5
Solved Example 4. Solve
d2y dy
x2 2
–x + 2 y = x log x
dx dx
Solution: Put x = ez or z = log x
d d
x = =D
dx dz

d2
x2 = D(D – 1)
dx 2
The given equation transforms to
[D (D–1) – D + 2] y = zez
The auxiliary equation m2 – 2m + 2 = 0
m =1+i
Complementary function is
u = ez [A cos z + B sin z]
= x [A cos (log x) + B sin (log x)]
Particular Integral is given by
1 1
v= [ zez ] = ez  z
D – 2D + 2
2
(D + 1) – 2(D + 1) + 2
2

 1 ax 1 
 e u = e ax –  u
 f ( D) f ( D + a) 

1
= ez z
D + 2D + 1 – 2D – 2 + 2
2

103
1
= ez  z
D +1
2

= ex . (1 + D2)–1. z
= e z [1 – D 2 + D 4 ]  z
= ez . z
= x log x, ( z = log x )
The complete solution is,
y = u + v = x [A cos (log x) + B sin (log x)] + x log x
d2y dy
Solved Example 5. Solve ( x + 3 ) – 4 ( x + 3 ) + 6 y = x.
2

dx 2 dx
Solution: Let x + 3 = ez. So that z = log (x + 3)
The given equation now becomes
[D(D – 1) – 4D + 6] y = ez – 3 = (D2 – 5D + 6)y = ez – 3
The auxiliary equation m2 – 5m + 6 = 0 gives m = 2, 3
C.F. = C1e2x + C2e3z
= C1e2 log(x + 3) + C2e3 log(x + 3)
2 3
= C1elog( x +3) + C2 elog( x +3)

= C1(x + 3)2 + C2(x + 3)3


1
P.I. = (e z – 3)
D – 5D + 6
2

1 1
= ez – 3 2 ez
D – 5D + 6
2
D – 5D + 6
1 z 1 1
= e – 3  = (e z – 1)
2 6 2
1 log( x +3) 1 1
= e – 1 = ( x + 3 – 1) + ( x + 2)
2 2 2
Hence the complete solution is
1
y = C1(x + 3)2 + C2(x + 3)3 + ( x + 2)
2

104
Exercise
Solve the following differential equations:
d3y 2
2 d y dy
1. x 3 3
– 3 x 2
+ 7x – 8 y = 0.
dx dx dx

d2y dy
2. x 2 2
+ 4x + 2y = ex .
dx dx

d2y dy
3. x 2
+2 = 6 x.
dx dx

[Hints: On multiplying both sides by x, the given equation can be written as


d2y dy
x2 2
+ 2x = 6x2 ]
dx dx

d2y dy
4. x 2 2
+x – 4y = x2 .
dx dx

d3y d2y  1
5. x 3 3
+ 2 x 2 2 + 2 y = 10  x +  .
dx dx  x

6. ( x2 D2 + xD –1) y = x3 .

7. ( x 2 D2 – xD + 2) y = x log x.

d2y dy
8. x 2 + 2x = log x.
dx 2 dx

9. ( x2 D2 – xD – 3) y = x2 log x.

d2y dy 1
10. x 2 +x – 3y = x 2 + 2
dx 2 dx x

d2y dy
11. x 2
+x + y = log x.
dx dx

12. ( x 2 D2 – 5xD + 8) y = x log x.

 1
13. ( x 2 D2 – xD + 1) y = 5  x +  .
 x 

Answers
1. y = x 2 [c1 + c2 log x + c3 (log x)2 ].

1 1 ex .
2. y = c1  + c2  +
x x2 x2

105
c2
3. y = c1 + + x2 .
x
c2 1 2
4. y = c1 x 2 + + x log x.
x2 4
c1
5. y = + (c2 cos log x + c3 sin log x ) + 5 x + 2 x –1 log x.
x

1 x3
6. y = c1 x + c2 + .
x 8

7. y = c1 x cos(log x + c2 ) + x log x.

1 1
8. y = c1 + c2 + (log x )2 – log x.
x 2
c1 1 2
9. y = + c2 x 3 – x 2 log x – x 2 .
x 3 9
C1 1 1
10. y = + C2 x 3 – x 2 + 2 .
x 3 5x

11. y = C1 cos(log x ) + C2 sin(log x ) + log x .

x 4
12. y = C1 x 2 + C2 x 4 +  log x +  .
3 3
 
5x 5 1
13. y = x (C1 + C2 log x ) + (log x )2 +  .
2 4 x
Euler’s Equation – Linear Equations with Variable Coefficients
An equation of the form
dny n –1 d
n –1
n– x d
n –2
dy
a0 x n + a1 x + a2 x + + an –1 x + an y = 0, an  0 ...(1)
dx n dx n –1 dx n –2 dx

where a0, a1, ... an are constants, is an Euler-type equation.


The transformation x = et reduces equation (1) to a linear equation with constant
coefficients.
Illustrated Examples
d2y dy
1. x 2 2
–x + y = 2 log x
dx dx
Under the transformation x = et we have
dy dy dx 1 dy 1 dy ,
= / = t =
dx dt dt e dt x dt

106
d 2 y d  dy  d  1 dy  dt
=   =
dx 2
dx  dx  dt  x dt  dx

 1 dx dy 1 d 2 y  dx
= – 2
+ 2 
/
 x dt dt x dt  dt

d2y 1  d 2 y dy 
=  – 
dx 2 x 2  dt 2 dt 

Equation becomes
d 2 y dy dy
– – + y = 2t
dt 2 dt dt
d2y dy
– 2 + y = 2t
dt 2 dt
Complementary solution is
y = et (C1 + C2t)
The particular integral is
2t , d
= 2
where D =
(D – 1) dt
= 2[1 – 2D + D2]–1 t = 2 (1 + 2D – D2)t
= 2t + 4
Hence the complete solution is
y = C.F. + P.I.
= et (C1 + C2 t) + 2t + 4
in terms of x, is
y = x (C1 + C2 log x) + 2 log x + 4.
d2y dy
Example 2. x 2 2
– 2x + 2y = 0
dx dx
Substituting x = et we have
dy dy dx 1 dy
= / =
dx dt dt x dt
d 2 y d  dy  d  dy  dt 1  d 2 y dy 
=   =   =  – 
dx 2 dx  dx  dt  dx  dx x 2  dt 2 dt 

Equation becomes
d2y dy
2
– 3 + 2y = 0
dt dt

107
Its complementary function is
y = C1 et + C2 e2t.
in terms of x is
y = C1 x + C2 x2.
d2y
Example 3. x 2 + y = 3x 2 .
dx 2
Substituting x = et we have
d2y 1  d 2 y dy 
=  –  (as similar to the previous example)
dx 2 x 2  dt 2 dt 

d 2 y dy
– + y = 3e2t
dt 2 dt
which in linear differential equation with constant coefficient (non-
homogeneous).
The auxiliary equation

1 3
m2 – m + 1 = 0, root = i
2 2
Its complementary function is
t
 3 3 
y = e 2  C1 cos t + C2 sin t
 2 2 

The particular integral is


3e2 t , where, D = d
=
D – D +1
2
dt

3e2 t 3e2 t
= = = e2 t
22 – 2 + 1 3
The complete solution is
y = C.F. + P.I.
t
 3 3 
y = e 2  C1 cos t + C2 sin t  + e2t
 2 2 

in terms of x,
1   3   3 
y = x 2 C1 cos  log x  + C2 sin  log x   + x 2
  
  2   2  

108
Exercises
d2y dy
1. x 2 2
– 3x + 3y = 0
dx dx

d2y dy
2. x 2 2
–x + 2y = 0
dx dx

d2y dy
3. x 2 2
+ 4x + 2y = ex
dx dx

d2y dy
4. x 2 +x – y = xn
dx 2 dx

d2y dy
5. x 2 2
+x – y = x3
dx dx

d2y dy
6. x 2 2
–x + 2 y = x ln x
dx dx

d2y dy
7. x 2 2
– 2x – 20 y = ( x + 1)2 .
dx dx

Answers
1. y = c1x + c2 x3
2. y = x(c1 cos log |x| + c2 sin log |x|)
3. y = c1x–1 + c2x–2 + x–2ex
xn
4. y = c1 x + c2 x –1 +
n2 – 1

x3
5. y = c1 x + c 2 x –1 +
8
6. y = x (c1 cos ln x + c2 sin ln x) + x ln x
1 2 1 1
7. y = c1 x 4 + c2 x –5 – x – x– .
14 9 20

109
Lesson 7
LINEAR DIFFERENTIAL EQUATIONS OF THE SECOND
ORDER WITH VARIABLE COEFICIENTS

Method I of solving y'' + Py' + Qy = R, when an integral in its complimentary functions


[C.F.] is known. The general form of the linear differential equation of second order is
y'' + Py' + Qy = R, ...(1)
where P, Q, R are function of x alone.
If the coefficients P and Q are constants this equation can be solved by the methods
discussed earlier; otherwise no general method is available. However if an integral
included in the C.F. of such an equation is known, the complete solution can be expressed
in terms of the known integral in the following way.
Suppose y = z be an integral included in the C.F. of (1); then substituting
y = vz ; y' = vz' + v'z
and y'' = vz'' + 2v'z' + v''z
in (1), we get
(vz'' + 2v'z' + v'z) + P(vz' + v'z) + Qvz = R
 v(z'' + Pz' + Qz) + v' (2z' + Pz) + v'' z = R ...(2)
As y = z is an integral of the C.F. of (1), therefore
z" + Pz' + Qz = 0.
Thus (2) reduces to
v'(2z' + Pz) + v''z = R
Put v' = p in it to get
dp  2 z   R
+ + P = ...(3)
dx  z  z

This is a linear differential equation.


 2 z 
  z
+ P  dx

Here I.F. = e

=e 
2 log z + P d x

= z 2e
Pd x

110
 The solution of the equation (3) is

p.z 2 e  +   Rz e   dx + c1
Pd x Pdx

 

dv
= p = z –2 e     Rz e   dx + c1 z –2 e 
– P dx Pdx – Pdx

dx  

Integrating this, we get

v =   z –2 e     Rz e  dx  dx + c1   z –2 e   dx + c2 ,
– Pdx Pdx – Pdx

     

where c1 and c2 are arbitrary constants.


The general solution of equation (1) is given by y = vz, where v is as found above.
Note: Given below are some rules which will help the reader to find an integral in the
C.F. of a given differential equation (1) merely by inspection.
I. If 1 + P + Q = 0, take Z = ex
II. If 1 – P + Q = 0, take Z = e–x
P Q
III. If 1 + + = 0, take Z = eax
a a2
(Note : I and II are special cases of III when a = 1 and –1 respectively)
IV. If P + Qx = 0, take z = x
V. If 2 + 2 Px + Qx2 = 0, take z = x2.
Given below some illustrative examples to explain these rules.
Example 4. Solve
xy' – y = (x – 1) (y'' – x + 1)
x 1
 y – y + y =x–1 ...(1)
x –1 x –1
x 1
Here P= and Q =
x –1 x –1
Clearly P + Qx = 0
y  x is solution (integral) included in the C.F. of (1).
Put y = vx
 y' = v + v'x and y'' = 2v' + v'' x
 (1) becomes
x 1
2v + vx – (v + vx ) + vx = x – 1
x –1 x –1

111
2 x  x –1
 v +  –  v = ...(2)
 x x –1 x

 (2) reduces to
dp  2 x  x –1
+ –  p = ...(3)
dx  x x – 1  x

This being linear, its


2 x 
  x – x –1  dx
I.F. = e
x 2e– x
=
x –1
 The solution of (3) is
x 2e– x
 xe dx + c1
–x
p =
x –1

= c1 – xe– x – e– x

dv  x –1 x –1
 = p = c1  2  e x – ( x + 1)  2
dx  x  x

1 1   1 
= c1  – 2  e x –  1 – 2 
x x   x 

1 1  x  1 
 v = c2 + c1   – 2 
e dx –   1 – 2  dx
x x   x 

ex  1
= c2 + c1 – x + .
x  x

 y = vx = c2x + c1ex – (x2+1) is required solution of (1).


Example 5. Solve
xy'' – y' + (1 – x)y = x2e–x.
1 1 
 y – y +  – 1  y = xe–x
x x 
1 1
Here P= – and Q = – 1
x x
As 1 + P + Q = 0, y = ex is an integral included in the C.F. of (1).
Put y = vex
 y' = vex + v'ex

112
and y'' = vex + 2v' ex + v'' ex
 (1) becomes
1 1 
(v + 2v + v)e x – (v + v)e x +  – 1  ve x = xe – x
x x 

 1
 v +  2 –  v = xe –2 x
 x

Take v' = p in it to get


dp  1
+  2 –  p = xe–2x ...(2)
dx  x

which being linear, its


 1
  2– x  dx
I.F. = e
= e2 x / x .
 The solution of (2) is
e2 x
p = x + c1
x
dv
 = p = x2e–2x + c1xe–2x
dx

 v = c2 + c1  xe –2 x dx +  x 2 e –2 x dx

 1 1   1 1 1 
= c2 + c1  – xe–2 x – e–2 x  +  – x 2 e–2 x – xe–2 x – xe–2 x 
 2 4   2 2 4 
 y = vx is the solution of (1).

Exercise I
Solve the following differential equations.
1. (1 – x2) y'' + xy' – y = x (1 – x2)3/2
2. y'' – x2y' + xy = x
3. xy'' + (1 – x) y' – y = ex
4. y'' – y' cot x – (1 – cot x) = ex sin x
5. (3 – x) y'' + (4x – 9) y' + (6 – 3x) y = 0

113
Answers

1. y = c2 x – c1 ( ) 1
1 – x 2 + sin –1 x – (1 – x 2 )3/2
9
2. y = c1 x – c2 x  x –2e x dx + 1.
3/3

3. y + e x log x + c1e x  x –1e – x dx + c2 e x .


1
4. y = c1e x + c2 e – x (cos x + 2sin x ) – e x cos x.
2
5. y = c1ex + c2e3 x (4x3 – 42x 2 + 150 x –183).

Method II. Changing the dependent Variable.


In the second order differential equation.
y'' + Py' + Q = R ...(1)
putting y = vz and proceeding as in Method I, we get
 z  1 R
v +  2 + P  v + ( z + Pz + Qz ) v =
 z  z z

 v + P1 v + Q1v = R1 ...(2)

z 1
where P1 = 2 + P, Q1 = ( z  + Pz  + Qz ) and R1 = R/z.
z z

Any desired value can be assigned to P1 or Q1 by the proper choice of z. The choice of z
so that Q1 = 0 was made in Method I. The from above
2z 
= P1 – P.
z

On integrating it, we get


1
2
log z = ( P1 – P)dx

1
z = e2 
( P1 – P ) dx
 ...(3)
Thus by assigning any arbitrary value to P1; the corresponding value of z can be obtained
from (3).
In particular, if P1 = 0, then z is given by
1
z = e 2
– Pdx

1
 1  – Pdx
 z" =  – P  e 2 
 2 

114
1
1 1  – Pdx
and z" =  P 2 – P  e 2 
4 2 

1 1
 Q1 = Q – P  – P 2
2 4
Thus with the value of z given by (3), equation (2) becomes
v" + Q1v = R1 ...(4)
Equation (4) is known as the normal form of the equation (1) and in quite a few cases, it
happens to be integrable. The transformation of equation (1) to the form (4) is known as
the Removal of the first derivative.
Example 6. Solve
cos2xy" – 2 sin x cos x . y" + cos2 x . y = 0 ...(1)
 y" – 2tan x . y' + y = 0
Here P = –2 tan x, Q = 1, R = 0.
1 1
 Q1 = Q – P – P2
2 4
= 1 + sec2x – tan2x = 2.
1
z = e 2  = e
– Pdx tan dx
and = sec x
 Normal form of the equation (1) is
v" + Q1v = R1
 v" + 2v = 0
 (D2 + 2) v = 0

 v = c1 cos 2 x + c2 sin 2 x

 ( )
y = vz = c1 cos 2 x + c2 sin 2 x sec x.

Exercises I
Solve the following equations
1. y" – 2 tan x . y' + 5y = 0.
2. y" – 2y' tan x + 5y = ex sec x.
2
3. y – 4 xy + (4 x 2 – 1) y = –3e x sin x.
4. y" + 2xy' cot nx + (m2 – n2)y = 0
5. x2y" – 2x (x + 1)y' + (x2 + 2x + 2)y = 0

115
Answers
1. y = ( c1e2 x + c2 e –2 x ) sec x.

( )
2. y = c1 cos 6 x + c2 sin 6 x sec x + e x sec x.
1
7
2  3 
3. y = e x  a cos x + b sin x + x cos x  .
 2 
4. y sin nx = a cos mx + b sin mx.
5. y = ex (ax2 + bx)
Method III. Changing the independent variable.
An equation can sometimes be transformed into an integrable form by changing the
independent variable.
Suppose that y" + Py' + Qy = R ...(1)
be any linear equation of second order and that the independent variable is to be changed
from x to z, by some given relation, z = f(x) connecting x and z; then
dy dy dz
y' = =  = y1  z 
dx dz dx

d2y d
and y" = = ( y1  z ) = y2  ( z ) 2 + y1 z 
dx 2 dx

dy d2y
where y1 = ; y2 = 2
dz dz
 Equation (1) becomes
y2(z')2 + y1z" + Py1 . z' + Qy = R
 z  + Pz   Q R
 y2 +   y1 + y =
 (z )  2

(z ) 2
( z)2

 y2 + P2y1 + Q2 = R2 ...(2)
z + Pz , Q R
where P2 = Q2 = and R2 =
(z) 2
(z) 2
(z)2

Here P2, Q2, R2 are functions of x but can be immediately expressed as functions of z by
means of z = f(x): Since z is arbitrary, it may be chosen to satisfy any assignable
condition.
(i) If by equating Q2 to a constant. P2 also becomes constant : Equation (2) can be
integrated i.e., when Q2 = a2,
 Q = a2 (z')2

116
 a dz = Q dx

 az =  Q dx

(ii) If z is so chosen that P2 = 0


then z" + Pz' = 0

z' = e 
– Pdx

z =   e   dx.
– Pdx

 

It with this value of z, Q2 becomes constant or constant divided by z2, equation (2) can
be integrated.
Example 7. Solve
y" + cot x . y' + 4y cosec2x = 0
Here P = cot x, Q = 4 cosec2x; R = 0.
Choose z such that P2 = 0; then z is given by

z =   e   dx =  e – logsin x dx
– Pdx

 

x
=  cosec x dx = log tan .
2

 z' = cosec x
 Q2 = Q/(z')2 = 4 and R2 = 0
 The given equation reduces to
y2 + 4y = 0
d
 (D2 + 4)y = 0 where D 
dz

 y = a cos 2z + b sin 2z
 x  x
= a cos  2 log tan  + b sin  2 log tan 
 2   2

Example 8. Solve
x6y" + 3x5y' + a2y = x–2
 y" + 3x–1y' + a2x–6y = x–8
Here P = 3x–1, Q = a2x–6 and R = x8

117
Choose z, so that P2 = 0; then z is given by
1
z =   e   dx =  e –3log x dx =  x –3 dx = – 2 .
– Pdx

  2x

 Q2 = Q/(z')2 = a2
1
and R2 = R / ( z )2 = = –2 z
x2
 The given equation reduces to
y2 + a2y = –2z
d
 (D2 + a2)y = –2z where D 
dx
 C.F. = c1 cos az + c2 sin az
1
and P.I. = (–2 z )
D + a2
2

–1
2 D2 
= – 2  1 + 2  z = –2 z /a2 .
a  a 

 y = C.F. + P.I

a   ax 2   1 
= c1 cos  x 2  – c2 sin  + 2 2 
2   2  a x 

Exercises II
Solve the following differential equations:
1. y" + 2x–1y' + a2x–4y = 0
2. xy" – y' – 4x3y = x5
3. y" + y' tan x + y cos2x = 0
4. y" cosx + y' sin x – 2y cos3x = 2cos5x.
5. xy" – y' + 4x3y = 8x3 sinx2
Answers
1. y = c1 cos (a/x) + c2 sin (a/x).
2 2 1
2. y = c1e x + c2 e – x – x 2 .
4
3. y = c1cos (sin x) + c2 sin (sin x).
4. y = c1e 2 sin x
+ c2 e – 2 sin x
+ sin 2x .
5. y = c1 cosx2 + c2 sinx2 + x2 cosx2.

118
Method of Variation of Parameters, Method of Undetermined
Coefficients, Euler’s Equations

Linear Differentials Equations of Second Order


Method of Variation of Parameters
In this lesson we will be concerned with the method of variation of parameters for solving
linear differential equations whose complementary function is known. Infact this method
is applicable to solve linear differential equations of any order.
Let us consider the differential equation
a0(x)y" + a1(x)y' + a2(x)y = f(x), ...(1)
where, a0(x), a1(x), a2(x) and f(x) are continuous with a0(x)  0 on an interval (a, b).
Let y1(x) and y2(x) be linearly indepedent solutions of the corresponding
homogeneous equation
a0(x)y" + a1(x)y' + a2(x)y = 0, ...(2)
then
y = C1y1(x) + C2y2(x), ...(3)
where C1 and C2 arbitrary constants will be the complementary functions (C.F.) of eqn.
(1). The method of variation of parameters always produces a particular solution of eqn.
(1). We first try to determine functions u1(x) and u2(x) such that
u1(x)y1(x) + u2(x)y2(x) ...(4)
satisfy eqn. (1).
let y = u1y1 + u2y2 ...(5)
then y' = u1 y1 + u1 y1 + u2 y2 + u2 y2

y' = (u1 y1 + u2 y2 ) + (u1 y1 + u2 y1 ) ...(6)

Since there are two functions u1 and u2 but only one condition [i.e. eqn. (6) satisfies
the eqn. (2)] on then, to get the functions u1 and u2, these functions u1 and u2 hence to
satisfy another condition, i.e. it does not violate the first one, i.e. let the second condition
be
u1 y1 + u2 y2 = 0 ...(7)

which leads to name variation of Parameters.


Then, 3y' = u1 y1 + u2 y2 [using (7) in eqn. (6)]

119
after differentiation
y" = u1 y1 + u1 y1 + u2 y2 + u2 y2 ...(8)

Substituting the values of y, y' and y" in eqn (1) from (5-8) we set,
a0 ( x) [u1y1 + u1 y1 + u2 y2 + u2 y2] + a1 ( x) [u1 y2] + a1 ( x) [u1 y1 + u2 y2 ]

+ a2 ( x ) [u1 y1 + u2 y2 ] = f ( x )

u1[a0 ( x) y1 + a1 ( x) y1 + a2 ( x) y1 ] + u2 [a0 y2 + y2 + a2 y2 ]

+a0 ( x)[u1y1 + u2 y2 ] = f ( x) ...(9)

Since y1(x) and y2(x) satisfy eqn. (2), hence


a0 ( x) y1 + a1 ( x) y1 + a2 ( x) y1 + 0 ...(10)

a0 ( x) y2 + a1 ( x) y2 + a2 ( x) y2 + 0 ...(11)

From (9), we have


f ( x)
u1 y1 + u2 y2 = = R (Since a0(x)  0) ...(12)
a0 ( x)

From eqns. (7) & (12), we have


u1 u 1
= 2 =
– Ry2 Ry1 y1 y2 – y1 y2
y2 ( x) f ( x)
or u1 ( x ) = – ...(13)
a0 ( x)w( x)
y1 ( x ) f ( x ) ,
u2 ( x ) = –
a0 ( x )w( x )
where w(x) = y1y'2 – y'1 y2 is the wronskian of y1 and y2
w(x)  0 since y1 and y2 are linearly indepedent. Integrating eqns. (13), we have
y2 ( x) f ( x)
u1(x) = –  dx + A ...(14)
a0 ( x )w( x)
y1 ( x) f ( x )
and u2(x) =  a ( x)w( x) dx + B
0

where A and B are aribtrary constants. The solution of eqn. (1) is obtained by substituting
(14) in (5).
Illustrated Examples
Example 1. Using variation of parameters, find the general solution of
y" + y = x ...(1)

120
Solution: Here the homogeneous eqn. of (1) is
y" + y = 0 ...(2)
The complementary functions,
y = C1 cos x + C2 sin x ...(3)
[Since (D2 + 1)y = 0  Auxiliary eqns. m2 + 1 = 0  m = i, –i]
and f(x) = x, a0(x) = 1, w(x) = y1y'2 – y2y'1
i.e. y1 = cos x, y2 = sin x
 w(x) = cos2x – sin x (– sin x) = 1
Equns. u1 y1 + u2 y2 = 0

and u1 y1 + u2 y2 = f ( x )

become, u1 ( x ) cos x + u2 ( x )sin x = 0 ...(4)

u1 ( x )(– sin x ) + u2 ( x ) cos x = 0 ...(5)

Solving u1 ( x ) = – x sinx

u2 ( x ) = x cosx

Integrating u1(x) = x cos x — sin x ...(6)


u2(x) = x sin x + cos x ...(7)
A particular solution y0 = u1y1 + u2y2
= (x cos x – sin x) cos x + (x sin x + cos x) sin x
y0 = x
Hence the general solution = C.F. + P.I.
y = C1 cos x + C2 sin x + x
Example 2. Using variation of parameters, find the general solution of
y" – 3y' + 2y = 2
Solution: The homogeneous equation of this equation is
y" – 3y' + 2y = 0
The complimentary function is
y = C1e2x + C2ex
Comparing with the standard equation
a0(x)y" + a1(x)y' + a2(x)y = f(x)

121
we have f(x) = 2, a0 = 1, a1 = –3, a2 = 2, y1 = e2x, y2 = e2,
w(x) = y1y'2 – y'1y2 = (1 – 2x) e3x = – e3x  0
We try to determine functions u1(x) and u2(x) such that
u1 y1 + u2 y2 = 0

and u1 y1 + u2 y2 = f(x)

i.e. u1 e2 x + u2 ex = 0

u1 e2 x 2 + u2 ex = 2

Solving u1 = 2e–2x, u 2 = –2e–x

Integrating u1 = –2e–2x, u2 = –2e–x


A particular solution
y0 = u1y1 + u2y2 + –e–2x e2x + 2e–xex = 1
Hence the general solution
y = Complementary function + Particular Integral
y – C1e2x + C2ex + 1
Example 3. y" + 4y = sin x
The homogeneous equation of this equation is
y" + 4y = 0
The complementary function of this differential equation is
y = C1 cos 2x + C2 sin 2x
After comparing with standard equation i.e.
a0y" + a1y' + a2y = f(x),
We get a0 = 1, a1 = 0, a2 = 4, f(x) = sin x
and y1 = cos 2x, y2 = sin 2x
 w(x) = y1 y2 – y1 y2 = cos 2x cos 2x2 – sin 2x sin 2x(–2)

= 2 (cos2 2x + sin2 s x) = 2  0
Now we determine two functions u1 and u2 such that
u1 y1 + u2 y2 = 0

u1 y1 + u2 y2 = f(x)

122
i.e. u1 cos 2x + u 2 sin 2x = 0

u1 sin 2x(–2) + u 2 ex = 2 sin x

1 1
Solving u1 = – sin x sin 2 x, u2 = sin x cos2 x
2 2
1 1
u1 = [cos3 x – cos x ], u2 = [sin 3 x – sin x ]
4 2

1  sin 3 x sin x 
Integrating u1 = –
4  3 1 

1  cos3 x 
u2 =  – + cos x 
4 3 
A particular solution is

y0 = u1y1 + u2y2 =  sin 3x –   


1 sin x cos3 x cos x
12 4  cos2 x +  – 12 + 4  sin 2 x
   
1 1
= (sin 5 x + sin x) – (sin 3 x – sin x)
24 8
1 1
– (sin 5 x – sin x) + (sin 3 x + sin x)
120 8
 2 1 1 8 1
= sin x  + +  = sin x = sin x
 24 8 8  24 3
1
P.I. = sin x
3
Hence the general solution is
y = C.F. + P.I.
1
= sin x + C1 cos2 x + C2 sin 2 x
3
Example 4. y" + y = cosec x.
The homogeneous equation of this equation is
y" + y = 0
The complementary function of this differential equation is
y = C1 cos x + C2 sin x
After comparing with the standard equations
we get, a0 = 1, a1 = 0, a2 = 1, f(x) = cosec x
and y1 = cos x, y2 = sin x

123
w(x) = y1 y2 – y1 y2 = cos2x + sin2x = 1  0
Now we determine two functions u1(x) and u2(x) such that
u1 y1 + u2 y2 = 0
and u1 y1 + u2 y2 = f(x)
i.e. u1 cos x + u 2 sin x = 0
– u1 sin x + u 2 cos x = cosec x
Solving u1 = –1, u 2 = cot x
u1 = –x, u2 = log sin x
A particular solution is
y0 = u1y1 + u2y2
y1 = –x cos x + (log sin x) sin x
Hence the general solution is
y = C.F. + P.I.
y = c1 cos x + c2 sin x + sin x log sin x – cos x
Exercises
Using the method of variation of parameters, find the general solution of the following
equation.
1. y" + 4y = ex
2. y" + y = tan x
3. y" + y = 2 sin x
4. (x + 2)y" – (2x + 5)y' + 2y = (x + 1)ex
5. x2y" – 2x(1 + x)y' + 2(1 + x)y = x3
6. x2y" + xy' – y = x2ex
7. xy' – y = (x – 1)y" + (x – 1)2
Answers
1 x
1. e + C1 sin 2 x + C2 cos2 x
5
2. C1 sin x + C2 cos x – cos x ln (sec x + tan x).
3. C1 sin x + C2 cos x – x cos x
4. C1(2x + 5) + C2e2x – e2
x2
5. C1 x + C2 x e2 x –
2
C 1
6. C1 x + 2 + e x – e x
x x
7. C1ex + C2x + x2 + x + 1

124
Lesson 8
SIMULTANEOUS DIFFERENTIAL EQUATIONS

In this lesson we will consider the case in which there is a set of relations consisting of as
many simultaneous linear differential equations as there are dependent variables. Here we
will consider a system of differential equations with two or more dependent variables and
one independent variable can be solved by eliminating all but one of the dependent
variables, thus obtaining a single ordinary differential equation. Solution obtained should
be checked by substitution into the original differential equations to insure that the proper
number of arbitrary constants are present.
Illustrated Examples
dx
1. – 7x + y = 0 ...(1)
dt
dy
– 2 x – 5y = 0 ...(2)
dt
we can write
d 
 dt – 7  x + y = 0
 
d 
–2 x +  – 5  y = 0
 dt 
or (D – 7)x + y = 0 ...(3)
–2x + (D – 5)y = 0 ...(4)
Eliminating y, i.e. operate on equation (3) by (D – 5) and subtracting the result from (4)
we get
[–2 – (D – 7) (D – 5)]x = 0
(D2 – 12D + 37)x = 0
It is homogeneous linear differential equation with constant coefficients the
complimentary function can be found very easily, i.e. the auxillary equation
m2 – 12m + 37 = 0,
12  144 – 4  37 12  –4
roots = = =6i
2 2
Hence the complimentary function
x = e6t(c1 sin t + c2 cos t)

125
the easiest way to get y in this example into use eqn. (1). Substituting for x in (1), we get
36 e6t (c1 sin t + c2 cos t) + e6t (c1 cos t – c2 sin t) – 7e6t (c1 sin t + c2 cos t) + y = 0
y = – e6t [(6c1 – c2 – 7c1) sin t + (6c2 + c1 – 7c2) cos t]
y = e6t [(c1 + c2) sin t + (c2 – c1) cos t]
Ans. x = e6t (c1 sin t + c2 cos t)
y = e6t [c1 + c2) sin t + (c2 – c1)cos t]
d2x
Example 2. – 3x – 4 y = 0 ...(1)
dt 2

d2y
+x+y =0 ...(2)
dt 2
we can write
(D2 – 3)x – 4y = 0 ...(3)
x + (D2 + 1)y = 0 ...(4)
Eliminating x, as in simultaneous linear equations of elementary algebra. To do this
we operate on equations (4) by (D2 – 3) and subtracting from (3), we get
[–4 – (D2 + 1) (D2 – 3)]y = 0
Solving this in the usual way, we get
[put D2 = m
(m + 1) (m – 3) + 4  m2 – 2m + 1 = 0  (m – 1)2 = 0
i.e., m = 1, 1
D2 – 1 = 0  D = + 1]
y = et (c1 + c2 t) + e–t(c2 + c4 t)
Now we will find x from eqn. (4), substituting for y in eqn. (4), we get
dy
= et c2 + e–t c4 + et (c1 + c2 t) –e–t(c2 + c4 t)
dt
d2y
2
= et c2 – c4 e–t + et c2 – c4 e–t + et (c1 + c2 t) + e–t (c1 + c4 t)
dt
d2y = 2c2 et – 2c4 e–t + et (c1 + c2t) + e–t (c1 + c2t)
(D2 + 1)y = 2[c2 et – c4 e–t + (c1 + c2t) e–t]
Hence
x = –2[((c1 + c2) + c2t) et – c4et]
y = et (c1 + c2t) + (c2 + c4t)e–t

126
dx dy
Example 3. 3 + 9 + 44 x + 49 y = t
dt dt
dx dy
3 + 7 + 34 x + 38 y = et
dt dt
we can write
(4D + 44)x + (9D + 49)y = t
d.
(3D + 34)x + (7D + 38)y = et, D =
dt
We can easily eliminate x and y using Cramers Rule
x y l
= =
t 9D + 49 4 D + 44 t 4 D + 44 9D + 49
et 7D + 38 3D + 34 et 3D + 34 7D + 38

4 D + 44 9 D + 49 t 9D + 49
x =
3D + 34 7 D + 38 et 7D + 39

4 D + 44 9D + 49 4 D + 44 t
and y =
3D + 34 7D + 38 3D + 34 et

solving
[(4D + 44) (7D + 38) – (9D + 49) (3D + 34)x = (7D + 38)t – (9D + 49)et
and
[(4D + 44) (7D + 38) – (9D + 49) (3D +34)]y = (4D + 44)et – (3D + 34)t
(D2 + 7D + 6)x = 7 + 38t – 58et
and (D2 + 7D + 6)y = 48et – 34t – 3
These are linear differential equation with constant coefficients, we will solve as to
the previous lesson
d.
(D2 + 7D + 6)x = 7 + 38t – 58 et, D =
dt
The complementary function
x = c1e–t + c2e–6t
1
P.I. = (7 + 38t – 58et )
D + 7D + 6
2

1  D2 + 7D  58
= 1 –  (7 + 38t ) – 2 et
6 6  D + 7 D + 6

127
1 7  58
=  7 + 38t – 38  – et
6 6  14

19t 56 29 t
= – – e
3 9 7
The solution is given by
x = C.F. + P.I.
19 29 t 56
x = c1e – t + c2 e –6t + t– e –
3 7 9
Solving for y
C.F. = y = c3 e–t + c4 e–6t
1
P.I. = (48 et – 38t – 3)
(D + 7D + 6)
2

24 t 17 55
= e – t+
7 3 9
The general solution is
y = C.F. + P.I.
24 t 17 55
= c3e – t + c4 e –6 t + e – t+
7 3 9
For the proper member of arbitrary constant, we will find the relation between c1, c2
and c3, c4 by substituting these solutions into the original set of differential equations.
i.e. c2 = –c1
c4 = 4c2
The required solution
19 29 t 56
x = c1e – t + c2 e –6t + t– e –
3 7 9
17 24 t 56
x = c1e – t + 4c2 e –6t – t+ e –
3 7 9
Example 4. (D – 1)x + Dy = 2t + 1 ...(1)
d
D=
dt
(2D + 1)x + 2Dy = t ...(2)
Multiplying eqn. (1) by 2 and subtracting from (2), we get,
2
3x = –3t – 2, x = –t –
3

128
Substituting the value of x in eqn. (1), we obtain
2 4
Dy = 2t + 1 – ( D – 1) x = 2t + 1 + 1 – t – = t +
3 3

t2 4
y= + t + c1
2 3
The complete solution is
2 1 4
x = –t – , y = t 2 – t + c1
3 2 3

D –1 D
Infact, here is of order 1 in D.
2D + 1 2D

( 2D2 – 2D – 2D2 – D = –3D)


Therefore there should be only one arbitrary constant in the final solution.
Exercises
1. (D + 2)x + (D + 1)y = 0
5x + (D + 3)y = 0
2. (D – 2)x + (2D + 2)y = 3et
(3D + 2)x + (D + 1)y = 4e2t
dy 2 dz
3. 2 – – 4y = 2x
dx 2 dx
dy dz
2 + 4 – 3z = 0
dx dx
dx
4. + 4 x + 3y = t
dt
dy
+ 2 x + 5 y = et
dx
5. Dx – (D + 1)y = –et
x + (D – 1)y = e2t
6. (D – 1)x + (D + 3)y = e–t – 1
(D – 2)x + (D + 1)y = e2t + t

129
Answers
1
1. x = [( A + 3)et + (3 – B)e – t ]
5
y = Aet + Be– t
6
– t 1 3
2. x = c1e + e 2 t – et
5
2 11
y 1 3
x + = c2 e – t + e2t – et
8 2 16
3
– x 1
3. y = (c1 + c2 x)e x + 3c3e 2
– x
2
3
– x 1
z = 2(3c2 – c1 – c2 x)e – c3e
x 2

3
31 5 1
4. x = c1e –2t + c2 e –7t – + t – et
196 14 8
2 9 1 5
y = – , c1e –2t + c2 e –7t + – t + et
3 93 5 24
3
5. x = (c1 – c2 )cos t + (c1 + c2 )sin t + e2 t
5
2 1
y = c1 cos t + c2 sin t + e2 t + et
5 2
–7t
5 2t 3 1
6. x = 2c1e 5 + e + t–
17 7 49
–7t
1 2t 1 26
y = 3c1e 5
– e + t–
17 7 49
Simultaneous Equations of the first order
The general type of a set of simultaneous equations of the first order form between
three variables is
P1dx + R1dy + R dz = 0
P2dx + Q2dy + R2dx = 0
where the coefficients are functions of x, y, z.
These equations can be expressed in the form
dx dy dz
= = ...(1)
P Q R

where P, Q and R are functions of x, y, z.


The solution of differential equation (1) consists of a pair of relations
f1(x, y, z) = c1 and f2(x, y, z) = c2

130
Method-I. By equating two out of three members in eqn. (1) if one of the variables be
absent from two members of eqn. (2) the method of procedure is obvious. For example
suppose that z is absent from P and Q, then the solution of
dx dy
=
P Q

gives relation between x and y say f1(x, y) = c1, which is one equation of the complete
solution. This equation may enable us to eliminate x or y from one of the other equations
in (1), and then another integral relation (solution) may be found, this will be the second
equation of the solution.
Method-II. (Multiplier Method)
Sometimes it is possible to find multipliers l, m, n and L, M, N such that at least one
of the equations
dx dy dz ldx + mdy + ndz Ldx + Mdy + Ndz
= = = =
P Q R lP + mQ + nR LP + MQ + NR

can be integrated easily.


Illustrated Examples
xdx dy dz
Example 1. = =
y2 z dz y2

The equation formed by first two fraction reduces to


x2dx = y2dy
integrating, x3 = y3 + 3c1 ...(1)
Similarly from first and last fractions give
x dx = 2 dz
 x2 = 32 + 2c2 ...(2)
The complete solution consists of these two solutions,
Ans. x2 = z2 + 2c2
x3 = y3 + 3c1
dx dy dz
Example 2. = =
x 2 y2 nxy

consider first two fractions, we have


dx dy
2
= 2
x y

131
x –1 y –1
Integrating = + c1
–1 –1
1 1
= – c1
x y

From last two fraction.


dy dz
2
=
y nxy

ndy dz
Then =
y x

1 
=  – c1  dz
y 

n
dz = dy
1 – c1 y

n
z= – ln(1 – c1 y) + c2
c1

nxy
z= ln(1 – c1 y) + c2 ...(2)
y–x

Hence equation (1) and (2) constitute the required solution.


dx dy dz
Example 3. = =
y2 x 2 x 2 y2 z 2

Consider first two fractions, we have


x2 dx = y2 dy
Integrating, x3 = y3 + 3c1, is its solution ...(1)
Consider last two fraction, we have
dz
y2dy =
z2

y3 1
Integrating, = – + c2
3 2
3
y3 + = 3c2 ...(2)
2
Hence eqn. (1) and (2) constitute the required solution.

132
a dx b dy c dz
Example 4. 2
= =
(b – c) y (c – a)zx (a – b) xy

we will solve by using multipliers method.


here P = (b – c)yz, Q = (c – a)zx, R = (a – b)xy
xP + yQ + zdz = xy z(b – c + c – a + a – b) = 0
a dx bdy axdx + bydy + czdz
Now, = = c dzR =
P Q xP + yQ + zdR = 0

 axdx + bydy + czdz = 0


Integrating, ax2 + by2 + cz2 + c1 ...(1)
Again, axP + byQ + czR = xyz (ab – ca + cb – ab + ac – bc) = 0
a dx b dy a2 xdx + b2 ydy + c 2 zdz
Now, = = c dzR =
P Q xP + byQ + czR = 0

 a2xdx + b2ydy + c2zdz = 0


On integrating, a2 x 2 + b2 y2 + c2 z 2 + c22

Hence, eqn. (1) and (2) constitute the required solution.


Exercises
dx dy dz
1. = =
y+z z+x x+y
dx dy dz
2. = =
mz – ny nx – lz ly – mx
dx dy dz
3. = =
yz zx xy

Answers
1. (x – y) = (y – z)c1
(x + y + z) (x – y)2 = c2
2. x2 + y2 + z2 = c1
ln + my + nz = c2
3. x2 – y2 = c1
y2 – z2 = c2

133
Lesson 9
TOTAL DIFFERENTIAL EQUATIONS

In this lesson we will consider single differential equations having one independent
variable and more than one dependent variables. In such equations the differential
coefficients of these dependent variables with respect to the single indepedent variable
appear. These equations are known as Total Differential Equations.
Consider the relation
(x, y, z) = c
where x, y and z are variables and c is an arbitrary constant.
Then d = 0
  
or dx + dy + dz = 0
x y z

or P dx + Q dy + R dz = 0
where P, Q and R are given function of x, y, z. This equation is called a total differential
equation.
Condition of Integrability
The condition of integrability of the total differential equation
P dx + Q dy + R dz = 0 ...(1)
 Q R   R P   P Q 
is P –  + Q –  + R –  =0
 z y   x z   y x 

Consider the equation


P dx + Q dy + R dz = 0
has an integral (x, y, z) = c ...(2)
so that the expression
  
d = dx + dy + dz
x y z

P, Q, R must be proprtional to
  
, , ;
x y z

134
  
i.e P = , Q = , R =
x y z

where  is some function of x, y and z.


These three conditions can be reduced to one involving the coefficients P, Q, R and
their derivatives.
On differentiating the first of the these three equations with respect to y and z, the
second with respect to z and x, and the third with respect to x and y, three results,
 P 2  Q ,
P + = =Q +
x y x y x x

 Q 2  R
Q + = =R + ,
z z  y z y z

 R 2  P
R + = =P + ,
x x  y z z z

Or rearranging,
 P Q   
 –  = Q –P ,
 y x  x y

 Q R   
 –  = R –Q ,
 z y  y z

 R P   
 –  = P –R ,
 x z  z x

Multiplying the first of the last three equations by R, the second by P, the third by Q,
and adding the results, we obtain
 Q R   R P   P Q 
P –  + Q –  + R – =0
 z y   x z   y x 

which is the condition to be satisfied by P, Q, R for the equation (2) to posses an integral
of the form (2).
Illustrated Examples [Solution by Inspection]
Example 1. Solve (y + z)dx + dy + dx = 0
Solution. Comparing with
P dx + Q dy + Q dz = 0
we have P = y + z, Q = 1, R = 1
Now the condition of integrability

135
 Q R   R P   P Q 
P –  + Q –  + R – =0
 z y   x z   y x 

L.H.S. (y + z) (0 – 0) + 1 (0 – 1) + 1 (1 – 0)
= –1 + 1 = 0
Thus the condition of integrability is satisfied.
The given equation can be written as
dy + dz
dx + =0
y+z

Integrating
x + (y + z) = c is required solution.
Example 2. Solve (3xz + 2y)dx + x dy + x dz = 0
P = 3xz + 2y, Q = x, R = x2
 Q R   R P   P Q 
P –  + Q –  + R – =0
 z y   x z   y x 

= (3xz + 2y) (0 – 0) + x(2x – 3x) + x2 (2 – 1)


= –x2 + x2 = 0
The equation is integrable.
Multiplying equation by x we have
3x2 zdx + 2xydx + x2dy + x2dx = 0
d(x3 z) + d(x2 y) = 0
Integrating x2z + x2y = c
x2(xz + y) = c, which is the required solution.
Example 3. Solve (y + z)dx + (z + x)dy + (x + y)dz = 0
Here P = y + z, Q = z + x, R = x + y
 Q R 
  P  z –  = (y + z) (1 – 1) + (z + x) (1 – 1) + (x + y) (1 – 1) = 0
y 

The given equation is integrable


Equation can be written as
(y dx + x dy) + (z dx + x dz) + (z dy + y dz) = 0
d(xy + zx + zy) = 0

136
Integrating
xy + zx + zy = c, which is the required solution.
Example 4. Solve zy dx = zx dy + y2dz
Here P = zy, Q = –zx, R = –y2
 Q R 
 P  z –  = zy (–x + 2y) – zx(0 – y) – y2(z + z)
y 

= –zxy + 2zy2xy – 2zy2
=0
The given equation is integrable
Equation can be written as
z(ydx – xdy) – y2dz = 0
Dividing by zy2 we get
ydx – xdy dx
– =0
y2 z
x
d   – d (ln z ) = 0
y
Integrating,
x
– ln z = C, which is the required solution.
y
Exercises
1. yz ln z dx – zx ln z dy + xy dz = 0
2. (a – z) (y dx + x dy) + sy dz = 0
3. (y2 + yz)dx + (zx + z2)dy + (y2 – xy)dz = 0
4. (yz + xyz)dx + (zx + xyz)dy + (xy + xyz)dz = 0
5. (x2 – y3 – y2z)dx + (xy2 – x2z – x3)dy + (xy2 + x2y)dz = 0
Answers
1. xz ln z = cy
2. xy = c (a – z)
3. y (x + z) = c (y + z)
4. ln xyz + x + y + z = c
y+z z+x
5. =C
x y

137
Method of finding the solution of the single integrable equation
Suppose that the condition for the integrability of
Pdx + Qdy + Rdz = 0
is satisfied
Consider any one of the variables, say z, as constant, that is
Step 1. z = constant  dz = 0
The given equation become Pdx + Qdy = 0 ...(1)
Step 2. Integrate the equation Pdx + Qdy = 0,
say its solution is
f(x, y) = c = (z) ...(2)
Step 3. From equations (1) and (2), we obtain
 
P = , Q= ...(3)
x y

 can be obtained from either of these equations.


Step 4. The given equation
 P dx +  Q dy +  R dz = 0 becomes
       
 dx + dy + dz  +   R – dz = 0
 x y z   z 

  
or d +   R – dz = 0
 z 

d  + Sdz = 0

where S = R –
z

Step 5. Solve the equation d  + sdz = 0 and its integral replace  by its values in x, y and
z. The resulting relation will be the required solution.
Remarks

(a) If x = constant, then d  + sdz = 0, S =  P –
x

(b) If y = constant, then d  + sdz = 0, S =  Q –
y

138
Illustrated Examples
The following examples satisfy the condition of intetgrability. The verification is left
to the students.
Example 1. Solve zy dx = zx dy + y2dz
Here P = zx, Q = –zx, R = –y2
Step 1. Let z = constant so that dz = 0
The given equation becomes
zy dx = zx dy
dx dx
Step 2. =  ln (x/y) =  (z) (say) ...(1)
x y

  x
Step 3. P = ,  z y= ln  
x x  y 

1 1
 xy =  =
x xyz

 1
Step 4. Now S =  R – = (– y2 ) – 0
z xyz

y
S= –
xz

The equation
d  + Sdz = 0 becomes
y dz
d – =0
x z

dz
or d – e–  =0
z

x
ln   = 
y

dz
e d  – =0
z

Integrating e – ln z = c
x
Hence – ln z = c which is the required solution.
y

139
Example 2. Solve (2x2 + 2xy + 2xz2 + 1) dx + dy + 2zdz = 0
Solution. P = 2x2 + 2xy + 2xz2 + 1
Q =1
R = 2z
Let us consider x = constant  dx = 0
Given equation becomes
dy + 2zdz = 0
Integrating y + z2 = constant = (x)

Q = = 1  1 = 1   = 1
y


S = P –
x
or S = 2x2 + 2xy + 2xz2 + 1 – 0
or S = 2x2 + 2x(y + z2) + 1 = 2x2 + 2x + 1
The given equation
d  + sdx = 0
d  + (2x2 + 2x  + 1) dx

+ 2x  = –(2x2 + 1)
x
which is a linear equation in x

I.F. = e  e x
2 xdx 2
It has

(  e x = –  (2 x 2 + 1)ex 2 dx + c
2
Solution.
2 2
 e x = – x e x + c (for integrating put) x 2 = t
2
( + x )e x = c
2
( x + y + z 2 )e x = c, which is the required solution.

Example 3. Solve 3x2dx + 3y2dy – (x3 + y3 + e2z)dz = 0


Solution. Here P = 3x2, Q = 3y2 and R = –(x3 + y3 + e2x)
Considering z = constant, i.e. dz = 0, the given differential equation becomes
3x2dx + 3y2dy = 0

140
Integrating x3 + y3 = c1 =  (say) ...(1)
we known

P =   3x 2 = 3x 2   = 1
z


Now S = R–
z

S = 1(x3 + y3 + e2x) – 0
S = –(x3 + y3) – e2x = – –e2x
The equation
d  + Sdz = 0 becomes
d  – ( + e2z)dz = 0

–  e2 z
z

which is a linear differential equation


I.F. = e = e– z
–1dz

Integrating
e–z =  e2 e– z dz + c
e–z = ez + c
(x3 + y3) = e2z + cez, is the required solution.
Note. If in the equation
Pdx + Qdy + Rdz = 0
Coefficients P, Q and R are homogeneous functions of x, y and z, then one of the
variable can be separated from the other two by substitutions.
x = uz and y = vz
Example 4. Solve
yz2 (x2 – yz)dx + zx2 (y2 – zx)dy + xy2(z2 – xy)dz = 0
Solution. Here P, Q and R are homogeneous functions of x, y and z.
Substitute
x = uz and y = vz
differentiating, dx = udz + zdu and dv = vdz + zdv
in the given equation, we obtain
v(u2 – v)du + u2(v2 – u)dv = 0
On dividing by u2v2, we get
du du udv
– + dv – 2 =0
v u2 v

141
vdu – udv du
– 2 + dv =0
v2 u

u  1
Integrating – – +v = c
v  u 

v 1
+ +v =c
u u

Using the above substitutions


x z y
+ + = c is the required solution
y x z

Exercises
1. (ezy + ez)dx + (eyz + ez)dy + (ey – ezy – eyx)dz = 0
2. x dy – y dx – 2x2z dz = 0
3. zy (1 + uzx)dx + xy (1 + 2xz)dz – xydz = 0
4. (x – 2y – z) dx + (2y – 3x) dy + (z – x)dz = 0
5. (y2 + z2 – x3)dx – 2xy dy – 2xy dz = 0
6. xz3dx – zdy + 2y dy = 0
7. 2x(y + z)dx + (2yz – x2 + y2 – z2)dy + (2yz – x2 – y2 + z2)dz = 0
8. (yz + z2)dx – xz dy + xy dz = 0

Answers
1. ezy + eyz + ezx + cez
2. y = x(c – y2)
yz
3. = cx
1 + 2 xz
4. x2 + 2y2 + z2 – 6xy – 2xy = c
y2 + z 2 x 2
5. + =c
x 2
x2 y
6. – =c
2 z2
7. x2 + y2 + z2 = c(y + z)
8. xz = c(y + z)

142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207

You might also like